HK1053037B - Acoustic correction apparatus and methods of acoustic correction - Google Patents

Description

Acoustic correction apparatus and acoustic correction method

Technical Field

The present invention relates generally to audio enhancement systems and, more particularly, to systems and methods designed to improve the realism of stereo sound reproduction. More particularly, the present invention relates to apparatus for overcoming acoustic imaging and frequency response deficiencies of acoustic systems as perceived by a listener.

Background

In a sound reproduction environment, there are a variety of factors that can act to reduce the quality of reproduced sound perceived by a listener. These factors render the reproduced sound distinct from that of the original sound field (sound stage). One such factor is the location of the speakers in the sound field, which if improperly set may result in distortion of the sound pressure response characteristics across the audio spectrum. The placement of the loudspeakers also affects the perceived width of the sound field. For example, speakers as point sources limit their ability to reproduce reverberant sounds that are easily perceived in a live sound stage. In fact, the perceived sound field width in many audio reproduction systems is limited to the distance separating a pair of loudspeakers placed in front of the listener. Another factor that degrades the quality of reproduced sound may come from a microphone that records sound in a manner different from the way the human auditory system perceives sound. In an attempt to overcome these factors that degrade the quality of reproduced sound, numerous efforts have been made to continually modify the characteristics of the sound reproduction environment to simulate the sound heard by a listener in a live sound stage.

Some efforts at enhancing stereo imaging have focused on the hearing capabilities and limitations of the human ear. The auditory response of the human ear is sensitive to the sound intensity, the phase difference between certain sounds, the frequency of the sound itself, and the direction from which the sound is emitted. Although the human auditory system is complex, the frequency response of the human ear is relatively stable from person to person.

When a sound wave having a stable sound pressure level at all frequencies is directed from a single location to a listener, the human ear will respond differently to the individual frequency components of the sound. For example, when equal-pressure sounds are directed to a listener from the front of the listener, a 1000 hertz sound will produce a different sound pressure level in the ears of the listener than a 2000 hertz sound.

In addition to sensitivity to frequency, the human auditory system responds differently to sounds striking the human ear from different angles. In particular, the sound pressure level in the human ear will vary with the direction of the sound. The shape of the outer ear or pinna and the inner ear canal is largely responsible for the frequency profiling (transducing) of directionally-varying sounds.

The human auditory response is sensitive to changes in the azimuth and elevation of the sound source. This is particularly true for complex sound signals, i.e., signals having multiple frequency components, which are generally more sensitive to higher frequency components. Sound pressure variations in the in-ear frequency components are interpreted by the brain to provide an indication of the source of the sound. Thus when reproducing a recorded sound, the indication of the direction of origin of the sound will depend on the actual position of the loudspeaker reproducing the sound, as the ear interprets from the sound pressure information.

The ears of the listener can obtain a stable sound pressure level, i.e., a "flat" sound pressure-frequency response characteristic, from the speakers disposed directly in front of the listener. This response characteristic is often desirable to achieve realistic sound images. However, the mass of a group of loudspeakers may be somewhat less than ideal and they may not be placed in the most ideal acoustic position. Both of these factors often result in a disrupted sound pressure characteristic. Prior art sound systems have disclosed methods for "correcting" the sound pressure emanating from the speakers to produce a spatially corrected response, thereby improving the resulting sound image.

In order to achieve a better spatially corrected response for a given sound system, it is known to select and apply Head Related Transfer Functions (HRTFs) to the audio signal. HRTFs are based on the acoustic effects of the human auditory system. The application of HRTFs is used to adjust the amplitude of portions of an audio signal to compensate for spatial distortion. HRTF-based principles can also be used to re-localize a stereo image produced from non-optimally placed speakers.

The second drawback also often occurs because it is difficult to sufficiently reproduce low-frequency sounds, such as bass sounds. Various conventional approaches to improving the output of low frequency sound include the use of higher quality speakers with larger cone areas, larger magnets, larger cabinets, or larger cone vibration capabilities. In addition, conventional systems attempt to reproduce low frequency sound using a resonant cavity and a horn to match the acoustic impedance of the speaker to the acoustic impedance of the free space surrounding the speaker.

However, not all systems can simply reproduce low frequency sound with more expensive or powerful loudspeakers. For example, some conventional sound systems, such as compact audio systems and multimedia computer systems, employ small speakers. In addition, many audio systems employ less accurate speakers in order to save costs. Such speakers typically do not have good ability to reproduce low frequency sounds, and thus such sounds are often less powerful or enjoyable than systems that more accurately reproduce low frequency sounds.

Some conventional enhanced systems attempt to compensate for poor reproduction of low frequency sounds by amplifying the low frequency signals before the signals are input to the speaker. Amplifying the low frequency signal sends a large amount of energy to the speaker, thereby driving the speaker with a greater force. However, such attempts to amplify low frequency signals can result in overdriving the speaker. Unfortunately, overdriving the speaker can increase background noise, create annoying distortions, and damage the speaker.

Yet another conventional system, when attempting to compensate for the lack of low frequencies, distorts the high frequency reproduction by adding undesirable acoustic coloration.

A third difficulty arises because sound emanating from multiple locations often cannot be properly reproduced in an audio system. A method of improving sound reproduction includes a surround sound system having a plurality of recording tracks. Spatial information associated with sounds emitted from a plurality of positions is recorded using a plurality of recording tracks.

For example, in a surround sound system, some recording tracks contain sound originating from the front of a listener, while other recording tracks contain sound originating from the back of the listener. When a plurality of speakers are provided around the listener, recording the audio information contained in the track makes the reproduced sound feel more realistic to the listener. However, such systems are generally more expensive than systems that do not employ multiple recording tracks and multiple speaker settings.

To save cost, many conventional dual speaker systems attempt to simulate the surround sound effect by introducing an unnatural time delay or phase shift between the left and right signal sources. Unfortunately, such systems often encounter unrealistic effects in reproducing sound.

Another commonly used prominence enhancement technique operates on so-called "sum" and "difference" signals. The sum signal, also called mono signal, is the sum of the left and right signals. This can be conceptualized as superimposing or combining the left and right signals (L + R).

On the other hand, the difference signal represents the difference between the left and right audio signals. It is most appropriately conceptualized as subtracting the right signal (L-R) from the left signal. The difference signal is also often referred to as an ambient signal.

It is well known that modifying certain frequencies in the difference signal can broaden the perceived sound projected from the left and right speakers. The widened sound image is usually obtained by changing the reverberant sound present in the difference signal.

However, the circuit that generates the sum signal and the difference signal by processing the left and right input signals. Furthermore, once the circuit generates the sum and difference signals, additional circuitry processes and recombines the sum and difference signals, respectively, to produce sound effects that are enhanced.

In general, generation and processing of the sum signal and the difference signal are realized by a digital signal processor, an operational amplifier, and the like. Such implementations typically require complex circuitry, thereby increasing the cost of such systems. Thus, despite the improvements in the prior art, there remains a need for a simplified audio enhancement system that reduces the costs associated with producing enhanced auditory effects.

Disclosure of Invention

The present invention addresses these and other problems by providing a signal processing technique that significantly improves the sound image size, bass performance, and dynamics of an audio system, placing listeners in the engaging and impressive performance of an audio performance. It improves the listening experience for a variety of applications including computers, multimedia, televisions, compact recorders, automobiles, home sound systems, and portable audio systems. In one embodiment, the sound correction system corrects for the extrinsic placement of the speakers, the sound image generated by the speakers, and the low frequency response characteristics produced by the speakers. In one embodiment, the sound correction system enhances the spatial and frequency response characteristics of sound reproduced by two or more speakers. The sound correction system includes a sound image correction module correcting a vertical sound image perceived by a listener of a sound reproduced by a speaker; a bass boost module to improve the bass response characteristics of the speaker as perceived by a listener; and a sound image enhancement module that enhances a horizontal sound image of the apparent sound field perceived by the listener.

In one embodiment, three processing techniques are employed. A Head Related Transfer Function (HRTF) is employed to equalize the spatial signal for localizing sounds outside the loudspeaker boundaries. These HRTF correction curves illustrate how the brain perceives the sound to be positioned on each side of the listener, even when played through speakers in front of the listener. Thus, with the addition of indirect and reflected sound surrounding the space, the appearance of the instrument and singer is now in place. The second set of HRTF correction curves expands and elevates the apparent size of the stereo image such that the sound field width appears in a large proportion to the position of the loudspeakers. Finally, bass performance is enhanced by psychoacoustic techniques that restore the perception of low frequency pitch by dynamically increasing the pitch that a loudspeaker can more easily reproduce.

The acoustic correction system and associated method of operation provide a sophisticated and effective system for improving vertical, horizontal and spectral sound images in poor reproduction environments. In one embodiment, the system first corrects the vertical sound image produced by the speakers, then boosts bass sounds, and finally corrects the horizontal sound image. Vertical sound image enhancement typically includes some enhancement of the lower frequency portions of the sound, and the vertical enhancement thus provided prior to bass enhancement has an effect on the overall effectiveness of the bass enhancement process. Bass enhancement provides some mixing of the common (common mode) portions of the left and right portions of low frequency information in the stereo signal. In contrast, horizontal image enhancement provides some enhancement and shaping of the difference (differential mode) between the left and right portions. Thus, in one embodiment, bass enhancement is advantageously provided prior to horizontal image enhancement in order to balance the common and differential mode portions of the stereo signal, thereby creating an effect that is pleasing to the listener.

In order to achieve an improved stereoscopic image on the vertical plane, the image-sound correcting apparatus divides an input signal into first and second frequency ranges that collectively contain substantially all of the audio spectrum. The frequency response characteristics of the input signals in the first and second frequency ranges are corrected and combined, respectively, to produce an output signal having a frequency response characteristic that is relatively smooth to a listener. The level of frequency correction, i.e., acoustic energy correction, depends on the reproduction environment and is suitable for overcoming the acoustic limitations of such an environment. The design of the acoustic correction apparatus makes it possible to easily and independently perform correction of input signals in respective frequency ranges, thereby achieving spatial correction and a relocated sound image.

Within an audio reproduction environment, the loudspeakers may be in unsuitable positions, thereby adversely affecting the sound image perceived by the listener. For example, headphones often produce an unpleasant acoustic image because the transducer is in close proximity to the listener's ear. The acoustic correction apparatus of the present invention repositions the acoustic image at a more desirable apparent location.

By applying the acoustic correction device, the stereo image produced by the playback of the audio signal can be spatially corrected so as to convey a perceived source whose vertical and/or horizontal position differs from the position of the loudspeakers. The exact source perceived by the listener will depend on the degree of spatial correction.

Once a perceived sound source is obtained by correction of spatial distortion, the corrected audio signal may be enhanced to provide an extended stereo image. According to one embodiment, the stereo image enhancement of the re-localized image takes into account the acoustic principles of human hearing, thereby leaving the listener in a realistic stage. In those sound reproduction environments where the listening position is relatively fixed (e.g., car interiors, multimedia computer systems, bookshelf speaker systems, etc.), the amount of stereo image enhancement applied to the audio signal is determined in part by the actual position of the speakers relative to the listener.

In loudspeakers that do not reproduce some low frequency sounds, the present invention creates the illusion that there is indeed lost low frequency sound. Therefore, the listener perceives low frequencies lower than the frequencies that the speaker can actually reproduce accurately. This illusive effect is achieved by processing the sound in a unique way using the human auditory system.

One embodiment of the present invention takes advantage of the way listeners psychologically perceive music or other sounds. The processing of sound reproduction is not limited to only the sound energy generated by the speaker, but also includes the ear, auditory nerve, brain, and mental processes of the listener. Hearing begins with the action of the ear and auditory nervous system. The human ear can be viewed as a sophisticated translation system that receives acoustic vibrations, converts these vibrations into nerve impulses, and ultimately into the "feel" or perception of sound.

Some embodiments of the present invention advantageously utilize the way the human ear processes the overtones and harmony of low frequency sounds, creating the perception that low frequency sounds are being emitted from a speaker and are not present. In some embodiments, frequencies in higher frequency bands are selectively processed to create the illusion of low frequency signals, while in other embodiments some of the higher frequency bands are modified with multiple filter functions.

In addition, some embodiments of the present invention are designed to enhance low frequency enhancement of popular audio program material (e.g., music). Most music is rich in harmony. Thus, these embodiments may modify a wide variety of music types to take advantage of the way the human ear processes low frequency sounds. Advantageously, music in an existing format can be processed to produce a desired effect.

This new method results in a number of distinct advantages. Because the listener perceives low frequency sounds that are not actually present, the need for large loudspeakers, larger cone wobble amplitudes or increased loudspeakers is reduced. Thus, in one embodiment, small speakers may feel as if they are low frequency sounds emitting larger speakers. As can be expected, this embodiment produces the perception of low frequency sounds, such as bass sounds, in a sound environment that is too small for large speakers. Large loudspeakers are also benefited by creating a perception that they produce enhanced low frequency sound.

Furthermore, with one embodiment of the present invention, small speakers in handheld and portable sound systems can create a more enjoyable perception of low frequency sound. Therefore, the listener does not need to sacrifice the quality of the low-frequency sound for portability.

In one embodiment of the invention, a low cost speaker creates the illusion of low frequency sound. Many low-cost speakers do not adequately reproduce low-frequency sound. Instead of using an expensive speaker cabinet, high performance components, and large magnets to actually reproduce low frequency sound, one embodiment uses higher frequency sound to create the illusion of low frequency sound. Thus, a lower cost speaker can be employed to create a more realistic and robust audio visual experience.

Furthermore, in one embodiment, the illusion of low frequency sound creates a higher audio-visual experience that increases the realism of the sound. Thus, instead of the blurred or shaken reproduction of low frequency sounds as is the case with many low cost prior art systems, one embodiment of the present invention reproduces sounds that are perceived to be more accurate and clear. Such low cost audio and audiovisual devices may include, for example, radios, mobile audio systems, computer games, speakers, Compact Disc (CD) players, Digital Versatile Disc (DVD) players, multimedia display devices, computer sound cards, and the like.

In one embodiment, creating the illusion of low frequency sounds requires less energy than actually reproducing the low frequency sounds. Thus, a system that operates using batteries, operates in a low power environment, small speakers, multimedia speakers, headphones, etc., can create the illusion of low frequency sounds without consuming as much valuable energy as a system that merely amplifies or boosts the low frequency sounds.

Other embodiments of the present invention utilize specialized circuitry to create the illusion of low frequency signals. These circuits are simpler than prior art low frequency amplifiers and therefore reduce manufacturing costs. Advantageously, these costs are lower than prior art sound enhancement devices that add complex circuitry.

Yet another embodiment of the present invention relies on a microprocessor that implements the disclosed low level enhancement techniques. In some cases, existing components that process audio may be reprogrammed to implement the unique low frequency signal enhancement techniques disclosed in one or more embodiments of the invention. Thus, the cost of adding low frequency enhancement to existing systems is significantly reduced.

In one embodiment, a sound enhancement device receives one or more input signals from a host system and generates one or more enhanced output signals. Specifically, the two input signals are processed to provide a pair of spectrally enhanced output signals that, when played through a speaker and heard by a listener, produce an extended bass sensation. In one embodiment, the low frequency audio information is modified in a different manner than the high frequency audio information.

In one embodiment, a sound enhancement device receives one or more input signals and generates one or more enhanced output signals. Specifically, the input signal includes a waveform having a first frequency range and a second frequency range. The input signal is processed to provide an enhanced output signal, which when played through a speaker and heard by a listener, produces an expanded bass sensation. Furthermore, the embodiment may modify the information in the first frequency range in a different manner than the information in the second frequency range. In some embodiments, the first frequency range may be bass frequencies that are too low for a desired speaker to reproduce, and the second frequency range may be bass frequencies that the speaker can reproduce.

One embodiment modifies the audio information common to both stereo channels in a manner that is different from the energy not common to both channels. The audio information common to both input signals is referred to as a combined signal. In one embodiment, the enhancement system spectrally shapes the amplitude of the phase and frequency of the combined signal to reduce clipping that may be caused by high amplitude input signals without removing the stereo perception of the audio information.

As discussed in more detail below, one embodiment of the sound enhancement system spectrally shapes the combined signal with various filters to produce an enhanced signal. By enhancing selected frequency bands within the combined signal, this embodiment provides a perceived speaker bandwidth that is wider than the actual speaker bandwidth.

One embodiment of the sound enhancement apparatus comprises feed forward signal paths for two stereo channels and three parallel filters for combining the signal paths. Each of the four parallel filters comprises a sixth order bandpass filter consisting of three biquad filters connected in series. The transfer functions of these four filters are specifically selected to provide phase and/or amplitude shaping of various harmonics of the low frequency content of the audio signal. This shaping unexpectedly increases the perceived bandwidth of the audio signal when played through the speaker. In another embodiment, the sixth order filter is replaced by a lower order Chebychev filter.

Because spectral shaping is performed on the combined signal, which is then combined with the stereo information in the feed-forward path, the frequencies in the combined signal can be changed, affecting two stereo channels, some signals in some frequency ranges being coupled from one stereo channel to the other. Thus, the various embodiments create enhanced audio sounds in a completely unique, novel, and unexpected manner.

Which in turn may be connected to one or more subsequent signal processing stages. These subsequent stages may enable improved sound field or spatial processing. The output signal may also be directed to other audio devices, such as recording devices, power amplifiers, speakers, etc., without affecting the operation of the sound enhancement device.

The present invention also provides a unique differential perspective correction system to improve the horizontal aspect of the sound field. The differential perspective correction system enhances sound in a manner completely different from other sound enhancing devices. Such perspective correction system embodiments may be advantageously used to enhance sound in a variety of low cost audio and audiovisual devices, which may include, for example, radios, mobile audio systems, computer games, multimedia display devices, and the like.

Generally, such differential perspective correction devices receive two input signals from a host system and then produce two enhanced output signals. Specifically, the two input signals are processed in unison to provide a pair of spatially corrected output signals. Further, one embodiment modifies the audio information common to the two input signals in a manner that is different from the audio information not common to the two input signals.

The audio information common to the two input signals is referred to as common mode information or a common mode signal. Common mode audio information differs from the sum signal in that it does not contain the sum of the input signals, but only the audio information present in both input signals at any given moment.

In contrast, audio information that is not common to both input signals is referred to as differential information or differential signal. Although differential information is processed differently than common mode information, the differential information is not a discrete signal. As discussed in detail below, the differential perspective correction apparatus spectrally shapes the differential signal using various filters to produce an equalized differential signal. By equalizing the selected frequency bands within the differential signal, the differential perspective correction apparatus widens the perceived sound image projected from a pair of speakers provided in front of the listener.

Because the crossover impedance network equalizes the frequency range in the differential inputs, the frequency in the differential signals can be changed without affecting the frequency in the common mode signal. Thus, the audio sound is enhanced in a completely unique and novel way.

Drawings

The above and other aspects, features and advantages of the present invention will become apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings.

Fig. 1 is a block diagram of a stereo image correction system operatively connected to a stereo enhancement system and a bass enhancement system for producing a realistic stereo image from a pair of input stereo signals.

Fig. 2 is a schematic diagram of a stereo system comprising a stereo receiver and two loudspeakers.

Fig. 3 is a schematic diagram of a typical multimedia computer system.

Fig. 4A is a graphical representation of a desired sound pressure-frequency characteristic of an audio reproduction system.

Fig. 4B is a graphical representation of sound pressure versus frequency characteristics corresponding to a first audio reproduction environment.

Fig. 4C is a graphical representation of sound pressure versus frequency characteristics corresponding to a second audio reproduction environment.

Fig. 4D is a graphical representation of sound pressure-frequency characteristics corresponding to a third audio reproduction environment.

Fig. 5 is a schematic block diagram of an energy correction system operatively connected to a stereo image enhancement system for producing a realistic stereo image from a pair of input stereo signals.

FIG. 6A is a graphical representation of signal modifications at various levels provided by the low frequency correction system, according to one embodiment.

FIG. 6B is a graphical representation of signal modifications at various levels provided by a high frequency correction system for boosting high frequency components of an audio signal, according to one embodiment.

FIG. 6C is a graphical representation of signal modifications at various levels provided by a high frequency correction system for attenuating high frequency components of an audio signal, according to one embodiment.

Fig. 6D is a graphical representation of a composite energy correction curve illustrating the possible range of sound pressure correction for repositioning the stereo image.

Fig. 7 is a graphical representation of the levels of equalization applied to the audio difference signal to achieve varying amounts of stereo image enhancement.

Fig. 8A is a schematic diagram illustrating the perceived and actual origin of sound heard by a listener from a loudspeaker disposed at a first location.

Fig. 8B is a schematic diagram illustrating the perceived and actual origin of sound heard by a listener from a speaker disposed at a second location.

Fig. 9 is a graph of the frequency response characteristic of a typical small speaker system.

Fig. 10 illustrates the actual signal spectrum and the perceived signal spectrum represented by two discrete frequencies.

Fig. 11 illustrates an actual signal spectrum and a perceived signal spectrum represented by continuous spectra.

Fig. 12A illustrates a time waveform of a modulated carrier.

Fig. 12B illustrates the time waveform of fig. 12A after detection by the detector.

Fig. 13A is a block diagram of a sound system having a bass enhancement processing function.

Fig. 13B is a block diagram of a bass enhancement processor that combines multiple channels into a single bass channel.

Fig. 13C is a block diagram of a bass enhancement processor that separately processes multiple channels.

Fig. 14 is a signal processing block diagram of a system that provides selectable frequency response characteristics for bass boost functions.

Figure 15 is a graph of the transfer function of the bandpass filter used in the signal processing schematic shown in figure 14.

Fig. 16 is a time domain graph showing the time amplitude response characteristic of the punch-through (punch) system.

FIG. 17 is a time domain plot of the signal and envelope portions of a typical bass note played by an instrument, where the envelope shows a heightened portion, a decaying portion, a sustained portion and a released portion.

Fig. 18 is a signal processing block diagram of a system for providing bass enhancement using a peak compressor and a bass pass-through system.

Fig. 19 is a time domain graph illustrating the effect of a peak compressor on an envelope with a rapid increase.

Fig. 20 is a conceptual block diagram of a stereo image (differential perspective) correction system.

Fig. 21 is a block diagram of a stereo image (differential perspective) correction system that does not derive explicit sum and difference signals.

FIG. 22 illustrates a graphical representation of the common-mode gain of a differential perspective correction system.

FIG. 23 is a graphical representation of the overall differential signal equalization curve of the differential perspective correction system.

FIG. 24 is a block diagram of one embodiment of a sound enhancement system that may be implemented on a single chip.

Fig. 25A is a schematic diagram of the left channel of the vertical sound image enhancement block suitable for the system shown in fig. 24.

Fig. 25B is a schematic diagram of the right channel of the vertical sound image enhancement block suitable for the system shown in fig. 24.

Fig. 26A and 26B are schematic diagrams of a bass enhancement block suitable for use in the system shown in fig. 24.

Fig. 27 is a schematic diagram of a filter system suitable for use in the bass enhancement system shown in fig. 26.

Fig. 28 is a schematic diagram of a compressor system suitable for use in the bass enhancement system shown in fig. 26A and 26B.

Fig. 29 is a schematic diagram of a horizontal sound image enhancing block suitable for the system shown in fig. 24.

Fig. 30 is a schematic diagram of a differential perspective correction system that can be used as a stereo image enhancement system.

FIG. 31 shows a differential perspective correction system using a crossover network.

Fig. 32 is a schematic diagram of a differential perspective correction apparatus using two crossover networks.

Fig. 33 shows a differential perspective correction device that allows a user to change the amount of overall differential gain.

FIG. 34 illustrates a differential perspective correction device that allows a user to vary the amount of common mode gain.

Fig. 35 illustrates a differential perspective correction apparatus having a first frequency-dividing network between the emitters of the transistors of a differential pair and a second frequency-dividing network between the collector stages of the differential pair.

FIG. 36 illustrates a differential perspective correction device with output buffers.

Fig. 37 illustrates a six-operational amplifier form of the image enhancement system.

FIG. 38 is a block diagram of a software embodiment of an acoustic correction system.

Figure 39 is a graph of the transfer function of a 40hz bandpass filter used in the block diagram of figure 38.

Figure 40 is a graph of the transfer function of the 60hz bandpass filter used in the block diagram of figure 38.

Figure 41 is a graph of the transfer function of the 100hz bandpass filter used in the block diagram of figure 38.

Figure 42 is a graph of the transfer function of the 150hz bandpass filter used in the block diagram of figure 38.

Figure 43 is a graph of the transfer function of the 200hz bandpass filter used in the block diagram of figure 38.

Fig. 44 is a graph of the transfer function of the low pass filter used in the block diagram of fig. 38.

Detailed Description

Fig. 1 is a block diagram of an acoustic correction apparatus 120, which includes a stereo image correction system 122, a bass enhancement system 101, and a stereo image enhancement system 124 in series. The sound image correction system 122 supplies the left stereo signal and the right stereo signal to the bass enhancement unit 101. The bass boost unit outputs left and right stereo signals to respective left and right input terminals of the stereo image enhancement device 124. The stereo image enhancement system 124 processes these signals and provides a left output signal 130 and a right output signal 132. The output signals 130 and 132 may in turn be connected to some other form of signal conditioning system, or they may be connected directly to speakers or headphones (not shown).

When connected to a speaker, the correction system 120 corrects for the placement of the speaker, the sound image generated by the speaker, and imperfections in the low frequency response characteristics produced by the speaker. The acoustic correction system 120 enhances the spatial and frequency response characteristics of the sound reproduced by the speaker. In the audio correction system 120, the sound image correction module 122 corrects a vertical sound image perceived by a listener of an apparent sound field reproduced by speakers, the bass enhancement module 101 improves a bass response characteristic perceived by the listener of sound, and the sound image enhancement module 124 enhances a horizontal sound image perceived by the listener of the apparent sound field.

The correction means 120 improves the sound reproduced by the speaker by compensating for defects in the sound reproduction environment and defects of the speaker. The apparatus 120 improves the reproduction of the original sound field by compensating for the positioning of the loudspeakers in the reproduction environment. Sound field reproduction is improved in the following way: while enhancing the apparent (i.e., reproduced) sound field in both the horizontal and vertical aspects of the audio spectrum. The device 120 advantageously modifies reverberant sounds that are readily perceived in a live stage, so that the listener also perceives the reverberant sounds in the reproduction environment, even though the loudspeakers only function as point sound sources of limited capability. The device 120 also compensates for the fact that the microphone often records sound in a manner different from the way the human auditory system perceives sound. The device 120 corrects the sound generated by the microphone using filters and transfer functions that simulate human hearing.

The sound system 120 utilizes characteristics of the human auditory response to adjust the apparent bearing and elevation points of the complex sound. This correction is used by the listener's brain to provide an indication of the source of the sound. The correction means 120 also corrects for speakers that are placed less optimally, such as speakers that are not in the optimal sound effect position.

In order to achieve a more spatially correct response characteristic for a given sound system, the acoustic correction device 120 uses certain aspects of the Head Related Transfer Function (HRTF) to correct the placement of the two speakers in conjunction with frequency response characteristic shaping of the sound information, correct the apparent width and height of the sound field, and correct deficiencies in the low frequency response of the speakers.

Therefore, the acoustic correction apparatus 120 provides a more natural and realistic sound field to the listener even if the positions of the speakers are slightly less ideal, and the speakers themselves are not sufficient to reproduce the desired sound well.

The various sound corrections provided by the correction device are provided in such an order that subsequent corrections do not interfere with previous corrections. In one embodiment, these corrections are provided in a desired order such that the prior corrections provided by the apparatus 120 enhance and assist the subsequent corrections provided by the apparatus 120.

In one embodiment, the correction device 120 simulates a surround sound system with improved bass response characteristics. The correction device 120 creates the illusion that the plurality of speakers are positioned around the listener and provides the plurality of speaker arrangements with the audio information contained in the plurality of recorded trajectories.

The acoustic correction system 120 provides a sophisticated and effective system to improve vertical, horizontal and spectral sound images in defective reproduction environments. The sound image correction system 122 first corrects the vertical sound image produced by the speakers, then the bass enhancement system 101 adjusts the low-frequency components of the sound signals in such a manner as to enhance the low-frequency output of small speakers that do not have sufficient low-frequency reproduction capability, and finally, the sound image correction system 124 corrects the horizontal sound image.

The vertical sound image enhancement provided by the sound image correction system 122 typically includes some emphasis of the lower frequency portions of the sound and thus provides vertical enhancement before the bass enhancement system 101 affects the overall effectiveness of the bass enhancement process. The bass enhancement system 101 provides some mixing of the common part (common mode) of the left and right parts of the low frequency information in the stereo signal. In contrast, the horizontal image enhancement provided by the image enhancement system 124 provides enhancement and shaping of the difference (differential mode) between the left and right portions of the signal. Thus, in the correction system 120, it is preferable to provide bass enhancement before horizontal image enhancement so as to balance the common mode portion and the differential mode portion of the stereo signal, thereby obtaining an effect that is pleasant to the listener.

As disclosed above, the stereo image correction system 122, the bass enhancement system 101, and the stereo image enhancement system 124 cooperate to overcome the acoustic deficiencies of the sound reproduction environment. The sound reproduction environment may be as large as a theater installation or as small as a portable electronic keyboard. The sound correction apparatus also provides great benefits for multimedia computer systems (see e.g. fig. 3), home stereo, television, headphones, pocket recorders, automobiles, etc.

Fig. 2 shows a stereo audio system with a receiver 220. Receiver 220 provides left channel signals to left speaker 246 and right channel signals to right speaker 247. Alternatively, receiver 220 may be replaced with a television, a portable stereo system (e.g., a portable radio), a radio alarm clock, etc. The receiver 220 also provides left and right channel signals to the headphones 250. A listener (user) 248 can listen to left and right channel signals using headphones 250 or speakers 246 and 247. Acoustic correction apparatus 120 may be implemented using analog devices in receiver 220 or by software running on a Digital Signal Processor (DSP) in receiver 220.

The speakers 246 and 247 are often not disposed at optimal positions to provide a desired stereoscopic image to the user, and thus, the listening pleasure of the listener is reduced. Likewise, headphones, such as headphone 250, often produce unpleasant sounds because the headphones are close to the ear, rather than in front of the listener. Furthermore, many small bookshelf speakers, multimedia speakers, and earphones have poor low frequency response characteristics, further reducing the listening pleasure of the listener. The acoustic correction device (or software) 120 within the receiver 220 corrects the left and right signals to produce more pleasing sounds when reproduced through the speakers 246 and 247 or the headphones 250. In one embodiment, the receiver 220 includes controls (e.g., width controls 3846 in FIG. 38 and/or bass controls 3827 in FIG. 38) that allow the listener 248 to adjust the sound produced in the left and right channels depending on whether the listener 248 listens to the speakers 246 and 247 or the headphones 250.

Fig. 3 illustrates an exemplary computer audio system 300 that may utilize embodiments of the present invention to improve the audio performance produced by speakers 246 and 247. Speakers 246 and 247 are typically connected to a sound card (not shown) in computer device 304. The sound card may be any computer interface card that produces an audio output, including a radio card, a television channel selection card, a PCMCIA (personal computer memory card international association) card, a built-in modem, a plug-in Digital Signal Processor (DSP) card, etc. The computer 304 causes the sound card to generate an audio signal that is converted to sound waves by the speaker 246.

Fig. 4A illustrates a graphical representation of a desired frequency response characteristic occurring at the outer ear of a listener in an audio reproduction environment. Curve 460 is the change in Sound Pressure Level (SPL) versus frequency measured in decibels. As shown in fig. 4A, the sound pressure level is relatively constant for all audio frequencies. Curve 460 may result from reproduction of pink noise by an ideal pair of speakers placed directly in front of the listener, approximately level with the ears. Pink noise refers to sound emitted over an audio spectrum with equal energy per octave. In practice, the smoothed frequency response of curve 460 may fluctuate with the inherent acoustic limitations of the speaker system.

Curve 460 represents the sound pressure level that existed prior to treatment by the listener's ear. Referring again to fig. 2, when the speakers are spaced apart and generally placed in front of the listener 248, the smooth frequency response characteristic represented by curve 460 is consistent with the sound emitted to the listener 248. The human ear processes the sound signal by applying its own auditory response characteristics to the sound, as represented by curve 460. The human auditory response characteristics are determined by the external pinna and the internal ear canal of the ear.

Unfortunately, the frequency response characteristics of many home and automotive sound reproduction systems do not provide the desired features shown in fig. 4A. Instead, the speakers may be placed in locations where sound effects are not ideal to accommodate other ergonomic requirements. It is possible that the location of the speakers 246 and 247 relative to the listener 248 alone may spectrally distort the sound emanating from the speakers 246 and 247. Furthermore, objects and surfaces in the audiovisual environment may cause the resulting sound signal to be absorbed or distorted in amplitude. This absorption is often common in higher frequencies.

Since both the spectrum and the amplitude are distorted, the stereo image perceived by the listener 248 is spatially distorted, thereby creating an undesirable listening experience. Fig. 4B-4D graphically illustrate the degree of spatial distortion for various sound reproduction systems and listening environments. The distortion characteristics illustrated in fig. 4B-4D represent sound pressure levels measured in decibels near the ear of a listener.

The frequency response curve 464 of fig. 4B drops in sound pressure level at frequencies above about 100 Hz. Curve 464 shows a possible sound pressure characteristic produced by a loudspeaker containing a woofer mounted below the listener. For example, assuming that speaker 246 of fig. 2 contains a tweeter, audio signals played only through such speaker 246 may exhibit the response of fig. 4B.

The particular slope associated with the droop curve 464 may vary and may not be exactly a straight line, depending on the audio-visual region, the quality of the speaker, and the particular placement of the speaker within the audio-visual region. For example, an audiovisual environment having a relatively hard surface may reflect audio signals, particularly at higher frequencies, more than an audiovisual environment having a relatively soft surface (e.g., cloth, carpet, acoustical tiles, etc.). The degree of spectral distortion will vary as the speakers are placed away from the listener.

Fig. 4C is a graphical representation of a sound pressure-frequency characteristic 468 in which a first frequency range of an audio signal has been spectrally distorted, but the higher frequency ranges of these signals have not been distorted. The characteristic 468 may be realized by the following speaker arrangement: the mid-low frequency speakers are arranged below the listener and the high frequency speakers are arranged near the horizontal line of the ears of the listener. The sound image resulting from the characteristic curve 468 will have low frequency components located below the listener 248 of fig. 2 and high frequency components located near the horizontal line of the listener's ears.

Fig. 4D is a graphical representation of a sound pressure versus frequency characteristic 470 that decreases in sound pressure level at lower frequencies and increases in sound pressure level at higher frequencies. Property 470 is implemented as follows: the mid-low frequency speakers are disposed below the listener and the high frequency speakers are disposed above the listener. As shown by curve 470 of fig. 4D, sound pressure levels at frequencies above 1000Hz may be significantly higher than lower frequencies, resulting in sound effects that are not ideal for nearby listeners. The sound image resulting from the characteristic curve 470 will have low frequency components located below the listener 248 of fig. 2 and high frequency components located above the listener 248.

The audio characteristics of fig. 4B-4D represent various sound pressure levels that are available in a typical listener environment and that are heard by the listener 248. The audio response curves of fig. 4B-4D are just a few examples illustrating how an audio signal located near a listener's ear may be distorted by various audio reproduction systems. The exact degree of spectral distortion at any given frequency varies widely depending on the reproduction system and reproduction environment. An apparent position may be generated for the speaker system defined by apparent elevation and azimuth coordinates that are different from the actual speaker position relative to a stationary listener.

Fig. 5 is a block diagram of the stereo image correction system 122, which inputs left and right stereo signals 126 and 128. The image correction system 122 corrects the distorted spectral density of various sound systems by advantageously separating the audio frequency spectrum into a first frequency component containing relatively low frequencies and a second frequency component containing relatively high frequencies. Each of the left and right signals 126 and 128 is separately processed by a respective low frequency correction system 580 and 582 and a high frequency correction system 584 and 586. It should be noted that in one embodiment, correction systems 580 and 582 will operate at a relatively "low" frequency range of approximately 100 to 1000 hertz, while correction systems 584 and 586 will operate at a relatively "high" frequency range of approximately 1000 to 10000 hertz. This is not to be confused with commonly used audio terminology, in general low frequencies referring to frequencies up to 100hz, medium frequencies referring to frequencies between 100 and 4khz and high frequencies referring to frequencies above 4 khz.

By separating the lower and higher frequency components of the input audio signal, the correction of the sound voltage level can be performed over a frequency range independent of each other. Correction systems 580, 582, 584 and 586 modify input signals 126 and 128 to correct for spectral and amplitude distortion of the input signals as reproduced by the speakers. The resulting signal and the original input signals 126 and 128 are combined at respective summing nodes 590 and 592. Corrected left stereo signal L_cAnd a corrected right stereo signal R_cAlong an output to the bass enhancement unit 101.

The corrected stereo signal provided to the bass unit 101 has a smooth, i.e. uniform, frequency response that appears at the ears of the listener 248 (as shown in fig. 2 and 3). This spatially corrected response creates an apparent sound source that appears to be directly in front of the listener 248 when played through the speakers 246 of fig. 2 or 3.

Once the sound source is properly located by energy correction of the audio signal, the bass enhancement unit 101 corrects for low frequency defects in the speakers 246 and provides the corrected bass left and right channel signals to the stereo enhancement system 124. The stereo enhancement system 124 adjusts the stereo signal to broaden the stereo image emanating horizontally from the apparent sound source. As discussed with reference to fig. 8A and 8B, the stereo image enhancement system 124 may be adjusted by the stereo directional device to compensate for the actual location of the sound source.

In one embodiment, the stereo enhancement system 124 equalizes different signal information present in the left and right stereo signals.

The left and right signals supplied from the bass enhancement unit 101 are input through the enhancement system 124 and supplied to the difference signal generator 501 and the sum signal generator 504. Difference signal (L) representing stereo content of corrected left and right input signals_c-R_c) Is provided at the output 502 of the difference signal generator 501 and represents the sum signal (L) of the sum of the corrected left and right input stereo signals_c+R_c) Is produced at the output 506 of the sum signal generator 504.

The sum and difference signals at outputs 502 and 506 are provided to optional level adjustment devices 508 and 510, respectively. The devices 508 and 510 are typically potentiometers or similar variable impedance devices. The adjustment of the devices 508 and 510 is typically performed manually, controlling the reference levels of the sum and difference signals in the output signal. This allows the user to customize the level and aspect of stereo enhancement according to the type of sound being reproduced and according to the user's personal preferences. Increasing the reference level of the sum signal emphasizes the audio information of the center sound field disposed between the pair of speakers. In contrast, increasing the reference level of the difference signal emphasizes the ambient sound information that creates a wider sound image perception. In certain audio settings where the music type and system configuration parameters are known or where manual adjustment is not possible, the adjustment means 508 and 510 may obviate the need to predetermine and fix the sum and difference signal levels.

The output of the means 510 is fed to a stereo enhancement equalizer 520 at an input 522. Equalizer 520 spectrally shapes the difference signal on input 522 as shown in fig. 7.

The shaped difference signal is provided to a mixer 542 which also receives the sum signal from the means 508. In one embodiment, the stereo signals 594 and 596 are also provided to a mixer 542. All of these signals are combined in mixer 542 to produce enhanced and spatially corrected left and right output signals 530 and 532.

Although the input signals 126 and 128 typically represent corrected stereo signals, they may also be generated synthetically from a mono sound source.

Acoustic image correction characteristic

Fig. 6A-6C are graphical representations of the spatial correction levels provided by the "low" and "high" frequency correction systems 580, 582, 584, 586 to obtain a repositioned sound image generated from a pair of stereo signals.

Referring initially to fig. 6A, the possible spatial correction levels provided by correction systems 580 and 582 are illustrated by curves having different amplitude versus frequency characteristics. The maximum level of correction or boost (boost), measured in dB, provided by systems 580 and 582 is represented by correction curve 650. Curve 650 gives the increased level of boost in a first frequency range of about 100Hz and 1000 Hz. At frequencies above 1000Hz, the boost level is maintained at a suitably constant level. Curve 652 represents a correction level close to zero.

To those skilled in the art, a typical filter is generally characterized by a pass band and a stop band of frequencies separated by a cutoff frequency. The calibration curves of fig. 6A-6C, while representing a typical signal filter, may be characterized in terms of pass bands, stop bands, and transition bands. A filter designed according to the characteristics of fig. 6A has a pass band above about 1000Hz, a transition band between about 100 and 1000Hz, and a stop band below about 100 Hz. The filter according to fig. 6B has a pass band above about 10kHz, a transition band between about 1kHz and 10kHz, and a stop band below about 1 kHz. The filter according to fig. 6C has a stop band above about 10kHz, a transition band between about 1kHz and 10kHz, and a pass band below about 1 kHz. In one embodiment, the filter is a first order filter.

As can be seen from fig. 6A-6C, the spatial correction of the audio signal accomplished by the systems 580, 582, 584, and 586 is substantially uniform within the pass band, but largely frequency dependent within the transition band. The amount of acoustic correction applied to the audio signal by adjustment of the stereo image correction system 622, which changes the filter slope of fig. 6A-6C, can be varied as a function of frequency. Thus, the frequency dependent correction is applied to a first frequency range between 100 and 1000 hertz and to a second frequency range of 1000 to 10,000 hertz. By independent adjustment of correction systems 580, 662, 584, and 586, an infinite number of correction curves are possible.

According to one embodiment, the spatial correction of the high frequency stereo signal components occurs between about 1000Hz and 10000 Hz. The energy correction of these signal components may be positive, i.e. boost, as shown in fig. 6B, or negative, i.e. decay, as shown in fig. 6C. The range of lift provided by correction systems 584 and 586 is characterized by a maximum lift curve 660 and a minimum lift curve 662. The curves 664, 666 and 668 indicate some additional boost levels that may be required for spatial correction of sound emitted from different sound reproduction systems. Fig. 6C illustrates an energy correction curve that is substantially the inverse of fig. 6B.

Because the lower and higher frequency correction factors represented by the curves of fig. 6A-6C are added together, there are a wide variety of possible spatial correction curves that may be applied between frequencies of 100 to 10000 Hz. Fig. 6D is a graphical representation illustrating the range of the composite spatial correction characteristic provided by the stereo image correction system 122. Specifically, the solid curve 680 represents the maximum level of spatial correction including the curve 650 (shown in FIG. 6A) and the curve 660 (shown in FIG. 6B). The correction for lower frequencies may vary according to the solid curve 680 for the range specified by θ 1. Likewise, the correction of high frequencies may be changed according to the solid curve 680 of the range specified by θ 2. Thus, the amount applied to the first frequency range of 100 to 1000 hertz varies between about 0 and 15dB, while the correction applied to the second frequency range of 1000 to 10000 hertz may vary between about 13dB to-15 dB.

Sound image enhancement feature

Turning now to the stereo image enhancement aspect of the present invention, a series of perspective enhancement or normalization curves are graphically represented in FIG. 7. Upper signal (L)_c-R_c) Representing the processed difference signal that has been spectrally shaped according to the frequency response characteristic of fig. 7. These frequency response characteristics are exploited by the equalizer 520 shown in fig. 5 and are based in part on the HRTF principle.

In general, selective amplification of the difference signal enhances any ambient or reverberant sound effects that may be present in the difference signal, but which are masked by the sound of the stronger direct sound field. To a suitable extent, these ambient sounds are readily perceived in the sound field of the scene. However, in the recorded performance, the ambient sound is attenuated relative to the live performance. By raising the level of the difference signal derived from a pair of left and right stereo signals, the projected sound image can be significantly broadened when the sound image is emitted from a pair of speakers provided in front of the listener.

The perspective curves 790, 792, 794, 796 and 798 of fig. 7 are shown as a function of gain-audio frequency shown in logarithmic format. Different levels of equalization between the curves of fig. 7 are required in view of various audio reproduction systems. In one embodiment, the level of different signal equalizations is a function of the actual placement of the speakers within the audio reproduction system relative to the listener. Curves 790, 792, 794, 796 and 798 generally show a frequency profiling characteristic in which the lower and higher difference signal frequencies are elevated relative to the mid-band of frequencies.

According to one embodiment, the range of the perspective curve of FIG. 7 is defined by a maximum gain of about 10-15dB at about 125 to 150 Hz. The maximum gain value represents the point of inflection of the curve of fig. 7 from which the slope of curves 790, 792, 794, 796, and 798 changes from a positive value to a negative value. In fig. 7, such turning points are indicated by points A, B, C, D and E. The gain of the perspective curve decreases below 125Hz at a rate of about 6dB per octave. Above 125Hz, the gain of the curve of fig. 7 also drops, but at a variable rate towards a minimum gain turning point of about-2 to +10 dB. The minimum gain turning points are significantly different between curves 790, 792, 794, 796 and 798. The minimum gain turning points are labeled as points A ', B ', C ', D ' and E ', respectively. The frequency at which these minimum gain breakpoints occur varies from about 2.1kHz for curve 790 to about 5kHz for curve 798. The gains of curves 790, 792, 794, 796 and 798 increase from their respective minimum gain frequencies up to about 10 KHz. Above 10KHz, the gain used by the perspective curve tends to flatten. However, up to about 20KHz, i.e., about the highest frequency audible to the human ear, the gain of all curves will continue to increase.

The above gain and frequency figures are for design purposes only, and the actual figures will vary from system to system. In addition, adjustment of the signal level devices 508 and 510 may affect the maximum and minimum gain values, as well as the gain interval between the maximum gain frequency and the minimum gain frequency.

The purpose of the equalization of the difference signal according to the curve of fig. 7 is to boost the difference signal components of statistically lower intensity without over-enhancing the difference signal components of higher intensity. The higher intensity difference signal component of a typical stereo signal is seen in the mid-range frequencies between about 1 to 4 kHz. The human ear is more sensitive to these same mid-range frequencies. Thus, the enhanced left and right output signals 530 and 532 produce much improved sound effects because ambient sounds are selectively emphasized to substantially surround the listener within the reproduced sound field.

As shown in fig. 7, the difference signal frequency below 125Hz results in a reduced boost through the use of a perspective curve (if any). This reduction is aimed at avoiding that very low frequencies, i.e. bass, are amplified too much. In many audio reproduction systems, amplifying the audio difference signal in such a low frequency range may produce an unpleasant and distorted sound image with too much bass response. Examples of such audio reproduction systems include near-field or low-power audio systems, such as multimedia computer systems and home stereo systems. The large power draw (draw) in these systems may cause the amplifier to "clip" during high boost, or may damage components of the audio system, including the speakers. Limiting the bass response of the difference signal also helps to avoid these problems in most near-field audio enhancement applications.

According to one embodiment, the level of the difference signal equalization in an audio environment with a fixed listener depends on the type of actual loudspeakers and their position relative to the listener. The acoustic principle based on this decision can be better explained in conjunction with fig. 8A and 8B. Fig. 8A and 8B are intended to illustrate this acoustic principle with respect to the change of orientation of the loudspeaker system.

Fig. 8A illustrates a top view of a sound reproduction environment having speakers 800 and 802 placed slightly forward of a listener 804 and directed to both sides of the listener. Speakers 800 and 802 are also placed below listener 804, at a similar height position to speaker 246 shown in fig. 2. The reference planes a and B are aligned with the ears 806 and 808 of the listener 804. As shown, planes A and B are parallel to the listener's line of sight.

The location of the speakers preferably corresponds to the location of speakers 810 and 812. In one embodiment, when the loudspeakers cannot be placed in the desired position, an enhancement of the apparent sound image can be achieved by selectively equalizing the difference signal, i.e. the gain of the difference signal will vary with frequency. Curve 790 of fig. 7 represents the ideal level of difference signal equalization for the actual speaker position corresponding to phantom speakers 810 and 812.

Bass enhancement

The invention also provides a method and system for enhancing an audio signal. The sound enhancement system improves the realism of sound with a unique sound enhancement process. In general, a sound enhancement process receives two input signals, a left input signal and a right input signal, and then produces two enhanced output signals, a left output signal and a right output signal.

The left and right input signals are processed together to provide a pair of left and right output signals. In particular, the enhancement system embodiments equalize the differences that exist between two input signals in a manner that broadens and enhances the perceived bandwidth of the sound. In addition, many embodiments adjust the level of sound that is common to both input signals in order to reduce clipping. Advantageously, some embodiments achieve sound enhancement through a simplified, low cost, and easy to manufacture analog system that does not require digital signal processing.

Although the embodiments herein are described with reference to a sound enhancement system, the invention is not limited thereto but may be used in various other situations where it is desirable to adapt different embodiments of a sound enhancement system to different situations.

A typical small speaker system for a multimedia computer, automobile, small stereo, portable stereo, headphone, etc. will have an acoustic output response that rolls off at about 150 Hz. Fig. 9 shows a curve 906 that approximately corresponds to the frequency response of a human ear. Fig. 9 also shows a measured response 908 for a typical small computer speaker system that employs a high frequency driver (tweeter) to reproduce high frequency, a four inch mid-range-bass driver (woofer) to reproduce mid-range and bass frequencies. Such a system employing two drives is commonly referred to as a two-way system. Speaker systems employing more than two drivers are also common in the art and may be used with embodiments of the present invention. Single driver speaker systems are also common and may be used with the present invention. The response characteristic 908 is plotted on rectangular coordinates, with the X-axis representing frequencies from 20Hz to 20 kHz. This band corresponds to the range of normal human hearing. The Y-axis of fig. 9 represents the normalized amplitude response from 0dB to-50 dB. Curve 908 is relatively smooth in the mid-range band of approximately 2kHz to 10kHz, while exhibiting some roll-off above 10 kHz. In the low frequency range, curve 908 shows a low frequency roll-off starting in the mid-bass frequency band between about 150Hz and 2kHz, up to below 150Hz, with the speaker system producing very little sound output.

The location of the frequency bands shown in fig. 9 is by way of example only and not by way of limitation. The actual frequency ranges of the ultra-bass, mid-bass and mid-range bands will vary depending on the speaker and the application in which the speaker is used. The term subwoofer is generally used to refer to frequencies within a frequency band where the speaker produces less accurate output than the speaker output at higher frequencies, such as in the mid-bass frequency band. The term mid-bass frequency band is generally used to refer to frequencies above the sub-bass frequency band. The term mid-range is generally used to refer to frequencies above the mid-bass band.

Many cone type drivers are inefficient at generating acoustic energy at low frequencies, where the cone diameter is smaller than the wavelength of the acoustic sound wave. When the cone diameter is smaller than the wavelength, to keep the sound pressure level of the sound output from the cone uniform, the cone swing needs to be increased by a factor of 4 (a factor of 2) for each octave of frequency reduction. If an attempt is made to improve the low frequency response by simply increasing the electrical power supplied to the driver, the maximum allowable swing of the driver cone is quickly reached.

Thus, the low frequency output of the driver cannot be increased beyond a certain limit, which indicates that the low and fair sound quality of most small loudspeaker systems is not good. Curve 908 is a typical characteristic of most small speaker systems that employ low frequency drivers of about 4 inches in diameter. Speaker systems with larger drivers tend to produce perceptible sound outputs down to frequencies lower than those shown in curve 908, while systems with smaller drivers do not typically produce as low an output as shown in curve 908.

As mentioned above, system designers have so far had little choice in designing speaker systems with extended low frequency response. Previously known solutions are expensive and elongated speakers that are too large for the table top. One popular solution to the low frequency problem is to use a subwoofer, typically placed on the floor near the computer system. Subwoofers can provide adequate low frequency output, but they are expensive and therefore less popular than inexpensive table top speakers.

Embodiments of the present invention do not employ a large direct cone driver or subwoofer, but rather utilize the features of the human auditory system to produce the perception of low frequency acoustic energy (even if such energy is not produced by the speaker system), thereby overcoming the low frequency limitations of small systems.

It is well known that the human auditory system is non-linear. In brief, a non-linear system means a system in which an increase in input does not follow a proportional increase in output. Thus, for example, in the ear, a doubled sound pressure level does not produce the sensation of doubling the volume of the sound source. In fact, the human ear is quite similar to a square law device, which responds not to the intensity of the acoustic energy, but to power. The non-linearity of this hearing mechanism produces intermodulation frequencies that sound like harmonics or overtones of the actual frequency of the sound wave.

The nonlinear intermodulation effect of the human ear is shown in fig. 10, which illustrates the idealized amplitude spectra of two pure tones. The spectral diagram in fig. 10 shows a first spectral line 1004 corresponding to acoustic energy produced by a speaker driver (e.g., a subwoofer) at 50 Hz. The second spectral line 1002 is shown at 60 Hz. Lines 1004 and 1002 are the actual spectral lines corresponding to the true acoustic energy produced by the driver, assuming no other acoustic energy is present. However, due to the inherent non-linearity of the human ear, it produces intermodulation products corresponding to the sum of the two actual spectral frequencies and the difference between the two spectral frequencies.

For example, a person listening to the acoustic energy represented by lines 1004 and 1002 may experience acoustic energy at 50Hz, as shown by line 1008, acoustic energy at 60Hz, as shown by line 1008, and acoustic energy at 110Hz, as shown by line 1010. The spectral line 1010 does not correspond to the true acoustic energy produced by the speaker, but rather corresponds to a spectral line in the human ear that is produced by the non-linearity of the human ear. Line 1010 occurs at a frequency of 110Hz, which is the sum of two actual lines (110 Hz-50Hz +60 Hz). Note that the non-linearity of the human ear also produces a line at another frequency of 10Hz (10 Hz-60 Hz-50Hz), but this line is not perceived because it is below the human auditory range.

Fig. 10 illustrates intermodulation processing in the human ear, but it is somewhat simplified over real program material, such as music. Typical program material, such as music, is rich in harmonics, so most music exhibits an almost continuous spectrum, as shown in fig. 11. Fig. 11 shows a comparison between actual and perceived acoustic energy of the same type as that shown in fig. 10, except that the graph in fig. 11 shows a continuous spectrum. Fig. 11 shows an actual acoustic energy curve 1120 and a corresponding perceived frequency spectrum 1130.

For most non-linear systems, the non-linearity of the human ear is more pronounced when the system produces a large swing (e.g., large signal level) rather than a small swing. Thus, the non-linearity is more pronounced at low frequencies for the human ear, where the eardrum and other parts of the human ear produce relatively large mechanical swings, even at lower volume levels. Thus, FIG. 11 shows that the difference between the actual acoustic energy 1120 and the perceived acoustic energy 1130 tends to be greatest in the lower frequency range and relatively small in the higher frequency range.

As shown in fig. 10 and 11, low frequency acoustic energy containing multiple tones or frequencies can create a perception in a listener that acoustic energy in the mid-bass range contains more spectral content than actually exists. The human brain, in the face of thought of missing information, will subconsciously attempt to "fill in" the missing information. This padding phenomenon is the basis for many illusions. In embodiments of the invention, the brain may be tricked into filling in low frequency information that is not really present by providing the brain with a mid-bass effect of such low frequency information.

In other words, if the brain gets harmonics to be generated by the ear, if low frequency acoustic energy is present (e.g., spectral line 1010), then the brain subconsciously fills in the low frequency spectral lines 1006 and 1008 that it deems "necessary" to be present, as appropriate. This padding process is amplified by another non-linear effect of the human ear, the so-called detector effect.

The non-linearity of the human ear also causes the ear to act as a detector, similar to a diode detector in an Amplitude Modulation (AM) receiver. If the mid-bass harmonic is an amplitude modulated wave modulated by a subwoofer, the ear will demodulate the modulated mid-bass carrier to reproduce the subwoofer envelope. Fig. 12A and 12B graphically illustrate modulated and demodulated signals. Fig. 12A shows a modulated signal including a carrier signal of a higher frequency (for example, a mid-bass carrier) modulated by a subwoofer signal on a time axis.

The amplitude of the higher frequency signal is modulated by the lower frequency sounds, whereby the amplitude of the higher frequency signal varies with the frequency of the lower frequency sounds. The non-linearity of the ear will partially demodulate the signal so that the ear detects the low frequency envelope of the higher frequency signal, producing the perception of low frequency sounds even though no actual acoustic energy is produced at the lower frequencies. For the intermodulation effects described above, the detector effect can be enhanced by appropriate signal processing of signals in the mid-bass frequency range. With appropriate signal processing, a sound enhancement system can be designed that produces the perception of low frequency sound energy, even if the system uses speakers that cannot or are inefficient in producing such energy.

The perception of the actual frequencies present in the acoustic energy produced by the loudspeaker may be considered a first order effect. The perception of additional harmonics not present in the actual audio, whether such harmonics are produced by intermodulation distortion or by detection, can be considered second order effects.

Bass enhancement expander

Fig. 13A is a block diagram of a sound system in which a bass enhancement unit 1304 provides sound enhancement functionality. The bass enhancement unit 1304 receives an audio signal from the signal source 1302. The signal source 1302 may be any signal source including the signal processing block 122 shown in fig. 1. The bass enhancement unit 1304 performs signal processing, modifying the received audio signal, thereby producing an audio output signal. The audio output signal may be provided to a speaker, amplifier, or other signal processing device.

FIG. 13B is a block diagram of the topology of a binaural bass enhancement unit 1304 having a first input 1309, a second input 1311, a first output 1317, and a second output 1319. The first input 1309 and the first output 1317 correspond to a first channel. The second input terminal 1311 and the second output terminal 1319 correspond to a second channel. The first input 1309 is connected to a first input of the combiner 1310 and to an input of the signal processing block 1313. The output of the signal processing block 1313 is provided to a first input of a combiner 1314. A second input 1311 is connected to a second input of the combiner 1310 and to an input of the signal processing block 1315. The output of the signal processing block 1315 is provided to a first input of a combiner 1316. The output of the combiner 1310 is provided to an input of a signal processing block 1312. An output of the signal processing block 1312 is provided to a second input of the combiner 1314 and a second input of the combiner 1316. The output of the combiner 1314 is provided to a first output 1317. The output of the second combiner 1316 is provided to a second output 1319.

The signals from the first and second inputs 1309 and 1311 are combined and processed by a signal processing block 1312. The output of signal processing block 1312 is a signal that, when combined with the outputs of signal processing blocks 1313 and 1315, respectively, produces bass enhancement outputs 1317 and 1319.

Fig. 13C is a block diagram of another topology of the binaural bass enhancement unit 1344. In fig. 13C, a first input 1309 is provided to an input of a signal processing block 1321 and to an input of a signal processing block 1322. The output of signal processing block 1321 is provided to a first input of a combiner 1325 and the output of signal processing block 1322 is provided to a second input of the combiner 1325. A second input 1311 is provided to an input of signal processing block 1323 and to an input of signal processing block 1324. The output of the signal processing block 1323 is provided to a first input of a combiner 1326 and the output of the signal processing block 1324 is provided to a second input of the combiner 1326. The output of the combiner 1325 is provided to a first output 1317 and the output of the second combiner 1326 is provided to a second output 1319.

Unlike the topology shown in fig. 13B, the topology shown in fig. 13C does not combine the two input signals 1309 and 1311, but maintains the separation of the two channels and performs bass enhancement processing on each channel.

FIG. 14 is a block diagram 1400 of one embodiment of the bass enhancement system 1304 shown in FIG. 13A. The bass enhancement system 1400 employs a bass pass-through unit 1420 to generate a time-dependent enhancement factor. Fig. 14 may also be used as a flow chart describing a program running on a DSP or other processor that implements the signal processing operations of an embodiment of the present invention. Fig. 14 shows two inputs, a left channel input 1402 and a right channel input 1404. As with the previous embodiments, left and right are used for convenience and are not limiting. Both inputs 1402 and 1404 are provided to an adder 1406, which produces a combined output of the two inputs.

The output of the adder 1406 is provided to a first band pass filter 1412, a second band pass filter 1413, a third band pass filter 1415, and a fourth band pass filter 1411. The output of the bandpass filter 1413 is provided to an input of a summer 1418.

The output of the bandpass filter 1415 is provided to a first throw of a Single Pole Double Throw (SPDT) switch 1416. The output of the bandpass filter 1411 is provided to a second throw terminal of the SPDT switch 1416. The single pole of switch 1416 is connected to an input of adder 1418.

The output of the band pass filter 1412 is provided to an input of a summer 1418.

The output of adder 1418 is connected to the input of the bass pass-through unit 1420. The output of the bass pass-through 1420 is provided to a first throw of a (SPDT) switch 1422. The second throw terminal of SPDT switch 1422 is set to ground. One pole of SPDT switch 1422 is connected to a first input of left channel adder 1424 and a first input of right channel adder 1432. The left channel input 1402 is provided to a second input of the left channel adder 1424, and the right channel input 1404 is provided to a second input of the right channel adder 1432. The outputs of the left channel adder 1424 and the right channel adder 1432 are the left channel output 1430 and the right channel output 1433, respectively, of the signal processing block 1400. Switches 1422 and 1416 are optional and may be replaced by fixed connections.

The switch 1416 allows the filters 1411-1415 to be configured in two different frequency ranges, namely 40-150Hz and 100-200 Hz.

The filtering operations provided by the filters 1411 and 1413, 1415 and the combiner 1418 may be combined into a combined filter 1407 as shown in figure 14. For example, in another embodiment, the filters 1411, 1413, 1415 are combined into a single bandpass filter having a passband extending from about 40Hz to 250 Hz. For processing bass frequencies, the pass band of the combined filter 1407 preferably extends from about 20 to 100Hz at the low end and from about 150 to 350Hz at the high end. The combining filter 1407 may have other filter transfer functions including, for example, a high pass filter, a shelving (shelving) filter, etc. The combined filter may also be configured to operate in a similar manner to a graphic equalizer and attenuate some frequencies within its pass band relative to other frequencies within its pass band.

As shown, fig. 14 corresponds approximately to the topology shown in fig. 13B, where signal processing blocks 1313 and 1315 have a transfer function of one, and signal processing block 1312 includes combining filter 1407 and bass pass-through unit 1420. However, the signal processing shown in fig. 14 is not limited to the topology shown in fig. 13B. The components of fig. 14 may also be used in the topology shown in fig. 13C, where the transfer functions of signal processing blocks 1321 and 1323 are unity, and signal processing blocks 1322 and 1324 include a combining filter 1407 and a bass pass-through unit 1420. Although not shown in fig. 14, the signal processing blocks 1313, 1315, 1321, and 1323 may provide additional signal processing, such as high pass filtering to remove bass frequencies, high pass filtering to remove frequencies processed by the bass pass-through unit 1420, high frequency emphasis to enhance high frequency sound, additional bass pass-through processing to supplement the bass pass-through system, and so on. Other combinations are also contemplated.

FIG. 15 is a frequency domain plot showing the general shape of the transfer functions of the bandpass filters 1411-1413, 1415. FIG. 15 shows bandpass transfer functions 1501-1504, which correspond to bandpass filters 1411-1413, 1415, respectively. The transfer functions 1501-1504 are represented as bandpass functions centered at 40, 100, 150, and 200Hz, respectively.

In one embodiment, the band pass filter 1411 is tuned to frequencies below 100Hz, such as 40 Hz. When switch 1416 is in a first position corresponding to a first throw, it selects bandpass filter 1411 and not bandpass filter 1415, thereby providing bandpass filters at 40, 100 and 150 Hz. When switch 1416 is in a second position corresponding to the second throw, it selects bandpass filter 1415 instead of bandpass filter 1411, thereby providing bandpass filters at 100, 150 and 200 Hz.

Thus, switch 1416 preferably allows the user to select the frequency range to be enhanced. Users of speaker systems having woofers providing small woofers, such as approximately 3 to 4 inch diameter woofers, typically select the higher frequency ranges provided by bandpass filters 1412-1413, 1415 tuned to 40, 60, 100, and 150Hz, respectively. Users of speaker systems having woofers arranged slightly larger, e.g., approximately 5 inch diameter, typically select the lower frequency ranges provided by bandpass filters 1411 and 1413, 1515 tuned to 40, 60, 100, and 150Hz, respectively. Those skilled in the art will recognize that more switches may be provided to enable selection of more band pass filters and more frequency ranges. Selecting different bandpass filters to provide different frequency ranges is a desirable technique because bandpass filters are inexpensive and different bandpass filters can be selected with a single throw switch.

In one embodiment, the bass pass-through unit 1420 employs Automatic Gain Control (AGC), which includes a linear amplifier with an internal servo feedback loop. The servo automatically adjusts the average amplitude of the output signal to match the average amplitude of the control input signal. The average amplitude of the control input is typically obtained by detecting the envelope of the control signal. The control signal may also be obtained by other methods including, for example, low pass filtering, band pass filtering, peak detection, RMS (root mean square) averaging, average averaging, and the like.

The servo loop increases the forward gain of the bass pass-through block 1420 in response to an increase in the amplitude of the envelope of the signal provided to the input of the bass pass-through block 1420. Conversely, in response to a decrease in the amplitude of the envelope of the signal provided to the input of the bass pass-through unit 1420, the servo loop will decrease the forward gain of the bass pass-through unit 1420. In one embodiment, the gain of the bass pass-through cell 1420 increases faster than the gain decreases. Fig. 16 is a time domain plot illustrating the gain of the bass punch-through cell 1420 in response to a unit step input. Those skilled in the art will appreciate that fig. 16 is a graph of gain as a function of time, rather than an output signal as a function of time. Most amplifiers have a fixed gain, so the gain is rarely plotted. However, Automatic Gain Control (AGC) in the bass pass-through unit 1420 changes the gain of the bass pass-through unit 1420 in response to the envelope of the input signal.

The unit step input is plotted as curve 1609 and the gain is plotted as curve 1602. In response to the leading edge of the input pulse 1609, the gain rises during a period 1604 corresponding to an increasing time constant. At the end of period 1604, gain 1602 reaches a steady state gain a 0. In response to the trailing edge of the input pulse 1609, the gain drops back to zero for a period of time corresponding to the decay time constant 1606.

The attack time constant 1604 and decay time constant 1606 are preferably chosen to provide enhancement of bass frequencies without overdriving other components of the system, such as the amplifier and speaker. FIG. 17 is a time domain plot 1700 of typical low notes played by a musical instrument (bass guitar, bass drum, electronic synthesizer, etc.). Curve 1700 shows a higher frequency portion 1744 whose amplitude is modulated by a lower frequency portion having a modulation envelope 1742. The envelope 1742 has a heightened portion 1746, followed by an attenuated portion 1747, followed by a sustained portion 1748, followed by a released portion 1749. The maximum amplitude of curve 1700 is at peak 1750, occurring at the point in time between the rising portion 1746 and the decaying portion 1747.

As noted, waveform 1744 is typical of many, if not most, instruments. For example, when pulling and releasing the string of a guitar, several large amplitude vibrations are initially generated, then the vibrations stabilize in a more or less stable state and slowly decay over a long period of time. The initial large amplitude vibration of the guitar string corresponds to the heightened portion 1746 and the damped portion 1747. The slowly decaying vibrations correspond to the continuing portion 1748 and the releasing portion 1749. The piano strings also act in a similar manner when struck by the hammer body which strikes the piano keys.

The piano string may have a more pronounced transition from the sustained portion 1748 to the released portion 1749 because the hammer body does not return to rest on the string until the piano key is released. When the key is depressed, the string is free to vibrate with relatively little attenuation for the duration 1748. During the release period 1749, when the piano keys are released, the felt covered hammer body rests on the piano keys and quickly dampens the string vibrations.

Likewise, when the tympanic membrane is impacted, a set of initially large amplitude vibrations is generated, corresponding to the elevated portion 1746 and the attenuated portion 1747. After the large amplitude vibration fades away (corresponding to the end of the attenuating portion 1747), the tympanic membrane continues to vibrate for a period of time corresponding to the duration portion 1748 and the release portion 1749. Many instrument sounds can be created by simply controlling the length of the time period 1746-1749.

As described in connection with fig. 12A, the amplitude of the higher frequency signal is modulated by the lower frequency tones (envelopes), whereby the amplitude of the higher frequency signal varies with the frequency of the lower frequency tones. The ear non-linearity will partially demodulate the signal so the ear will detect the low frequency envelope of the higher frequency signal, creating the perception of low frequency sounds even though no actual acoustic energy is generated at low frequencies. This detector effect can be enhanced by appropriate signal processing of the signals in the mid-bass frequency range, typically 50-150Hz at the low end of the range and 200-500Hz at the high end of the range. With appropriate signal processing, a sound enhancement system can be designed that produces the perception of low frequency sound energy even when speakers that do not produce such energy are used.

The perception of the actual frequencies present in the acoustic energy produced by the loudspeaker may be considered a first order effect. The perception of additional harmonics not present in the actual audio, whether such harmonics are generated by intermodulation distortion or by detection, can be considered second order effects.

However, if the amplitude of the peak 1750 is too high, the speaker (and possibly the power amplifier) will be overdriven. An overdriven speaker will cause considerable distortion and may damage the speaker.

The bass feedthrough unit 1420 preferably provides enhanced bass in mid-bass regions while reducing the overdrive effect of the peak 1750. The increased time constant 1604 provided by the bass pass-through cell 1420 limits the rise time of the gain through the bass pass-through cell 1420. The attack time constant of the bass pass-through unit 1420 has a relatively small effect on the wavelength of the long attack period 1746 (slow envelope rise time) and a relatively large effect on the wavelength of the short attack period 1746 (fast envelope rise time).

Bass feedthrough with peak compression

The heightened portion of a note played by a bass instrument (e.g., a bass guitar) typically begins with an initial pulse of relatively high amplitude. In some cases, such peaks may overdrive the amplifier or speaker, resulting in distorted sound, and possibly damaging the speaker or amplifier. The bass enhancement processor provides flattening of peaks in the bass signal while increasing the energy of the bass signal, thereby enhancing the overall bass perception.

The energy in the signal is a function of the amplitude of the signal and the duration of the signal. In other words, the energy is proportional to the area under the signal envelope. Although the initial pulse of bass reflex may have a relatively large amplitude, the pulse tends to contain less energy because of its short duration. Thus, the initial pulse of small energy often does not produce a noticeable bass sensation. Thus, the amplitude of the initial pulse can generally be reduced without significantly affecting the bass perception.

Fig. 18 is a signal processing block diagram of a bass enhancement system 1800 that provides bass enhancement using a peak compressor to control pulse amplitude, e.g., initial pulse, bass. In system 1800, peak compressor 1802 is interposed between combiner 1718 and pass-through unit 1720. The output of combiner 1718 is provided to the input of peak compressor 1802, and the output of peak compressor 1802 is provided to the input of bass pass-through unit 1720.

The above description relating to fig. 14 to 13B and 13C also applies to the topology shown in fig. 18. For example, as shown, fig. 18 approximately corresponds to the topology shown in fig. 13B, where signal processing blocks 1313 and 1315 have a transfer function of one, and signal processing block 1312 includes a combining filter 1707, a peak compressor 1802, and a bass pass-through unit 1720. However, the signal processing shown in fig. 18 is not limited to the topology shown in fig. 13B. The components of fig. 18 may also be used in the topology shown in fig. 13C. Although not shown in fig. 18, the signal processing blocks 1313, 1315, 1321, and 1323 may provide additional signal processing, such as high pass filtering to remove bass frequencies, high pass filtering to remove frequencies processed by the bass feedthrough 1720 and compressor 1802, high frequency emphasis to enhance high frequency sound, additional bass processing to supplement the bass feedthrough system 1720 and peak compressor 1802, and so forth. Other combinations are also contemplated.

Peak compressor 1802 "flattens" the envelope of the signal provided at its input. For large amplitude input signals, the apparent gain of the compression device 1802 is reduced. For small amplitude input signals, the apparent gain of the compression device 1802 is raised. Thus, the compression apparatus reduces the peaks of the input signal envelope (fills the recesses in the input signal envelope). Whatever signal is provided at the input of the compressor device 1802, the envelope (e.g., average amplitude) of the signal output from the compressor device 1802 has a relatively uniform amplitude.

Fig. 19 is a time domain graph illustrating the effect of a peak compressor on the envelope of an initial pulse having a relatively high amplitude. Fig. 19 shows a time domain plot of an input envelope 1914 with an initial large amplitude pulse followed by a longer time lower amplitude signal. Output envelope 1916 represents the effect of bass pass-through unit 1720 on input envelope 1914 (the case where peak compressor 1802 is not used). The output envelope 1917 represents the effect of passing the input signal 1914 through the peak compressor 1802 and through-cell 1720.

As shown in fig. 19, the bass feedthrough does not limit the maximum amplitude of the input signal 1914, assuming that the amplitude of the input signal 1914 is sufficient to cause overdrive of the amplifier or speaker, so the output signal 1916 is also sufficient to overdrive the amplifier or speaker.

However, the pulse compression device 1802 used for the pin pair signal 1917 compresses the large amplitude pulse (reduces the amplitude of the large amplitude pulse). The compression device 1802 detects large amplitude amplitudes of the input signal 1914 and compresses (reduces) the maximum amplitude so that there is little likelihood that the output signal 1917 will overdrive the amplifier or speaker.

Because compression device 1802 reduces the maximum amplitude of the sign, it is possible to increase the gain provided by feedthrough cell 1420 without significantly reducing the likelihood that output signal 1917 overdrives the amplifier or speaker. Signal 1917 corresponds to an embodiment in which the gain of bass pass-through unit 1420 has been increased. Thus, during the longer decay portion, the signal 1917 has a larger amplitude than the curve 1916.

As described above, the energy in the signals 1914, 1916, and 1917 is proportional to the area under the curve representing each signal. The signal 1917 has more energy because even though it has a smaller maximum amplitude, the area under the curve representing the signal 1917 is larger than the signals 1914 or 1916. Because the signal 1917 contains more energy, the listener will perceive more bass in the signal 1917.

Thus, the use of a peak compressor in combination with the bass feedthrough 1420 allows the bass enhancement system to provide more energy in the bass signal while reducing the likelihood that the enhanced bass signal overdrives the amplifier or speaker.

Stereo image enhancement

The present invention also provides a method and system for improving the realism of sound, particularly in terms of the level of the sound field, using a unique differential perspective correction system. In general, a differential perspective correction device receives two input signals, a left input signal and a right input signal, and then generates two enhanced output signals, a left output signal and a right output signal as shown in fig. 5.

The left and right input signals are processed together to provide a pair of spatially corrected left and right output signals. In particular, one embodiment equalizes the difference that exists between two input signals in a manner that broadens and enhances the sound perceived by a listener. In addition, one embodiment also adjusts the level of sound that is common to both input signals in order to reduce clipping. Advantageously, one embodiment achieves sound enhancement using simplified, low cost, and easy to manufacture circuitry, without the need for separate circuitry to process the common and difference signals shown in FIG. 5.

Although the embodiments are described herein with reference to various sound enhancement systems, the present invention is not so limited and may be used in a variety of other applications where it is desirable to adapt different embodiments of sound enhancement systems to different situations. To facilitate a thorough understanding of the invention, the remainder of the detailed description is organized into the following sections and subsections:

fig. 20 is a block diagram of a differential perspective correction apparatus 2002 according to a first input signal 2010 and a second input signal 2012. In one embodiment, the first and second input signals 2010 and 2012 are stereo signals; the first and second input signals 2010 and 2012 need not be stereo signals but may include various audio signals. As described in detail below, the differential perspective correction device 2002 modifies the audio acoustic information common to the first and second input signals 2010 and 2012 in a manner that is different from the audio acoustic information not common to the first and second input signals 2010 and 2012.

The audio information common to the first and second input signals 2010 and 2012 is referred to as common mode information or a common mode signal (not shown). In one embodiment, the common mode information is not present in the form of discrete signals. Thus, the term common mode signal is used generically for this detailed description and conceptually refers to audio information that is present in both the first and second input signals 2010 and 2012 at any instant in time. For example, if a 1 volt signal is applied to the first and second input signals 2010 and 2012 simultaneously, the common mode signal includes 1 volt.

The adjustment of the common mode signal is conceptually illustrated in the common mode characteristic block 2020. Common mode characteristic block 2020 represents a change in a common mode signal. One embodiment reduces the amplitude of the frequency of the common mode signal in order to reduce clipping that may result from high amplitude input signals.

In contrast, audio information that is not common to the first and second input signals 2010 and 2012 is referred to as differential information or differential signal (not shown). In one embodiment, the differential signal is not a discrete signal, but in this detailed description, the differential signal refers to audio information representing the difference between the first and second input information 2010 and 2012. For example, if the first input signal 2010 is 0 volts and the second input signal 2012 is 2 volts, then the differential signal is 2 volts (the difference between the two input signals 2010 and 2012).

The modification of the differential signal is conceptually illustrated in the differential mode characteristic block 2022. As described in detail below, the differential perspective correction device 2002 equalizes selected frequency bands in the differential signal. That is, one embodiment equalizes audio information in a differential signal differently than audio information in a common-mode signal.

The differential perspective correction device 2002 spectrally shapes the differential signal in the differential module 2022 using various filters to generate an equalized differential signal. The differential perspective correction apparatus 2002 widens the perceived sound image projected from a pair of speakers provided in front of the listener by equalizing the selected frequency band within the differential signal.

Also, while the common mode characteristic block 2020 and the differential mode characteristic block 2022 are conceptually represented as separate blocks, one embodiment utilizes a single specially adapted system to accomplish these functions. Thus, one embodiment processes both common mode and differential audio information. Advantageously, one embodiment does not require complex circuitry to separate the audio input signal into a common mode signal and a differential signal. In addition, an embodiment generates a set of enhanced output signals without the need for a mixer to recombine the processed common mode signals and the processed differential signals.

The differential perspective correction device 2002 is then connected to one or more output buffers 2006. The output buffer 2006 outputs an enhanced first output signal 2030 and a second output signal 2032. As discussed in detail below, the output buffer 2006 isolates the differential perspective correction device 2002 from other elements connected to the first and second output signals 2030 and 2032. For example, the first and second output signals 2030 and 2032 may be accessed by other audio devices, such as a recording device, a power amplifier, a pair of speakers, etc., without changing the operation of the differential perspective correction device 2002.

FIG. 21 is a block diagram of a system for providing the differential perspective correction shown in FIG. 20 using a differential amplifier. In fig. 21, a first input 2010 is provided to the non-inverting input of the first differential amplifier 2102 and the first input of the crossover impedance block 2106. The second input 2012 is connected to a non-inverting input of the second differential amplifier 2104 and a second terminal of the crossover impedance block 2106. The inverting input of the first differential amplifier 2102 is connected to a first terminal of a crossover impedance block 2107 and a first terminal of a first feedback impedance 2108. The output of the first differential amplifier 2102 is provided to a first output 2030 and a second terminal of the first feedback impedance 2108. The inverting input of the second differential amplifier 2104 is connected to a second terminal of the crossover impedance block 2107 and a first terminal of a second feedback impedance 2109. The output of the second differential amplifier 2104 is provided to a second output 2032 and a second terminal of the second feedback impedance 2109.

The impedance of each block 2106, 2107, 2108 and 2109 is typically frequency dependent and may constitute a filter, for example, using resistors, capacitors and/or inductors. In one embodiment, impedances 2108 and 2106 are not frequency dependent.

Fig. 22 is an amplitude-frequency graph illustrating the common mode gain at the left and right output terminals 2030 and 2032. The common mode gain is represented by a first common mode gain curve 2200. As shown by common mode gain curve 2200, frequencies below about 130 hertz (Hz) are de-emphasized more than frequencies above about 130 Hz. For frequencies above about 130Hz, the frequency is uniformly reduced by about 6 decibels.

Figure 23 illustrates an overall correction curve 2300 resulting from the combination of the first and second frequency dividing networks 2106 and 2107. The approximate correlation gain values for the various frequencies within the overall correction curve 2300 may be measured with reference to zero (0) dB.

With this reference, the overall correction curve 2300 is defined by two turning points (labeled as point a and point B). At point a (about 125Hz in one embodiment), the slope of the correction curve changes from a positive value to a negative value. At point B (about 2kHz in one embodiment), the slope of the correction curve changes from negative to positive.

Thus, frequencies below about 125Hz are de-emphasized relative to frequencies near 125 Hz. Specifically, below 125Hz, the gain of the overall correction curve 2300 decreases at a rate of about 6dB per octave. De-emphasis of signal frequencies below 125Hz prevents very low frequencies (i.e., bass) from being over emphasized. In many audio reproduction systems, over-emphasizing the audio signal in this low frequency range relative to higher frequencies may produce an unpleasant and distorted sound image with a bass response that is too strong. Moreover, excessive emphasis on these frequencies can damage various audio components, including speakers.

Between points a and B, the slope of an overall correction curve is negative. That is, frequencies between about 125Hz and about 2kHz are de-emphasized relative to frequencies near 125 Hz. Thus, the gain related to the frequency between points A and B decreases at a variable rate to a maximum equalization point-8 dB at about 2 kHz.

Above 21.8kHz, the gain increases at a variable rate up to about 120kHz, about the highest frequency audible to the human ear. That is, frequencies above about 21.8kHz are emphasized over frequencies near 21.8 kHz. Therefore, the gain associated with frequencies above point B increases at a variable rate toward 120 kHz.

These relative gain and frequency values are for design purposes only, and the actual numbers may change as the system changes. Also, the gain and frequency values may vary depending on the type of sound or user preference without departing from the spirit of the present invention. For example, changing the number of crossover networks and changing the resistance and capacitance values within each crossover network allows the overall perspective correction curve 2300 to be tailored to the type of sound being reproduced.

Selective equalization of the differential signal enhances the ambient or reverberant sound effect present in the differential signal. As described above, frequencies in the differential signal are easily perceived at appropriate levels in the live sound field. Unfortunately, sound images do not provide the same 360 degree effect of live performance when recorded performances are played. However, equalizing the frequencies of the differential signals using the differential perspective correction device 2002 can significantly broaden the projected sound image, thereby reproducing the experience of live performance through a pair of speakers placed in front of the listener.

The equalization of the differential signal according to the overall correction curve 2300 is to de-emphasize signal components of statistically lower intensity relative to signal components of higher intensity. The higher intensity differential signal component of a typical audio signal occurs at mid-range frequencies between about 2 to 4 kHz. In this frequency range, the human ear has an increased sensitivity. Thus, the enhanced left and right output signals produce a more improved audio effect.

In other embodiments, the number of frequency dividing networks and the elements within the frequency dividing networks may be varied in order to simulate a so-called head transfer function (HRTF). The head transfer function describes different signal equalization techniques for adjusting the sound produced by a pair of speakers to account for the time it takes for the sound to be perceived by the left and right ears. Advantageously, HRTF-based transfer functions can be applied to the differential signals to localize the sound effect of the bleed-through, in order to create a localized sound field that is sufficiently bleed-through.

Examples of HRTF transfer functions that can be used to achieve a certain perceived azimuth angle are described in the article entitled "conversion of sound pressure level from free field to eardrum in horizontal plane" in e.a.b.shaw at volume 56, vol.6, 1974, and "conversion characteristics of human outer ear" by s.mehrgardt and v.mellert at volume 61, vol.6, 1977, in a l.12, vol.56, and in an article entitled "conversion characteristics of human outer ear" in e.g. 6, acout.soc.am, vol.s.mehrgart and v.mellert, as given intact.

Single chip implementation

Fig. 24 is a block diagram of one embodiment of a sound enhancement system 2400 that can be implemented on a single chip. As described in conjunction with fig. 1-23 above, the system 2400 includes a vertical sound image enhancement block 2402, a bass enhancement block 2404, and a horizontal sound image enhancement block 2406. External connections to system 2400 are provided through connection pins P1-P27. A positive power supply voltage is provided to pin P25, a negative power supply voltage is provided to pin P26, and ground is connected to pin P27. A first terminal of compressive coupling capacitance 2421 is provided to pin P10 and a second terminal of compressive coupling capacitance 2421 is provided to pin P11. A first terminal of compression delay capacitance 2420 is provided to pin P13 and a second terminal of compression delay capacitance 2420 is provided to pin P14. A first terminal of width control resistor 2430 is provided at pin P19, and a second terminal of width control resistor 2430 is provided at pin P20. A first terminal of width control resistor 2431 is provided at pin P21, and a second terminal of width control resistor 2431 is provided at pin P22. In one embodiment, width control resistors 2430 and 2431 are variable resistors.

Fig. 25A is a schematic diagram of the left channel of the vertical sound image enhancement block 2402. Fig. 25B is a schematic diagram of the right channel of the vertical sound image enhancement block 2402. In FIG. 25A, the left channel input is provided to pin P2 and the left channel bypass input is provided to pin P1. Pin P1 is provided to a first terminal of resistor 2501. A second terminal of the resistor 2501 is provided to a first terminal of the resistor 2502 and a first terminal of the capacitor 2503. Pin P2 is provided to a first terminal of resistor 2504 and a first terminal of capacitor 2505. A second terminal of the capacitor 2505 is provided to a first terminal of the resistor 2506 and a first terminal of the resistor 2507. A second terminal of the resistor 2506 is connected to ground.

A second terminal of the resistor 2502 is connected to a second terminal of the capacitor 2503, a second terminal of the resistor 2504, a second terminal of the resistor 2507, a first terminal of the resistor 2508, and an inverting input of an operational amplifier (opamp) 2510. The non-inverting input of opamp 2510 is grounded. A second terminal of the resistor 2508 is connected to a first terminal of the resistor 2509 and a first terminal of the capacitor 2612. A second terminal of the resistor 2509 is connected to a second terminal of the capacitor 2512, an output of the opamp 2510 and the left channel output 2511.

In one embodiment, resistor 2501 is 9.09 kohms, resistor 2502 is 27.4 kohms, capacitor 2503 is 0.1 μ f, resistor 2504 is 22.6 kohms, capacitor 2505 is 0.1 μ f, resistor 2506 is 3.01 kohms, resistor 2507 is 4.99 kohms, resistor 2508 is 9.09 kohms, resistor 2509 is 27.4 kohms, capacitor 2512 is 0.1 μ f and opamp 2510 is type TL074 or equivalent.

The right channel shown in fig. 25B is similar to the left channel shown in fig. 25A, with a bypass input from pin P3, a right channel input from P4, and a right channel output 2514.

Fig. 26 is a schematic diagram of bass boost block 2404. The left channel output 2511 from fig. 25A is connected to a first terminal of the resistor 2601 and a first terminal of the resistor 2611. The right channel output 2514 from fig. 25B is connected to a first terminal of the resistor 2602 and a first terminal of the resistor 2614.

A second terminal of the resistor 2601 is provided to a second terminal of the resistor 2602, a first terminal of the resistor 2625, and a first terminal of the capacitor 2603. The second terminal of the capacitor 2603 is connected to ground. A second terminal of the resistor 2625 is provided to an inverting input of the opamp 2606, a first terminal of the capacitor 2605, and a first terminal of the resistor 2604. The non-inverting input of opamp 2606 is connected to ground. The output of opamp 2606 is connected to the second terminal of resistor 2604, the second terminal of capacitor 2605, and the input of filter block 2607 (shown in detail in fig. 27). The first, second, and third outputs of filter block 2607 are connected to the inverting input of opamp 2608 and a first terminal of resistor 2609. The non-inverting input of opamp 2608 is connected to ground. The output of opamp 2608 is connected to a second terminal of resistor 2609 and pin P10.

Pin P19 is also coupled to the input of a compressor 2610 (shown in detail in fig. 28). The output of the compressor 2610 is connected to pin P12. Pin P12 is connected to pin P16. Pin P16 is connected to a first terminal of resistor 2612 and a first terminal of resistor 2613.

A second terminal of the resistor 2612 is coupled to the second terminal of the resistor 2611, an inverting input of the opamp 2620, and a first terminal of the resistor 2619. The non-inverting input of opamp 2620 is grounded. An output of the opamp 2620 is connected to a second terminal of the resistor 2619 and a first terminal of the resistor 2621. A second terminal of resistor 2621 is connected to pin P17. The output of opamp 2620 is also provided as left channel output 2630.

A second terminal of the resistor 2613 is connected to a second terminal of the resistor 2614, an inverting input of the opamp 2615, and a first terminal of the resistor 2617. The non-inverting input of opamp 2615 is connected to ground. The output of opamp 2615 is connected to a second terminal of resistor 2617 and a first terminal of resistor 2618. A second terminal of resistor 2618 is connected to pin P18. The output of opamp 2615 is also provided as right channel output 2631.

In one embodiment, resistors 2601, 2602, and 2604 are 43.2k ohms, capacitor 2603 is 0.022 μ f, resistor 2625 is 21.5k ohms, and capacitor 2605 is 0.01 μ f. In one embodiment, the resistor 2609 is 100k ohms, the resistors 2611, 2612, 2613, 2614, 2617, and 2619 are 10k ohms, and the resistors 2618 and 2621 are 200 ohms. In one embodiment, opamps 2606, 2608, 2615, and 2620 are of TL074 type or equivalent.

Fig. 27 is a schematic diagram of a filter system 2607. In fig. 27, an input is provided to a first terminal of resistors 2701 and 2704. A second terminal of resistor 2701 is connected to a first terminal of resistor 2710, a first terminal of capacitor 2721, and a first terminal of capacitor 2720. A second terminal of the capacitor 2721 is connected to a first terminal of the resistor 2722 and an inverting input of the opamp 2732. The non-inverting input of opamp 2732 is grounded. The output of opamp 2732 is connected to the second terminal of capacitor 2720, the second terminal of resistor 2722, and the first terminal of resistor 2723. A second terminal of resistor 2723 is connected to the first filter output.

A second terminal of resistor 2702 is connected to a first terminal of resistor 2712 and pin P5. A second terminal of resistor 2712 is connected to ground.

A second terminal of resistor 2703 is connected to a first terminal of resistor 2713 and pin P7. A second terminal of resistor 2713 is connected to ground.

Pin P6 is coupled to a first terminal of capacitor 2724 and a first terminal of capacitor 2728. A second terminal of capacitor 2728 is connected to a first terminal of resistor 2725, to a first terminal of resistor 2726, and to an inverting input of opamp 2729. The non-inverting input of opamp2729 is connected to ground. An output of opamp2729 is connected to a second terminal of capacitor 2724, a second terminal of resistor 2725, a second terminal of resistor 2726, and a first terminal of resistor 2730. A second terminal of the capacitor 2724 is connected to pin P8. A second terminal of resistor 2725 is connected to pin P9. A second terminal of the resistor 2730 is connected to the second filter output.

When pin P5 is shorted to pin P6 and pin P8 is disconnected from P9, the second filter output is a low frequency output (e.g., 40 Hz). When pin P7 is shorted to pin P6 and pin P8 is shorted to P9, the second filter output is a high frequency output (e.g., 150 Hz).

A second terminal of the resistor 2704 is connected to a first terminal of the resistor 2714, a first terminal of the capacitor 2731, and a first terminal of the capacitor 2735. A second terminal of capacitor 2735 is connected to a first terminal of resistor 2734 and an inverting input of opamp 2736. The non-inverting input of opamp 2736 is connected to ground. The output of opamp 2736 is connected to the second terminal of capacitor 2731, the second terminal of resistor 2734, and the first terminal of resistor 2737. A second terminal of resistor 2737 is connected to the third filter output.

In one embodiment, the first filter output is a bandpass filter centered at 100Hz, the third filter output is a bandpass filter centered at 60Hz, and the second filter output is a bandpass filter centered at 40Hz or 150Hz (as described above).

In one embodiment, the resistor 2701 is 31.6k Ω, the resistor 2702 is 56.2k Ω, the resistor 2703 is 21k Ω, the resistor 2704 is 37.4k Ω, the resistor 2710 is 4.53k Ω, the resistor 2712 is 13k Ω, the resistor 2713 is 3.09k Ω, the resistor 2714 is 8.87k Ω, the resistor 2722 is 63.4k Ω, the resistor 2723 is 100k Ω, the resistor 2725 is 57.6k Ω, the resistor 2726 is 158k Ω, the resistor 2730 is 100k Ω, the resistor 2734 is 107k Ω, and the resistor 2737 is 100k Ω. In one embodiment, capacitances 2720, 2721, 2724, 2728, 2731, and 2735 are 0.1 μ f. In one embodiment, opamps 2732, 2729, and 2736 are TL074 type or equivalent.

Fig. 28 is a schematic diagram of a compressor 2610. The compressor 2610 includes a peak detector 2804, a bias circuit 2802, a gain control block 2806, and an output buffer 2810. The peak detector is built around diode 2810 and diode 2811. The bias circuit is built around the transistor 2820 and the zener diode 2816. The gain control circuit is built around a FET (field effect transistor) 2814. The output buffer is built around opamp 2824.

The input to compressor 2610 is provided at pin P10. Pin P10 is connected to a first terminal of resistor 2827. A second terminal of resistor 2827 is connected to the drain of FET2814 and a first terminal of resistor 2822. A second terminal of resistor 2822 is connected to an inverting input of opamp 2824 and a first terminal of resistor 2823. The non-inverting input of opamp 2824 is grounded. The output of opamp 2824 is connected to the second terminal of resistor 2823 and pin P12. Pin P12 is the output of compressor 2616.

The source of FET2814 is connected to ground. A gate of FET2814 is connected to a first terminal of resistor 2813, a first terminal of resistor 2815, and pin P13. Pin P14 is connected to a second terminal of resistor 2815.

A second terminal of resistor 2813 is coupled to a cathode of diode 2811. The anode of diode 2811 is connected to the cathode of diode 2810 and pin P11. An anode of the diode 2810 is coupled to a first terminal of the resistor 2812. A second terminal of resistor 2812 is connected to pin P14.

Pin P14 is also connected to a first terminal of resistor 2818 and the emitter of PNP transistor 2820. A second terminal of resistor 2818 is connected to ground. A base of PNP transistor 2820 is connected to a first terminal of resistor 2817 and a first terminal of resistor 2819. A second terminal of resistor 2817 is connected to ground. The collector of PNP transistor 2820 is connected to the second terminal of resistor 2819, the anode of zener diode 2816, and pin P15. The cathode of the zener diode 2816 is grounded. Pin P15 is set so that the current limiting bias resistor can be connected between the zener diode and the negative supply voltage.

A capacitor 2421 connected between P10 and P11 ac couples the input to the peak detector circuit. A capacitor 2420 connected between P13 and P14 provides a delay time constant for compressor startup.

In one embodiment, diodes 2810 and 2811 are 1N4148 type or equivalent. In one embodiment, FET2814 is 2N3819 type or equivalent, PNP transistor 2820 is 2N2907 type or equivalent, and zener diode 2816 is a 3.3 volt zener diode (1N746A or equivalent). In one embodiment, opamp 2824 is TL074 type or equivalent. Capacitor 2420 is a DC block and capacitor 2421 sets the compression delay. In one embodiment, resistor 2812 is 1k Ω, resistor 2813 is 10k Ω, resistor 2815 is 100k Ω, resistor 2817 is 4.12k Ω, resistor 2818 is 1.2k Ω, resistor 2819 is 806 Ω, resistor 2822 is 10k Ω, resistor 2827 is 1k Ω, and resistor 2823 is 100k Ω.

The gain control block 2806 acts as a voltage divider that is controlled by a voltage. The voltage divider is formed by resistor 2827 and the drain-source resistance of FET 2814. The drain-source resistance of the FET2814 is controlled by the voltage applied to the gate of the FET 2814. The output buffer 2810 amplifies the voltage produced by the voltage controlled voltage divider (i.e., the voltage on the drain of FET 2814) and provides an output voltage on pin P12. The bias circuit 2802 biases the FET2814 into the linear operation region. The peak detection circuit 2804 detects the peak amplitude of the signal provided at pin P10 and reduces the "gain" of the gain control 2806 (by changing the drain-source resistance of the FET 2814) in response to the peak amplitude increasing.

Fig. 29 is a schematic diagram of the horizontal sound image enhancing block 2406. In block 2406, the left channel signal 2630 from the bass module 2404 is connected to a first terminal of the resistor 2903 and a first terminal of the resistor 2901. A second terminal of the resistor 2901 is connected to ground. Right channel signal 2631 from bass module 2404 is provided to a first terminal of resistor 2904 and a first terminal of resistor 2902. A second terminal of resistor 2902 is connected to ground.

A second terminal of resistor 2903 is connected to a first terminal of resistor 2905 and to a non-inverting input of opamp 2914. A second terminal of resistor 2904 is connected to a first terminal of capacitor 2906 and to a non-inverting input of opamp 2912. A second terminal of capacitor 2906 is connected to a second terminal of resistor 2905.

An inverting input of opamp2912 is connected to the first terminal of capacitor 2911, the first terminal of capacitor 2907, the first terminal of capacitor 2910, and pin P21. An output of opamp2912 is connected to a first terminal of resistor 2913, pin P22, and a second terminal of capacitor 2911.

An inverting input of opamp 2914 is connected to a first terminal of capacitor 2915, pin P19, a first terminal of resistor 2908, and a first terminal of resistor 2909. A second terminal of resistor 2909 is connected to a second terminal of capacitor 2910. A second terminal of resistor 2908 is connected to a second terminal of capacitor 2907. An output of opamp 2914 is connected to a first terminal of resistor 2917, pin P20, and a second terminal of capacitor 2915.

A second terminal of resistor 2913 is connected to pin P24 as the right channel output. A second terminal of resistor 2917 is connected to pin P23 as the left channel output. A variable resistor 2430 connected between pins P19 and P20 controls the apparent spatial image width of the left channel. A variable resistor 2431 connected between pins P21 and P22 controls the apparent spatial image width of the right channel. In one embodiment, the variable resistors 2930 and 2931 are mechanically coupled such that changing one resistance changes the other.

In one embodiment, resistors 2901 and 2902 are 100k Ω, resistors 2903 and 2904 are 10k Ω, resistor 2905 is 8.66k Ω, resistor 2908 is 15k Ω, resistor 2909 is 30.1k Ω, and resistors 2917 and 2913 are 200 Ω. In one embodiment, the capacitance 2906 is 0.018 μ f, the capacitance 2907 is 0.001 μ f, the capacitance 2910 is 0.082 μ f, and the capacitances 2915 and 2911 are 22 pf. In one embodiment, variable resistors 2430 and 2431 have a maximum resistance of 100k ohms. In one embodiment, the opamp is of the TL074 type or equivalent.

Fig. 30 is a schematic diagram of a correction system 3000 that may be used as the stereo image enhancement system 124. System 3000 includes a differential amplifier that provides a common mode characteristic 3020 and a differential mode characteristic 3022.

System 3000 includes two transistors 3010 and 3012; a plurality of capacitors 3020, 3022, 3024, 3026, and 3028; and a plurality of resistances 3040, 3042, 3044, 3046, 3048, 3050, 3052, 3054, 3056, 3058, 3060, 3062 and 3064. Located between transistors 3010 and 3012 are three divider networks 3070, 3072, and 3074. First frequency-dividing network 3070 includes a resistor 3060 and a capacitor 3024. The second frequency-dividing network 3072 includes a resistor 3062 and a capacitor 3026, and the third frequency-dividing network 3074 includes a resistor 3064 and a capacitor 3028.

The LEFT input terminal 3000(LEFT IN) supplies a LEFT input signal to the base of the transistor 3010 through a capacitor 3020 and a resistor 3040. A power supply Vcc3040 is connected to the base of the transistor 3010 through a resistor 3046. The power supply Vcc3040 is also connected to the collector of the transistor 3010 through a resistor 3046. The base of the transistor 3010 is also connected to ground 3041 through a resistor 3044, while the emitter of the transistor 3010 is connected to ground 3041 through a resistor 3048.

Capacitor 3020 is a decoupling capacitor that provides Direct Current (DC) isolation of the input signal on the left input 3000. On the other hand, the resistors 3042, 3044, 3046, and 3048 constitute a bias circuit which provides stable operation of the transistor 3010. Specifically, resistors 3042 and 3044 set the base voltage of the transistor 3010. Resistor 3046, in combination with third divider network 3074, sets the dc value of the collector-to-emitter voltage of transistor 3010. Resistor 3048, in combination with first and second frequency-dividing networks 3070 and 3072, sets the DC current of the emitter of transistor 3010.

In one embodiment, the transistor 3010 is an NPN 2N2222A transistor, which is generally available from various transistor manufacturers. The capacitance 3020 was 0.22 microfarads. Resistance 3040 is 22 kilo-ohms (k Ω), resistance 3042 is 41.2k Ω, resistance 3046 is 10k Ω, and resistance 3048 is 6.8k Ω. However, those skilled in the art will appreciate that various transistors, capacitors, and resistors having different values may be employed.

The right input terminal 3002 provides a right input signal to the base of the transistor 3012 through a capacitor 3022 and a resistor 3050. A power supply Vcc3040 is connected to the base of the transistor 3012 through a resistor 3052. The power supply Vcc3040 is also connected to the collector of the transistor 3012 through a resistor 3056. The base of the transistor 3012 is also connected to ground 3041 through a resistor 3054, and the emitter of the transistor 3012 is connected to ground 3041 through a resistor 3058.

Capacitor 3022 is a decoupling capacitor that provides Direct Current (DC) isolation of the input signal at the right input terminal 3002. On the other hand, the resistors 3052, 3054, 3056, and 3058 constitute a bias circuit which provides stable operation of the transistor 3012. Specifically, resistors 3052 and 3054 set the base voltage of transistor 3012. Resistor 3056, in combination with third divider network 3074, sets the dc value of the collector-to-emitter voltage of transistor 3012. Resistor 3058, in combination with first and second frequency-dividing networks 3070 and 3072, sets the DC current of the emitter of transistor 3012.

In one embodiment, the transistor 3012 is an NPN 2N2222A transistor, which is generally available from various transistor manufacturers. The capacitance 3022 was 0.22 microfarads. The resistor 3050 is 22 kilo-ohms (k Ω), the resistor 3052 is 41.2k Ω, the resistor 3056 is 10k Ω, and the resistor 3048 is 6.8k Ω. However, those skilled in the art will appreciate that various transistors, capacitors, and resistors having different values may be employed.

System 3000 produces two voltage gains: common mode voltage gain and differential voltage gain. The common mode voltage gain is a change in voltage common to the left and right input terminals 3000 and 3002. The differential gain is a change in the output voltage due to a difference between voltages applied to the left and right input terminals 3000 and 3002.

In system 3000, the common mode gain is designed so as to reduce clipping that may be caused by high amplitude input signals. In one embodiment, the common mode gain at the left output 3004 is primarily defined by the resistors 3040, 3042, 3044, 3046, and 3048. In one embodiment, the common mode gain is about 6 decibels.

Frequencies below about 30 hertz (Hz) are de-emphasized more than frequencies above about 30 Hz. For frequencies above about 30Hz, each frequency is reduced uniformly by about 6 decibels.

However, for a given implementation, the common mode gain may be changed by changing the values of the resistances 3040, 3042, 3044, 3050, 3052, and 3054.

The differential gain between the left and right outputs 3004 and 3006 is defined primarily by the ratio of resistors 3046 and 3048, the ratio of resistors 3056 and 3058, and the three divider networks 3070, 3072, and 3074. As discussed in more detail below, one embodiment equalizes certain frequency ranges in a differential input. Therefore, the differential gain varies according to the frequency of the left and right input signals.

Because the crossover networks 3070, 3072, and 3074 equalize the various frequency ranges in the differential inputs, the frequencies in the differential signals can be changed without affecting the frequencies in the common mode signal. Thus, one embodiment may create enhanced audio sounds in a completely unique and novel manner. In addition, the differential perspective correction device 102 may implement many other audio enhancement systems more simply and less expensively.

Referring now to the three crossover networks 3070, 3072 and 3074, crossover networks 3070, 3072 and 3074 act as filters that spectrally shape the differential signal. A filter is typically characterized by a cut-off frequency that separates the pass band of frequencies from the stop band of frequencies. The cut-off frequency is the frequency marking the edge of the pass band and the beginning of the transition to the stop band. Typically, the cutoff frequency is a frequency that is de-emphasized by three decibels relative to other frequencies in the passband. The pass bands of frequencies are those frequencies that pass through the filter without substantial equalization or attenuation. On the other hand, stopbands of frequencies are those frequencies that are equalized or attenuated by the filter.

Fig. 31 shows an embodiment of the invention with only a first frequency-dividing network 3070. First frequency-dividing network 3070 includes a resistor 3060 and a capacitor 3024, which interconnect the emitters of transistors 3010 and 3012. Because the first frequency-dividing network 3070 equalizes frequencies in the lower portion of the spectrum, it is referred to as a high-pass filter. In one embodiment, resistor 3060 has a value of about 27.01k Ω and capacitor 3024 has a value of about 0.68 microfarads.

The values of resistor 3060 and capacitor 3024 are selected to define a cut-off frequency in the low frequency range. In one embodiment, the cutoff frequency is about 78Hz, the stop band is below about 78Hz, and the pass band is above about 78 Hz. Frequencies below about 78Hz are de-emphasized relative to frequencies above about 78 Hz. However, because the first frequency-splitting network 3070 is only a first order filter, each frequency that defines a cutoff frequency is a design goal. The actual eigenfrequency may vary for a given implementation. In addition, other values may be selected for resistor 3060 and capacitor 3024 to change the cutoff frequency to deemphasize other desired frequencies.

Fig. 32 is a schematic diagram of a differential perspective correction unit 3200 with second and third crossover networks. Like the first frequency-dividing network 3070, the second frequency-dividing network 3072 is preferably a filter that equalizes certain frequencies in the differential signal. However, unlike the first frequency-dividing network 3070, the second frequency-dividing network 3072 is a high-pass filter that de-emphasizes lower frequencies in the differential signal relative to higher frequencies in the differential signal.

As shown in fig. 32, a second frequency-dividing network 3072 interconnects the emitters of transistors 3010 and 3012. In addition, the second frequency-dividing network 3072 includes a resistor 3062 and a capacitor 3026. Preferably, resistor 3062 has a value of about 1k Ω and capacitor 3026 has a value of about 0.01 microfarads.

These values are selected to define a cut-off frequency in the high frequency range. In one embodiment, the cutoff frequency is approximately 15.9 kilohertz (kHz). Frequencies in the stop band below about 15.9kHz are de-emphasized relative to frequencies in the pass band above 15.9 kHz.

However, because the second frequency-dividing network 3072 is a first order filter, as is the first frequency-dividing network 3070, each frequency defining the passband is the design objective. The actual eigenfrequency may vary for a given implementation. Still further, other values may be selected for resistor 3062 and capacitor 3026 to deemphasize other desired frequencies.

Referring now to fig. 33, a third divider network 3074 interconnects the collectors of transistors 3010 and 3012. The third crossover network 3074 includes a resistor 3064 and a capacitor 3028 that are selected to form a low pass filter that deemphasizes frequencies above the mid-range frequency. In one embodiment, the cut-off frequency of the low-pass filter is about 795 Hz. Preferably, resistor 3064 has a value of about 9.09k Ω and capacitor 3028 has a value of about 0.022 microfarads.

In the correction produced by the third crossover network 3074, frequencies in the stop band above about 795Hz are de-emphasized relative to frequencies in the pass band below about 795 Hz. As described above, since the third dividing network 3074 is only a first order filter, it is the frequency defining the low pass filter in the third dividing network 3074 that is the design objective. These frequencies may vary for a given implementation. Again, other values may be selected for resistor 3064 and capacitor 3028 to change the cutoff frequency to deemphasize other desired frequencies.

In operation, the first, second and third crossover networks 3070, 3072 and 3074 combine to spectrally shape the differential signal.

The overall correction curve 2300 (shown in FIG. 23) is defined by two inflection points (labeled as point A and point B). At point a (about 125Hz in one embodiment), the slope of the correction curve changes from a positive value to a negative value. At point B (about 1.8kHz in one embodiment), the correction curve changes from a negative value to a positive value.

Thus, frequencies below about 125Hz are de-emphasized relative to frequencies near 125 Hz. Specifically, below 125Hz, the gain of the overall correction curve 2300 decreases at a rate of about 6dB per octave. De-emphasis of signal frequencies below 125Hz prevents very low frequencies (i.e., bass) from being over emphasized. In many audio reproduction systems, over-emphasizing the audio signal in such a low frequency range relative to higher frequencies may produce an unpleasant and distorted sound image with a too strong bass response. Moreover, excessive emphasis on these frequencies can damage various audio components, including speakers.

Between points a and B, the slope of an overall correction curve is negative. That is, frequencies between about 125Hz and about 1.8kHz are de-emphasized relative to frequencies near 125 Hz. Thus, the gain related to frequency between points A and B is reduced at a variable rate to a maximum equalization point-8 dB at about 1.8 kHz.

Above 1.8kHz, the gain is raised at a variable rate up to about 20kHz, about the highest frequency audible to the human ear. That is, frequencies above about 1.8kHz are emphasized over frequencies near 1.8 kHz. Thereby, the gain associated with frequencies above point B increases at a variable rate towards 20 kHz.

These relative gain and frequency values are merely for design purposes, and the actual numbers may vary from system to system. Also, the gain and frequency values may be changed according to the sound type or user preference without departing from the spirit of the present invention. For example, changing the number of crossover networks and changing the resistance and capacitance values within each crossover network makes the overall perspective correction curve 2300 appropriate for the type of sound being reproduced.

Selective equalization of the differential signal enhances the ambient or reverberant sound effect present in the differential signal. As described above, frequencies in the differential signal are easily perceived at appropriate levels in the live sound field. Unfortunately, when playing a recorded performance, the sound image does not provide the same 360 degree effect as a live performance. However, by equalizing the frequencies of the differential signals, the projected sound image can be significantly broadened to reproduce the experience of a live performance through a pair of loudspeakers positioned in front of the listener.

The equalization of the differential signal according to the overall correction curve 2300 is to de-emphasize signal components of statistically lower intensity relative to signal components of higher intensity. The higher intensity differential signal components of a typical audio signal are found in the mid-range frequencies between about 1 to 4 kHz. In this frequency range, the human ear has an enhanced sensitivity. Thus, the enhanced left and right output signals produce a more improved audio effect.

The number of frequency dividing networks and the components within the frequency dividing networks may be varied in another embodiment to simulate head transfer functions (HRTFs). Advantageously, HRTF-based transfer functions can be applied to differential signals to locate an immersive sound effect (ambient) in order to create a sufficiently osmotically located sound field.

Fig. 33 shows a differential perspective correction device 3300 that allows the user to change the amount of overall differential gain. In this embodiment, a fourth frequency-splitting network 3301 interconnects the emitters of transistors 3010 and 3012. In this embodiment, the fourth frequency-dividing network 3301 includes a variable resistor 3302.

The variable resistor 3302 acts as a level adjustment device, preferably a potentiometer or a variable resistor-like device. Changing the resistance value of variable resistor 3302 raises and lowers the relative balance of the entire perspective correction circuit. The adjustment of the variable resistor is generally performed manually so that the user can change the level and aspect (aspect) of the differential gain according to the type of reproduced sound and according to the personal preference of the user. In general, a reduction in the overall level of the differential signal reduces the ambient sound information, thereby creating a narrowed acoustic image sensation.

FIG. 34 illustrates a differential perspective correction device 3400 that allows a user to change the amount of common mode gain. The differential perspective correction device 3400 includes a fourth frequency-division network 3401. The fourth frequency-splitting network 3401 includes a resistor 3402, a resistor 3404, a capacitor 3406, and a variable resistor 3408. The capacitor 3406 removes differential information and allows the variable resistance 3402 and the resistance 3404 to change the common mode gain.

Resistors 3402 and 3404 may be of various values depending on the desired common mode gain range. On the other hand, the variable resistor 3408 functions as a level adjustment device to adjust the common mode gain within a desired range. Desirably, the variable resistance 3408 is a potentiometer or similar variable resistance device. Changing the resistance of variable resistor 3408 affects transistors 3010 and 3012 to the same extent, thereby raising and lowering the relative equalization of the overall common mode gain.

The adjustment of the variable resistance is typically done manually so that the user can customize the level and characteristics of the common mode gain. The increase in common mode gain emphasizes the audio information common to the input signals 3002 and 3004. For example, increasing the common mode gain in a sound system will emphasize audio information at the center stage between a pair of speakers.

Fig. 35 illustrates a differential perspective correction apparatus 3500 having a first frequency-dividing network 3501 located between the emitters of transistors 3010 and 3012 and a second frequency-dividing network 3502 located between the collector stages of transistors 3010 and 3012.

The first frequency-splitting network 3501 is a high-pass filter that de-emphasizes frequencies in the lower portion of the spectrum. In this embodiment, the first frequency-dividing network 3501 includes a resistor 3510 and a capacitor 3512. The values of resistor 3510 and capacitor 3512 are selected to define a high pass filter with a cutoff frequency of about 350 Hz. Thus, the value of resistor 3510 is about 27.01k Ω and the value of capacitor 3512 is about 0.15 microfarads. In operation, frequencies below 30Hz are de-emphasized relative to frequencies above 350 Hz.

A second frequency-dividing network 3502 interconnects the collectors of the transistors 3010 and 3012. The second frequency dividing network 3502 is a low pass filter that de-emphasizes frequencies in the lower part of the spectrum. In this embodiment, the second frequency-dividing network 3502 includes a resistor 3520 and a capacitor 3522.

The values of resistor 3520 and capacitor 3522 are selected to define a low pass filter with a cutoff frequency of about 27.3 kHz. Thus, resistor 3520 has a value of about 9.09k Ω and capacitor 3522 has a value of about 0.0075 microfarads. In operation, frequencies above 27.3kHz are de-emphasized relative to frequencies below 27.3 kHz.

The first and second frequency-dividing networks 3501 and 3502 combine to spectrally shape the differential signal. Frequencies below about 5kHz are de-emphasized relative to frequencies near 5 kHz. Specifically, below 5kHz, the gain of the overall correction increases at a rate of about 5dB per octave. Further, above 5kHz, the gain of the overall correction curve 1400 also decreases at a rate of about 5dB per octave.

The above-described embodiment of the differential perspective correction apparatus may further include an output buffer 3630 as shown in fig. 36. The output buffer 3630 is designed so as to isolate the perspective correction differential device from load variations present in the circuits connected to the left and right output terminals 3004 and 3006. For example, when the left and right output terminals 3004 and 3006 are connected to a pair of speakers, the impedance load of the speakers does not change the way the differential perspective correction device equalizes the differential signals. Therefore, without the output buffer 3630, the circuitry, speakers, and other components may affect the way the differential perspective correction device 102 equalizes the differential signals.

In one embodiment, left output buffer 3630A includes left output transistor 3601, resistor 3604, and capacitor 3604. The power supply Vcc3040 is directly connected to the collector of the transistor 3601. The collector of the transistor 3601 is connected to ground 3041 through a resistor 3604 and to the left output terminal 3004 through a capacitor 3602. Further, a base of the transistor 3601 is connected to a collector of the transistor 3010.

In one embodiment, transistor 3601 is an NPN 2N2222A transistor, resistor 3604 is 1k Ω, and capacitor 3602 is 0.22 microfarads. The resistor 3604, the capacitor 3602, and the transistor 3601 constitute a unity gain. That is, the left output buffer 3630A mainly transfers the enhanced sound signal to the left output terminal 3004 without further equalizing the enhanced sound signal.

Similarly, a right output buffer 3630B includes a right output transistor 3610, a resistor 3612, and a capacitor 3614. The power supply Vcc3040 is directly connected to the collector of the transistor 3610. The collector of the transistor 3610 is connected to ground 3041 through a resistor 3612 and to the right output terminal through a capacitor 3614. Further, a base of the transistor 3610 is connected to a collector of the transistor 3012.

In one embodiment, transistor 3610 is an NPN 2N2222A transistor, resistor 3612 is 1k Ω, and capacitor 3614 is 0.22 microfarads. Resistor 3612, capacitor 3614 and transistor 3610 produce unity gain. That is, the right output buffer 3630B mainly transmits the enhanced sound signal to the right output terminal 3006 without equalizing the enhanced sound signal.

Those skilled in the art will appreciate that output buffer 3630 may also be implemented with other amplifiers, such as opamps, etc.

Fig. 37 shows another embodiment of the stereoscopic image enhancement processor 124. In fig. 37, left input 2630 is provided to a first terminal of resistor 3710, a first terminal of resistor 3716, and a first terminal of resistor 3740. A second terminal of resistor 3710 is connected to a first terminal of resistor 3711 and an inverting input of opamp 3712. Right input 2631 is provided to a first terminal of resistor 3713, a first terminal of resistor 3741, and a first terminal of resistor 3746. A second terminal of resistor 3713 is connected to a first terminal of resistor 3714 and a non-inverting input of opamp 3712. A second terminal of resistor 3714 is connected to ground. A second terminal of resistor 3740 and a second terminal of resistor 3741 are connected to a non-inverting input of opamp3744 and a first terminal of resistor 3742. A second terminal of resistor 3742 is connected to ground.

The output of opamp3744 is provided to a first terminal of resistor 3761. A second terminal of resistor 3761 is connected to the inverting input of opamp 3744. A second terminal of resistor 3743 is connected to ground. Returning to opamp 3712, the output of opamp 3712 is provided to the second terminal of resistor 3711. The output of opamp 3712 is also provided to a first terminal of resistor 3715. A second terminal of resistor 3715 is connected to a first terminal of capacitor 3717. A second terminal of capacitor 3717 is connected to a first terminal of resistor 3718, a first terminal of resistor 3719, a first terminal of capacitor 3721, and a first terminal of resistor 3722. A second terminal of resistor 3718 is connected to ground. A second terminal of resistor 3719 is connected to a second terminal of resistor 3720 and to a second terminal of resistor 3725. A second terminal of capacitor 3721 is connected to a first terminal of resistor 3720 and a first terminal of resistor 3023. A second terminal of resistor 3722 is connected to a first terminal of resistor 3725 and a first terminal of capacitor 3724. A second terminal of resistor 3023 and a second terminal of capacitor 3024 are both grounded.

A second terminal of resistor 3719 is also provided to a first terminal of resistor 3726 and an inverting input of opamp 3727. The non-inverting input of opamp 3727 is grounded. A second terminal of resistor 3726 is connected to an output of opamp 3727. The output of opamp 3727 is connected to a first fixed terminal of potentiometer 3728. A second fixed terminal of potentiometer 3728 is connected to ground. The wiper of potentiometer 3728 is connected to a second terminal of resistor 3747 and a first terminal of resistor 3729.

The output of opamp3744 is connected to a first fixed terminal of potentiometer 3745. A second fixed end of potentiometer 3745 is connected to ground. The wiper of potentiometer 3745 is connected to a first terminal of resistor 3730 and a first terminal of resistor 3751. A second terminal of resistor 3747 is connected to a first terminal of resistor 3748 and an inverting input of opamp 3749.

The non-inverting input of opamp 3749 is grounded. An output terminal of opamp 3749 is connected to a second terminal of resistor 3748 and a first terminal of resistor 3750. A second terminal of resistor 3750 is connected to a second terminal of resistor 3729. A second terminal of resistor 3730 is connected to a non-inverting input of opamp 3753. The first terminal of resistor 3731 is also connected to the non-inverting input of opamp 3735. A second terminal of resistor 3731 is connected to ground. An inverting input of opamp 3735 is connected to a first terminal of resistor 3734 and a first terminal of resistor 3732. A second terminal of resistor 3732 is connected to ground. The output of opamp 3735 is connected to the second terminal of resistor 3734. A second terminal of resistor 3750, a second terminal of resistor 3751, a second terminal of resistor 3746, and a first terminal of resistor 3752 are connected to a non-inverting input of opamp 3755. A second terminal of resistor 3752 is connected to ground. A non-inverting input of opamp3755 is connected to a first terminal of resistor 3753 and a first terminal of resistor 3754. The output of opamp3755 is connected to a second terminal of resistor 3754.

The output of opamp 3735 is provided as the left channel output and the output of opamp3755 is provided as the right channel output.

Resistors 3710, 3711, 3713, 3714, 3740, 3741, 3742, 3743, 37 and 3761 are all 33.2k Ω resistors. Resistors 3716 and 3746 are each 80.6k Ω. Potentiometers 3745 and 3728 are each 10.0K linear potentiometers. Resistance 3715 is 1.0K, capacitance 3717 is 0.47 μ f, resistance 3718 is 4.42K, resistance 3719 is 121K, capacitance 3721 is 0.0047 μ f, resistance 3720 is 47.5K, resistance 3722 is 1.5K, resistance 3723 is 3.74K, resistance 3725 is 33.2K, and capacitance 3724 is 0.47 μ f. Resistor 3726 is 121K. Resistors 3747 and 3748 are each 16.2K. Resistors 3729 and 3750 are both 11.5K. Resistors 3730 and 3751 are each 37.9K. Resistors 3731, 3732, 3752, and 3753 are all 16.2K. Resistors 3734 and 3754 are each 38.3K. opamp 3712, 3744, 3727, 3749, 3735 and 3755 are all TL074 types or equivalents.

Digital signal processor implementation

The acoustic correction system can also be easily implemented in software as described in connection with fig. 3. Suitable processors include general purpose processors, Digital Signal Processors (DSPs), and the like.

Fig. 38 is a block diagram of a software embodiment of acoustic correction system 120. In fig. 38, the left channel input 3801 is provided at the input of the 10db attenuator 3803. The output of the attenuator 3803 is provided to an input terminal of the filter 3804 and a first throw terminal of a DPDT (double pole double throw) switch 3805. The output of the filter 3804 is provided to a second throw of the switch 3806. The right channel input 3802 is provided to an input of a 10db attenuator 3806. The output of the attenuator 3806 is provided to an input of the filter 3807 and a first throw of the switch 3805. The output of the filter 3807 is provided to a second throw of the switch 3805.

A first pole of the switch 3805 is connected to a first input of the adder 3828 and a first input of the adder 3808. A second pole of the switch 3805 is connected to a first input of the adder 3829 and a second input of the adder 3808. The output of the adder 3808 is provided to an input of the low pass filter 3809. The output of low pass filter 3809 is provided to an input of a dual-band bandpass filter 3810, an input of a dual-band bandpass filter 3811, and an input of a 100Hz bandpass filter 3812.

The output of filter 3810 is provided to a first input of adder 3821, the output of filter 3811 is provided to a second input of adder 3821, and the output of filter 3812 is provided to a third input of adder 3812. The output of adder 3821 is provided to an input of a 2.75dB amplifier 3863, a first input of multiplier 3824, and an input of absolute value block 3822. The output of the absolute value block 3822 is provided to the input of a fast increase slow decay (FASD) compressor 3823. The output of FASD compressor 3823 is provided to a second input of multiplier 3824.

The output of amplifier 3863 is provided to the positive input of subtractor 3825. The output of multiplier 3824 is provided to the negative input of subtractor 3825. The output of subtractor 3825 is provided to a first input of multiplier 3826. The output of bass control 3827 is provided to a second input of multiplier 3826. The output of multiplier 3826 is connected through an SPDT switch 3860 to a second input of adder 3828 and a second input of adder 3829.

The output of adder 3828 is provided to a first input of adder 3830, an input of 9dB attenuator 3833, a positive input of subtractor 3837, and a first throw of DPDT switch 3836. The output of adder 3829 is provided to the negative input of subtractor 3837, to the second input of adder 3830, to the input of 9dB attenuator 3834, and to the first throw of switch 3836.

The output of summer 3830 is provided to the input of a 5dB attenuator 3832. The output of attenuator 3832 is provided to a first input of adder 3835 and a first input of adder 3866. The output of attenuator 3833 is provided to a second input of adder 3835. The output of attenuator 3834 is provided to a second input of adder 3866. The output of summer 3835 is provided to a second throw terminal of switch 3836. The output of summer 3866 is provided to a second throw terminal of switch 3836.

The output of subtractor 3837 is provided to the input of a 48Hz high pass filter 3838. The output of the high pass filter 3838 is provided to the input of a 6dB attenuator 3840, to the input of a 7kHz high pass filter 3841 and to the input of a 200Hz low pass filter 3842. The output of attenuator 3840 is provided to a first input of an adder 3844, the output of high pass filter 3841 is provided to a second input of adder 3844, and the output of low pass filter 3842 is provided to a third input of adder 3844 through a 3db attenuator 3843. The output of adder 3844 is provided to a first input of multiplier 3845. The output of width control 3846 is provided to a second input of multiplier 3845. The output of multiplier 3845 is provided to a third input of adder 3835 and through an inverter (i.e., a gain of-1) to a third input of adder 3866.

The first switch of switch 3836 is connected to the left channel output 3850. The second pole of switch 3836 is connected to right output 3851.

As shown in fig. 38, left and right stereo input signals are provided to left and right inputs 3801 and 3802, respectively. For the processed bass enhancement portion (corresponding to the bass enhancement block 101 shown in fig. 1), the left and right channels are added together by adder 3808, processed as a mono signal, and then added back to the left and right channels by adders 3828 and 3829, thereby forming an enhanced stereo signal. The bass information is processed as a mono signal, and because there is usually little stereo separation in the bass frequency signal, there is little need to process both channels repeatedly.

FIG. 38 illustrates that software user controls include: software control 3827 for controlling an amount of bass enhancement; software control 3846 for controlling the width of the apparent sound field; and software switches 3805, 3860, and 3836 for enabling or disabling vertical, bass, and width sound image enhancement, respectively. These user controls may be either dynamically altered or fixed to a particular configuration, depending on the application. The user controls may be "connected" to controls, such as sliders, check boxes, etc. in a dialog box, so that the user may control the operation of the acoustic correction system.

In fig. 38, the left and right inputs 3801 and 3802 are first processed with a-10 dB gain to set the bypass level and prevent saturation of the signal during later processing. The channels are then processed through boosting filters (filters 3804 and 3807 for the left and right channels, respectively) that perform the boosting and expansion of the sound field described in connection with fig. 4-6.

After the boost filter, the left and right channels are mixed together and passed through a low pass filter 3809 and then through a band pass filter bank 3810 and 3812. The low pass filter 3809 has a cutoff frequency of 284 Hz. The following three filters 3810-3812 are all second-order bandpass filters. The filter 3810 may be selected to be 40Hz or 150 Hz. The filter 3811 may be selected to be 60Hz or 200 Hz. Thus, there are three configurations available for the specification of the speaker: small, medium and large. All three configurations employ three band pass filters, but the center frequencies of filters 3810 and 3811 are different.

The outputs of the three active filters are then summed together by summer 3821 and provided to the bass control stage.

The bass control stage includes an expander circuit with an absolute value detector 3822, a fast-up slow-decay peak detector 3823, and a multiplier 3824. The output of the peak detector 3823 is used as a multiplier for the expander input signal to expand the dynamic range of the signal.

The second portion of the bass control stage subtracts an expanded version of the input signal of the stage from the same input signal with a 2.75dB gain applied by amplifier 3863. This has the effect of limiting the level of high amplitude signals while adding a small constant gain to lower amplitude signals.

The output of the bass control stage is added to the left and right channel signals by adders 3828 and 3829, respectively. The amount of enhanced bass signal mixed into the left and right channels is determined by bass control 3827.

The resulting left and right channel signals are then added together by adder 3830 to form an L + R signal, which is subtracted by subtractor 3837 to form an L-R signal. The L-R signal is processed through a perspective curve (see fig. 7) to spectrally shape it, which is implemented with a filter network and gain adjustments as follows. First, the signal passes through a 48Hz high pass filter 3838. The output of this filter is then split and passed through a 7kHz high pass filter 3841 and a 200Hz low pass filter 3842. The three filter outputs are then summed by summer 3844 to form a perspective curve signal with the following gain adjustments: -6dB for a 48Hz high pass filter 3838, 0dB (without adjustment) for a 7kHz high pass filter 3841 and +3dB for a 200Hz low pass filter 3842. Width control 3846 determines the amount of the perspective curve signal that passes through the last adders 3835 and 3866.

Finally, the left, right, L + R, and L-R signals are mixed by summers 3835 and 3866 to produce the final left and right channel outputs, respectively. The left channel output is formed by mixing the-5 dB gain adjusted L + R signal, the-9 dB gain adjusted left channel signal, and the perspective curve signal without gain adjustment (other than that provided by width control 3846). The right channel output is formed by mixing the-5 dB gain adjusted L + R signal, the-9 dB gain adjusted right channel, and the inverse perspective curve signal without gain adjustment (except for width control).

The algorithm for fast-increase slow-decay (FASD) peak detector 3823 is represented in pseudo code as follows:

if[in＞out(previous)]then

out＝in-([in-out(previous)]*attack}

else

out＝in+{[out(previous)-in]*decay}

endif

where out (previous) represents the output from the previous sample period.

The values of boost and decay are related to the sample rate, since the slew rate must be related to real time. Their respective formulas are as follows:

increase 1- (1/(0.01X sample rate))

Attenuation is 1- (1/(0.1 x sample rate)) where the sample rate is in units of number of samples/second.

The input to the FASD peak detector 3123 is always greater than or equal to zero because it is from the output of the absolute value function 3122.

The filters 3809-3812 are implemented as Infinite Impulse Response (IIR) filters at a sampling frequency of 44.1 kHz. These filters are designed using a bilinear transform. Each filter is a second order filter with one section. These filters are implemented using a 32-bit fractional fixed-point algorithm. Specific information for each filter is given in table 1 below. Further, transfer functions of the filters 3810 to 3812 are shown in fig. 39 to 43, respectively. The transfer function of the low pass filter 3809 is shown in fig. 44.

Band-pass filter

Filter frequency (Hz)	-3dB Low (Hz)	Center (Hz)	-3dB high (Hz)	Bandpass gain	Band-pass gain (dB)
						40	30	38.7	50	1.43	3.12
60	45	58.1	75	1.43	3.12
						100	78	96.8	129	1.00	0.0
150	116	145.1	192	1.00	0.0
						200	150	193.6	250	0.71	-2.93
Low-pass filter
								-3dB(Hz)	-15dB(Hz)	Bandpass gain	Band-pass gain (dB)
		285	1021	1.00	0.0

TABLE 1

The bass control 3827 determines the amount of bass boost applied to the audio signal and provides a value between 0 and 1 to the multiplier 3826.

Width control 3846 determines the amount of stereo width enhancement to apply to the final output. The width control provides a value between 0 and 2.82(9dB) to multiplier 3845.

OTHER EMBODIMENTS

The entire acoustic correction system disclosed herein can be readily implemented by software running on a DSP or personal computer, by discrete circuit elements, such as hybrid circuit structures, or within a semiconductor substrate having terminals for adjusting appropriate external elements. Adjustments made by the user currently include levels of low and high frequency energy correction, adjustments of various signal levels including the levels of the sum and difference signals, and orientation adjustments.

From the foregoing description and drawings, it has been shown that the present invention has significant advantages over current acoustic correction and stereo enhancement systems. While the above detailed description has shown, described, and pointed out the fundamental novel features of the invention, it will be understood that various changes, substitutions, and alterations in the form and details of the device described may be made by those skilled in the art without departing from the spirit of the invention. Accordingly, the invention should be limited only by the scope of the following claims.

Claims (28)

1. A sound enhancement system comprising:

a sound image correction module configured to correct a perceived height of an apparent sound field reproduced by the plurality of speakers;

a bass boost module configured to correct a perceived bass response of the speaker; and

a sound image enhancement module configured to correct a perceived width of the apparent sound field,

wherein the correction provided by the lip image correction module precedes the correction provided by the bass enhancement module, and wherein the correction provided by the bass enhancement module precedes the correction provided by the lip image enhancement module.

2. The sound enhancement system of claim 1, wherein the image correction module is further configured to correct a perceived vertical position of the apparent sound field.

3. The sound enhancement system of claim 1 wherein the image correction module includes a left channel filter to filter sound in a left signal channel and a right channel filter configured to filter sound in a right signal channel.

4. A sound enhancement system as claimed in claim 3, wherein the left channel filter and the right channel filter are configured so as to filter the left and right channels in accordance with a change in the frequency response of the human auditory system as a function of the vertical position of the sound source.

5. The sound enhancement system of claim 3 wherein the left channel filter and the right channel filter are configured to emphasize lower frequencies relative to higher frequencies.

6. The sound enhancement system of claim 1, wherein the bass enhancement module is configured to emphasize lower frequency portions relative to higher frequencies.

7. The sound enhancement system of claim 1 wherein the bass enhancement module is configured to receive a plurality of input signals and emphasize a common mode portion of lower frequencies of the input signals relative to higher frequencies of the input signals.

8. The sound enhancement system of claim 1, wherein the bass enhancement module comprises:

a first combiner configured to combine at least a portion of the left channel signal with at least a portion of the right channel signal, thereby generating a combined signal;

a filter configured to select a portion of the combined signal to produce a filtered signal;

a variable gain module configured to adjust the filtered signal in response to an envelope of the filtered signal to produce a bass enhancement signal;

a second combiner configured to combine at least a portion of the bass enhancement signal with the left channel signal; and

a third combiner configured to combine at least a portion of the bass enhancement signal with the right channel signal.

9. The sound enhancement system of claim 8, wherein the variable gain module comprises an expander.

10. The sound enhancement system of claim 8, wherein the variable gain module comprises a compressor.

11. The sound enhancement system of claim 1, wherein the image enhancement module is configured to receive an input signal comprising a left channel input and a right channel input, the image enhancement module further configured to provide a common mode characteristic in response to a common mode portion of the input signal and to provide a differential mode characteristic in response to a differential mode portion of the input signal.

12. The sound enhancement system of claim 1, wherein the image enhancement module is configured to provide a common mode transfer function and a differential mode transfer function.

13. A sound enhancement system as in claim 12, wherein said differential mode transfer function emphasizes lower frequencies relative to higher frequencies.

14. The sound enhancement system of claim 12, wherein the differential mode transfer function is configured to provide first de-emphasis on frequency components in a first frequency band, second de-emphasis on frequency components in a second frequency band, third de-emphasis on frequency components in a third frequency band, and fourth de-emphasis on frequency components in a fourth frequency band, the first frequency band being lower than the second frequency band, the second frequency band being lower than the third frequency band, the third frequency band being lower than the fourth frequency band, the second de-emphasis being less than the first de-emphasis and the third de-emphasis.

15. A method for enhancing an audio sound to improve a perceived sound field and to improve a perceived bass component of the sound, comprising the operations of:

height correction of the sound signal so as to increase the perceived height of the apparent sound field reproduced by the plurality of loudspeakers;

bass boost a sound signal, enhancing a perceived bass response of the loudspeaker;

width enhancing a multi-channel sound signal, widening a perceived width of an apparent sound field produced by the multi-channel sound signal,

wherein the height correction precedes the bass boost, and wherein the bass boost precedes the width boost.

16. The method of claim 15, wherein the height correction operation comprises filtering the sound signal to change a perceived vertical position of the apparent sound field as heard by a listener.

17. The method of claim 16, wherein the height correction operation comprises the operations of filtering the signal in the left signal channel and filtering the signal in the right signal channel.

18. The method of claim 7, wherein the filtering operation comprises adjusting frequency components of the left signal channel and the right signal channel in accordance with changes in a vertical spatial frequency response of human hearing.

19. The method of claim 17, wherein the filtering operation comprises emphasizing lower frequencies relative to higher frequencies.

20. The method of claim 15 wherein the bass boost operation includes emphasizing lower frequency portions relative to higher frequencies.

21. The method of claim 15, wherein the bass enhancement operation includes emphasizing common mode portions of lower frequencies of a multi-channel input signal relative to higher frequencies of the multi-channel input signal.

22. The method of claim 15, wherein the bass enhancement operation comprises operations of:

combining at least a portion of the left channel signal with at least a portion of the right channel signal, thereby producing a combined signal;

filtering the combined signal, thereby generating a filtered signal;

amplifying the filtered signal according to an envelope of the filtered signal, thereby producing a bass enhancement signal;

combining at least a portion of the bass enhancement signal with the left channel signal; and

combining at least a portion of the bass enhancement signal with the right channel signal.

23. The method of claim 22, wherein the amplifying operation comprises compressing the filtered signal for an elevated period.

24. The method of claim 22, wherein the amplifying operation comprises expanding the filtered signal over a decay period.

25. The method of claim 15, wherein the width enhancement operation comprises the operations of: identifying a common mode portion of the multi-channel sound signal and adjusting the common mode portion according to common mode characteristics; and identifying a differential mode portion of the multi-channel sound signal and adjusting the differential mode portion according to a differential mode characteristic.

26. The method of claim 15, wherein the width enhancement operation comprises applying a common-mode transfer function and applying a differential-mode transfer function to the multi-channel sound signal.

27. The method of claim 26, wherein said applying a differential-mode transfer function comprises emphasizing lower frequencies relative to higher frequencies.

28. The method of claim 26, wherein the operation of applying a differential-mode transfer function comprises:

de-emphasis the frequency components in the first frequency band according to the first de-emphasis value;

de-emphasis frequency components in a second frequency band according to a second de-emphasis value, the second frequency band having frequencies higher than the first frequency band;

deemphasizing the frequency components in a third frequency band according to a third de-emphasis value, the third frequency band having frequencies higher than the frequencies of the second frequency band, the second de-emphasis value being relatively smaller than the first de-emphasis value and the third de-emphasis value; and

de-emphasis frequency components in a fourth frequency band according to a fourth de-emphasis value, the fourth frequency band having a frequency higher than the third frequency band, the fourth de-emphasis value being relatively smaller than the first de-emphasis value and the third de-emphasis value.