CN1290068C - Karaoke Scoring Device - Google Patents

Abstract

A karaoke scoring device for a karaoke system includes: a feature extraction element, a similarity measurement unit, and a scoring unit. The karaoke system can accept a target audio input to compare and score with a reference audio input via a scoring device. The reference audio input and the target audio input are sampled and sequentially converted into corresponding multi-frame reference sampling signals and target sampling signals, which are temporarily stored in a storage element. The temporarily stored frame reference sampling signals and target sampling signals are automatically correlated by the feature extraction element to generate a set of reference feature values and a set of target feature values. These feature values are compared and processed by the similarity measurement unit to generate similarity results. The scoring unit calculates the corresponding similarity results obtained from the multi-frame sampling signals to output the final total score.

Description

Karaoke Scoring Device

technical field

The present invention relates to a karaoke scoring apparatus, in particular to a karaoke scoring apparatus applied in a karaoke system to score a singer's singing ability.

Background technique

Existing karaoke system (karaoke) is provided with the main purpose of karaoke scoring apparatus (karaoke scoring apparatus), is to be used for making karaoke system have the ability of scoring for the singer's singing ability.

By using the karaoke system, the singer can easily get the accompaniment of various musical instrument sound effects to sing songs. Compared with the accompaniment by a real person playing a musical instrument, the use of a karaoke system enables a singer who does not have the ability to play a musical instrument to obtain the accompaniment of various musical instrument sound effects without the assistance of other people's musical instrument performance. Moreover, the effect is like a professional musical accompaniment to a commercially available music album.

Due to the convenience of the karaoke system and the fun of singing songs, the karaoke system is mostly used as a tool for providing leisure and entertainment. The main purpose of making the karaoke system capable of scoring the singer's singing ability is also to provide the singer with more fun when using the karaoke system. Thereby there is the setting of karaoke scoring device.

Although the existing karaoke scoring device can give points when the singer sings the end song, because its starting point is to provide fun, therefore, the scoring method of multiple existing karaoke scoring devices is to use random numbers within a high score range. Determining the points to be awarded, the decisions of which often come across as absurd or unacceptable.

US Patent 5,565,639, US Patent 5,567,162, US Patent 5,719,344, US Patent 5,889,224 and US Patent 6,326,536 propose various more accurate scoring methods, but each has its own shortcomings.

Contents of the invention

The object of the present invention is to provide a karaoke scoring system that can reliably give a score to a singer's singing ability.

Another object of the present invention is to provide a karaoke scoring system with reasonable scoring criteria.

According to a kind of karaoke scoring apparatus (karaoke scoring apparatus) of the present invention, it is used in karaoke system, to be used for scoring the singer's singing ability. The karaoke system includes a predetermined reference audio input and accepts a target audio input for comparison and scoring with the reference audio input via the karaoke scoring device.

The karaoke scoring device includes a storage element, a feature extraction element (feature extraction element), a similarity measurement element (similarity measurement element), and a scoring unit.

The reference audio input and the target audio input are sequentially sampled into corresponding multi-frame reference sampling signals (a plural frames of reference sampling signals) and multi-frame target sampling signals (a plural frames of target sampling signals).

The memory element is used to temporarily store at least one frame of reference sampling signal and at least one frame of target sampling signal currently sampled.

The feature extraction element (feature extraction element) is used to perform autocorrelation calculation (autocorrelation calculation) on the frame reference sampling signal temporarily stored in the storage element and a plurality of frame reference sampling signals with different delays to obtain a set of reference features value. The feature extraction component can also be used to automatically correlate the frame target sampling signal temporarily stored in the storage component with a plurality of frame target sampling signals with different delays, so as to obtain a set of target feature values.

A similarity measurement element can perform similarity comparison processing according to the set of target feature values and the set of reference feature values to generate a similarity result corresponding to the frame reference sample signal and the frame target sample signal.

The scoring unit (scoring element) can calculate the corresponding similarity results obtained from multi-frame sampling signals to output a final score (final score).

According to the present invention, the karaoke scoring device can extract the features of the reference human voice input in the reference audio input, that is, the pitch of the human voice in each frame of the reference audio input, to be used as the scoring standard for the target audio input, and the The decimated audio input is converted to corresponding quantized features for detailed comparison. Moreover, reasonable scoring criteria are provided so that when the singer uses the karaoke system to sing songs, on the singing pitch of each frame of audio input, there are corresponding different scores for hits, misses, consecutive hits, and consecutive misses. And the degree of consecutive hits or consecutive mistakes is different, and the degree of bonus or deduction will also change accordingly. Therefore, the present invention provides a karaoke scoring system that can surely give a score to a singer's singing ability. And, the karaoke scoring system provided by the present invention has reasonable scoring criteria.

The advantages and spirit of the present invention can be further understood through the following detailed description of the invention and the accompanying drawings.

Description of drawings

Fig. 1 is a schematic diagram of the karaoke scoring device of the present invention.

Fig. 2 is a schematic diagram of the central frequency of each pitch.

Fig. 3 is the value of τ corresponding to the central frequency of each pitch in Fig. 2 sampled at 44.1K Hz.

Detailed ways

Please refer to FIG. 1, which is a schematic diagram of the karaoke scoring device of the present invention. The karaoke scoring device 10 of the present invention comprises a storage element (memory element) 14. a feature extraction element (featureextraction element) 16. a similarity measurement unit (similarity measurement element) 18. and a scoring unit (scoring element) 20.

The music scoring device 10 is used in a karaoke system (karaoke system) to score a singer's singing ability. When the singer uses the karaoke system to sing a song, the karaoke system can receive the singing voice of the singer as the target audio input 22 . The music scoring device 10 then compares and scores the target audio input 22 with predetermined reference audio inputs 24 included in the karaoke system.

The target audio input 22 is the target vocal input (target vocal input) received by the singer via a microphone or other audio input devices when singing. The reference audio input 24 is formed by mixing reference instrumental input and/or reference vocal input to provide accompaniment music data for the karaoke system. Generally speaking, the reference audio input 24 can be stored in a commercially available music CD, music cassette, or stored in the hard disk of the karaoke system. Existing vocal cassettes, for example, only have reference instrument inputs and typically do not include reference vocal inputs. And some karaoke systems can also use commercially available music CD albums mixed with reference vocal input and reference instrument input as accompaniment. In addition, more advanced accompaniment CDs or accompaniment DVDs can store the reference vocal input and the reference instrument input separately for the convenience of users.

In this embodiment, considering that the target audio input 22 may be in analog or digital form, as shown in Figure 1, an analog/digital conversion unit 12 can be used to convert the target audio input 22 to be sampled into a digital form that is convenient for calculation . In addition, the audio decoding unit 42 is used to perform audio decoding on the reference audio input 24 . The storage element 14 is used for temporarily storing at least one frame of target sampling signal 26 and at least one frame of reference sampling signal 28 currently being sampled. The storage element 14 includes a first storage element 46 and a second storage element 48 . The first storage element 46 and the second storage element 48 can be registers or other storage elements.

The music scoring device 10 samples the reference audio input 24 according to a predetermined sampling frequency of about 44.1K Hz, and sequentially converts into corresponding multi-frame reference sampling signals 28 and temporarily stores them in the first storage element 46 . The music scoring device 10 samples the target audio input 22 according to the sampling frequency, and sequentially converts the target audio input 22 into corresponding multi-frame target sampling signals 26 and temporarily stores them in the second storage element 48 .

And each frame of the reference sampling signal 28 and each frame of the target sampling signal 26 have N sampling points, N=1024. The sampling signal of each frame as mentioned above can be expressed as X(k), k=0 ^- N-1, and can be delayed by different delay times τ to be X(k+τ).

The feature extraction element 16 performs automatic correlation calculation on the frame reference sampling signal 28, X(k) temporarily stored in the storage element 14, and a plurality of frame reference sampling signals 28, X(k+τ) with different delays. (autocorrelation cal culation). In the automatic correlation calculation, X(k) and X(k+τ) are subjected to the following calculation to obtain the automatic correlation function r _xx (τ).

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; xx</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; )</span>$

The feature extraction unit 16 can also be used to perform the automatic correlation calculation on the frame target sampling signal 26 temporarily stored in the storage unit 14 and a plurality of frame target sampling signals 26 with different delays.

After obtaining the autocorrelation function r _xx (τ) relative to the frame reference sampling signal 28, the feature extraction component 16 will also select a set of τ values according to the reference feature value selection criteria (selection criteria for thereference characteristic value) , τ ₀ ~τ _Nr-1 , as a set of reference characteristic values 30. The selection criteria for the reference characteristic values are as follows:

r _xx (τ)≥r _xx (τ-1), r _xx (τ)≥r _xx (τ+1)

r _xx (τ)≥α*(MAX(r _xx (τ))-MIN(r _xx (τ)))+MIN(r _xx (τ))

τ _lowerbound <τ≤τ _upperbound

In the reference feature value selection standard, α is a predetermined constant, MAX(r _xx (τ)) is the maximum value of the autocorrelation function r _xx (τ) when the value of τ is not 0, and MIN( r _xx (τ)) is the minimum value of the function value in the autocorrelation function r _xx (τ) when the value of τ is not 0, and τ _lowerbound is the lower bound of the predetermined value of τ, and τ _upperbound is the upper bound of the predetermined value of τ boundary.

In this embodiment, the selection criterion of the reference feature value is to select the first three maximum values of the function value in the autocorrelation function r _xx (τ) when the value of τ is not 0, that is, N _r =3. Considering that the pitch of the main string of most music is about between 100 Hz and 900 Hz, according to this embodiment, 1024 sampling points are taken, and the automatic correlation calculation is performed by sampling at 44.1K Hz, then the values of τ The range is approximately between 49 (44100/900=49) and 441 (44100/100=441).

Similarly, after obtaining the automatic correlation function r _xx (τ) relative to the target sampling signal 28 of the frame, the feature extraction component 16 will also select a target characteristic value according to the selection criteria for the target characteristic value. A set of τ values, τ ₀ ˜τ _Nm-1 , serves as a set of target characteristic values 32 . In this embodiment, the criterion for selecting the target characteristic value is to select the maximum value of the function value in the auto-correlation function r _xx (τ) when the value of τ is not 0, that is, N _m =1.

The feature extraction component 16 is also provided with a reference feature buffer 35 (feature buffer of reference input) for temporarily storing the reference feature value 30. The reference audio input 24 is the stored music data. According to experience, humans cannot distinguish the change of music within 100 ms, so the features of the reference audio input 24 at least 100 ms are usually extracted and temporarily stored in the reference feature buffer 35 .

Please refer to FIG. 2 and FIG. 3 . FIG. 2 is a schematic diagram of the central frequency of each pitch, and FIG. 3 is a value of τ corresponding to the central frequency of each pitch in FIG. 2 sampled at 44.1K Hz. Each pitch has a corresponding central frequency. For example, the central frequency of central C is 261.626 Hz, and sampling is performed at 44.1K Hz in this embodiment, so the τ value corresponding to central C is 169.

The reference audio input 24 and the target audio input 22 are sound signals, each containing a plurality of different pitches. In the present invention, the respective τ values of the reference audio input 24 and the target audio input 22 are obtained to obtain quantized samples that can be compared with the target human voice input and the reference human voice input. The three τ values representing the reference feature value 30 as described above are used to represent the three pitches in the reference audio input 24 of the frame. A τ value representing the target feature value 32 is used to represent a pitch in the segment of the target audio input 22 .

The similarity measurement unit (similarity measurement element) 18 of Fig. 1 can carry out similarity comparison processing according to the target characteristic value 32 and the reference characteristic value 30, to generate the reference sampling signal 28 corresponding to the frame and the target sampling signal 26 of the frame similarity results.

The similarity comparison process subtracts the target feature value 32 and three reference feature values 30 respectively, and when any one of the absolute values of the subtraction results is less than a given threshold (threshold), then the similarity result is a hit ( Hit), otherwise it is a mistake (Miss). In this embodiment, three reference feature values 30 are selected in each frame of reference audio input 24, and _Nr =3, which is based on the reference audio input 24 that may be mixed with reference instrument input and reference human voice input, the extracted pitch Characteristic In addition to the pitch of the main rhythm of the music, it may contain the pitch of the accompanying temperament. In order to ensure that the pitch of the main chord rhythm that is sufficient to be used as a similarity measurement benchmark is selected, it is usually a reference human voice input, so the selected number is set to three.

In different implementations, the number Nm of the selected target feature values 32 and the number Nr of the reference feature values 30 can be changed according to the format of the reference audio input 24 . For example, when a CD or DVD for accompaniment, which can store reference vocal input and reference instrument input separately, is used as the source of the reference audio output 24, only the reference vocal input can be sampled separately, and Nr can be reduced at this time. However, when using an old-fashioned backing cassette with only reference instrument inputs as the source of the reference audio output, Nr can be increased to select the individual chord pitches of the accompaniment strings, which can include scores that can be used as the target vocal input 22. Baseline tonic pitch. However, this embodiment considers a commercially available music CD album mixed with reference human voice input and reference musical instrument input. According to the experimental results, better scoring effect can be obtained when Nr=3.

It should be noted that changes in the number of selected target characteristic values 32 and reference characteristic values 30 caused by various implementations should also fall within the scope of the claims of the present invention.

The above given threshold (threshold) varies with pitch. The τ value in each group of reference characteristic values 30 has a corresponding threshold value (TH _τ ), which is obtained by the following formula:

${\mathrm{<span\; class="notranslate">\; TH</span>}}_{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{{<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; |</span>$

Among them, FS represents a predetermined sampling frequency, which is 44.1K Hz in this embodiment, FC represents the central frequency of the pitch corresponding to the τ value, and FC _upper and FC _lower represent the two most corresponding pitches corresponding to the τ value. The center frequency adjacent to the upper and lower pitches. For example, please refer to Figure 3 and Figure 2, the reference characteristic value with a τ value of 169 can correspond to a frequency of 261.626K Hz, and its corresponding threshold value is 44100/|1/(293.665+261.626)-1/(246.942 +261.626)|=7.296.

The scoring unit 20 (scoring element) in FIG. 1 can calculate the corresponding similarity results obtained from multi-frame sampling signals to output a final total score 34 (final score). The scoring unit 20 includes a hit count module 36 (hitcount module) and a miss count module 38 (miscount module). The hit counting module 36 will cumulatively count the hits according to the similarity results transmitted from the similarity measurement unit 18, and output a hit count value (hitcount value), expressed as HitCount. The error counting module 38 will accumulatively count the errors according to the similarity result, and output the error count value (miscount value), expressed as MissCount.

The final total score 34 is between the predetermined maximum score Score _Max and the predetermined minimum score Score _Min , and is calculated by the following formula:

$\mathrm{<span\; class="notranslate">\; FinalScore</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; Score</span>}}_{\mathrm{<span\; class="notranslate">\; MAX</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; Score</span>}}_{\mathrm{<span\; class="notranslate">\; Min</span>}}<span\; class="notranslate">\; )</span>\frac{\mathrm{<span\; class="notranslate">\; HitCount</span>}}{\mathrm{<span\; class="notranslate">\; MissCount</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; HitCount</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; S</span>}{\mathrm{<span\; class="notranslate">\; core</span>}}_{\mathrm{<span\; class="notranslate">\; Min</span>}}$

In this way, the karaoke scoring device 10 can compare the target audio input 22 with the reference audio input 24 to obtain a final total score 34 (FinalScore).

The hit counting module 36 accumulatively counts hits according to the similarity results transmitted from the similarity measurement unit 18 . When the similarity result is a hit, the hit count module 36 will add the current hit count value to the hit increment value (hit-increase value), expressed as HitIncrease, and become the updated hit count value, and reset the miss count The value is the default value (default value). When the similarity results are consecutive hits, the hit increment value will also increase accordingly. In other words, when the pitch of one frame of the target audio input 32 continuously matches the pitch of the reference audio input 24 , the karaoke scoring device 10 will give a higher score.

In the same way, regarding the cumulative counting of the error counting module 38, it means that when the similarity result is a mistake, the error counting module 38 will add the current error counting value to the error increment value (miss-increasevalue), which is expressed as MissIncrease, and becomes the updated miss count value and resets the hit count value to the default value. When the similarity results are all errors consecutively, the error increment value will also increase accordingly.

In another embodiment, the similarity comparison process performed by the similarity measurement unit 18 may have the following implementation manner. The reference audio input 24 and the target audio input 22 include audio with a plurality of different pitches, each pitch has a corresponding central frequency, and has a predetermined frequency range, and the similarity comparison process is to find out each set of reference features The similarity result is generated by determining whether the frequency corresponding to the group of target feature values falls within a predetermined frequency range of the same pitch. For example, please refer to FIG. 3 and FIG. 2, the reference characteristic value with a τ value of 169 can correspond to a frequency of 261.626K Hz, then the corresponding frequency range is (246.942+261.626)/2=254.284K Hz to (277.183+ 261.625)/2=269.404K Hz. In this embodiment, if the frequency corresponding to the target feature value falls within the frequency range (254.284KHz˜269.404KHz), it is a hit, otherwise it is a miss.

According to the present invention, the karaoke scoring device 10 can extract the pitch feature of the main chord in the reference audio input 24, so as to be used as the scoring standard of the target audio input 22, and convert the extracted audio input into corresponding quantization characteristics for a detailed comparison. It may refer to the characteristics of the human voice input or the characteristics of the reference music, and provide reasonable scoring criteria, so that when the singer uses the karaoke system to sing songs, on the singing pitch of each frame of audio input, hit, miss, continuous There are corresponding different scores for hits and consecutive misses, and the degree of consecutive hits or consecutive misses is different, and the degree of bonus points or deductions also changes accordingly. Therefore, the present invention provides a karaoke scoring system that can surely give a score to a singer's singing ability. And, the karaoke scoring system provided by the present invention has reasonable scoring criteria.

Through the detailed description of the preferred embodiments above, it is hoped that the features and spirit of the present invention can be described more clearly, rather than limiting the scope of the present invention by the preferred embodiments disclosed above. On the contrary, the intention is to cover various changes and similar arrangements within the scope of the claims of the present invention. Therefore, the scope of the claims of the present invention should be interpreted in the broadest way based on the above description so as to cover all possible changes and similar arrangements.

Claims (15)

1. A karaoke scoring device, used in a karaoke system, to be used for scoring the singer's singing ability, the karaoke system includes a predetermined reference audio input and can accept target audio input, so as to pass through the karaoke scoring device to compare and score with the reference audio input, the karaoke scoring device samples the reference audio input and the target audio input respectively and sequentially converts them into corresponding multi-frame reference sampling signals and multi-frame target sampling signals, The karaoke scoring device contains: A storage element, used to temporarily store at least one frame of reference sampling signal and at least one frame of target sampling signal currently sampled; The feature extraction component is used to perform automatic correlation calculation between the frame reference sampling signal temporarily stored in the storage component and a plurality of frame reference sampling signals with different delays, so as to obtain a set of reference feature values, and the feature extraction component is also It can be used to perform the automatic correlation calculation on the target sampling signal of the frame temporarily stored in the storage element and a plurality of target sampling signals of the frame with different delays, so as to obtain a set of target characteristic values; The similarity measurement unit can perform similarity comparison processing according to the set of target feature values and the set of reference feature values, so as to generate a similarity result corresponding to the frame reference sample signal and the frame target sample signal; and The scoring unit can calculate the corresponding similarity results obtained from multi-frame sampling signals to output a final total score. 2. The karaoke scoring device as claimed in claim 1, wherein the predetermined reference audio input includes a reference instrument input and/or a reference vocal input, and the target audio input is the singer's target vocal input through a microphone . 3. The karaoke scoring device as claimed in claim 1, wherein the storage element comprises a first register and a second register, and the karaoke scoring device samples the reference audio input according to a predetermined sampling frequency, and converts it into phase The corresponding multi-frame reference sampling signal is temporarily stored in the first register, and the karaoke scoring device samples the target audio input according to the sampling frequency, and sequentially converts the corresponding multi-frame target sampling signal and temporarily stores it in the in the second register. 4. The karaoke scoring device as claimed in claim 3, wherein the predetermined sampling frequency is approximately 44.1K Hz, and each frame of the reference sampling signal and each frame of the target sampling signal have N sampling points, wherein N=1024. 5. The karaoke scoring device as claimed in claim 4, wherein each frame sampling signal can be expressed as X(k), k=0-N-1, and can be delayed as X(k+τ through different delay times τ ), the autocorrelation calculation takes X(k) and X(k+τ) through the following calculation to obtain the autocorrelation function r _xx (τ). ${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; xx</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; )</span>$ 6. The karaoke scoring device as claimed in claim 5, wherein after obtaining the autocorrelation function r _xx (τ) with respect to the frame reference sampling signal, a selection criterion will be selected according to the reference characteristic value. A group of τ values, τ ₀ ~τ _Nr-1 , is used as the reference characteristic value of the group, and the selection criteria of the reference characteristic value are as follows: r _xx (τ)≥r _xx (τ-1), r _xx (τ)≥r _xx (τ+1) r _xx (τ)≥α(MAX(r _xx (τ))-MIN(r _xx (τ)))+MIN(r _xx (τ)) τ _lowerbound <τ≤τ _upperbound where α is a predetermined constant, MAX(r _xx (τ)) is the maximum value of the function value in the autocorrelation function r _xx (τ) when the value of τ is not 0, and MIN(r _xx (τ)) is The minimum value of the function value in the autocorrelation function r _xx (τ) when the τ value is not 0, and τ _lowerbound is the predetermined lower bound of the τ value, and τ _upperbound is the predetermined upper bound of the τ value. 7. The karaoke scoring device as claimed in claim 6, wherein the reference characteristic value selection standard can select the maximum value of the function value in the auto-correlation function r _xx (τ) when the τ value is not 0 The first three, that is, N _r =3, and the range of the value of τ is between 49 and 441, where N _r is the number of reference characteristic values. 8. karaoke scoring device as claimed in claim 5, wherein after obtaining this auto-correlation function r _xx (τ) with respect to this frame target sampled signal, also can select a standard to select one according to target characteristic value A set of τ values, τ ₀ ˜τ _Nm-1 , are used as the set of target feature values, where N _m is the number of target feature values. 9. The karaoke scoring device as claimed in claim 8, wherein the target feature value selection criterion is to select the maximum value of the function value in the autocorrelation function r _xx (τ) when the τ value is not 0 , that is, N _m =1. 10. The karaoke scoring device as claimed in claim 5, wherein the similarity comparison process subtracts the group of target feature values and the group of reference feature values, and when any one of the absolute values of the subtraction results is less than a given threshold, the similarity result is a hit, otherwise it is a miss. 11. The karaoke scoring device as claimed in claim 10 , wherein the reference audio input and the target audio input include audio with a plurality of different pitches, each pitch has a corresponding central frequency, and corresponds to at least one τ The value of τ in each group of reference characteristic values has a corresponding threshold (TH _τ ), which is obtained by the following formula: ${\mathrm{<span\; class="notranslate">\; TH</span>}}_{\mathrm{<span\; class="notranslate">\; \&tau;</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{{<span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; |</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; FS</span>}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; FC</span>}}_{\mathrm{<span\; class="notranslate">\; upper</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; FC</span>}<span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; |</span>$ Wherein, FS represents a predetermined sampling frequency, FC represents the central frequency of the pitch corresponding to the τ value, and FC _upper and FC _lower represent the central frequencies of the two most adjacent upper and lower pitches of the pitch corresponding to the τ value. 12. The karaoke scoring device as claimed in claim 10, wherein the scoring unit includes a hit counting module and a mistake counting module, and the hit counting module can accumulate counts for hitters according to the similarity result, and output a hit count value, representing HitCount, the error counting module will accumulatively count the mistakes based on the similarity results, and output the error count value, expressed as MissCount, the final total score is between the predetermined maximum score Score _Max and the predetermined minimum score Score _Min , and is calculated by the following formula: $\mathrm{<span\; class="notranslate">\; FinalScore</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; Score</span>}}_{\mathrm{<span\; class="notranslate">\; MAX</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Scor</span>}{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; Min</span>}}<span\; class="notranslate">\; )</span>\frac{\mathrm{<span\; class="notranslate">\; HitCount</span>}}{\mathrm{<span\; class="notranslate">\; MissCount</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; HitCount</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; Score</span>}}_{\mathrm{<span\; class="notranslate">\; Min</span>}}$ 13. The karaoke scoring device as claimed in claim 12, wherein when the similarity result is a hit, the hit count module will add the hit increment value to the current hit count value, expressed as HitIncrease, and become an updated hit count value, and reset the miss count value to a default value, and when the similarity results are all hits consecutively, the hit increment value will also increase accordingly. 14. The karaoke scoring device as claimed in claim 12, wherein when the similarity result is an error, the error counting module will add the error increment value to the current error count value, which is represented as MissIncrease, and becomes an updated error count value, and reset the hit count value to a default value, and when the similarity results are consecutively false, the error increment value will also increase accordingly. 15. The karaoke scoring device as claimed in claim 5, wherein the reference audio input and the target audio input include audio with a plurality of different pitches, each pitch has a corresponding central frequency, and has a predetermined frequency range The similarity comparison process is to find out whether the frequency corresponding to each set of reference feature values and the set of target feature values falls within a predetermined frequency range of the same pitch, so as to generate the similarity result.

Images