CN101485212A - Method and system for wireless real-time transmission of multi-channel audio or video data - Google Patents

Abstract

Methods and systems for streaming multi-channel digital synchronization data are described in the present application. The method is used for streaming multi-channel digital isochronous data, e.g. audio data, in a standard for local area network transmission system where bandwidth is reserved for contention-based and contention-free traffic, and audio data (10) formed from samples (9) are organized in audio frames (174) and transmitted to receivers (6) within successive beacon intervals (137) using multicast. According to the invention, the contention-free traffic (138) of the beacon interval (137) is adjusted to an optimal value, and the length of the beacon interval (137) is adjusted such that the required amount of audio data (9) can be transmitted to the receiver (6) with minimal system delay.

Description

Method and system for wireless real-time transmission of multi-channel audio or video data

technical field

The invention relates to a method for wireless transmission mainly for audio purposes according to the preamble of claim 1 .

The invention also relates to a system for wireless transmission for audio purposes according to the preamble of claim 7 .

The present invention relates to error control methods and systems and synchronization methods and systems.

Typically, the subject of the present invention is a system and associated apparatus and method for synchronized, electromagnetic interference resistant, wireless delivery of the highest studio-quality multi-channel digital audio. This same approach can also be used as the basis for multicast transmission of other digital information with the same type of real-time and bandwidth requirements, such as live digital video and television signal distribution over local areas.

Background technique

introduce

Under currently known technology, studio quality multi-channel digital audio signals are first converted to analog form and then passed to speakers with cables per channel. Furthermore, digital transmission methods in cables or optical fibers are known. If the speakers are passive, additional amplifiers are required for each speaker to amplify the signal power at the transmitter and feed enough power to the transmission cable to drive the speakers at a suitable sound level. All these manipulations, as well as the physically analog transmission path itself, inject several degrading effects such as noise, interference, distortion, group delay, amplitude and phase errors into the quality of the original signal. Cables are often unwieldy and look messy. With careful design of the cables and their wiring arrangements, these effects can be limited to some extent, but seldom completely overcome. The number and bulk of cables required for careful design and tedious installation work add to the expense as well as the skill and time required. Cables and their electromechanical connectors are also prone to mechanical failures, which are difficult to find and repair. These problems are especially detrimental in public performances when the performers and often even the audience move in the cable. In these situations, there can be a real risk of injury and injury from the cables. These problems, impacts and costs are multiplied by the frequent installation and removal of audio installations to and from changing environments during artistic touring.

Using active speakers (with integrated and optimized amplifiers inside them) simplifies the situation to some extent. Analog signals can now have lower power levels and can easily employ differential signaling which is more resistant to noise and interference. However, the generation of multi-channel differential signals requires relatively expensive high quality analog electronics plus expensive differential cables and connectors.

Typically, currently available wireless audio distribution systems are non-standard radio or infrared solutions that use lossy audio compression methods, resulting in compromised performance. They are therefore mainly used for support purposes such as rear loudspeakers in home theaters.

Contents of the invention

The purpose of the present invention is to solve the problems associated with the above-mentioned technologies and related to the synchronized real-time transmission of the highest studio quality streaming digital audio data by constructing a novel international standard compliant with wireless local area network (WLAN) based digital Communications systems, transmitters, receivers plus necessary firmware and software for efficient restricted area distribution of digital audio and testing, configuration, management and control of such systems.

The invention is based on the idea of using multicast in a studio-quality digital format via electromagnetic radio waves or via power line wiring, typically using internationally standardized and mass-produced Wireless Local Area Network (WLAN) components without dedicated signal cables , to pass digital information from a central transmitter station to individual loudspeakers. The digital signal is converted into analog form only at the active loudspeaker unit itself, and finally fed locally to optimized electronic low-pass, band-pass and high-pass filters, amplifiers and finally to high-quality loudspeaker elements. This guarantees ultimate sound quality. Due to the application of mass-produced WLAN technology and its commercial components and a very small number of additional standard integrated circuits, the development effort and the cost of the actual system can be kept to a very reasonable level.

The approach presented here replaces wired lines with standard commercial wireless local area network technology as specified in the IEEE 802.11 series of standards. The special features required for uncompressed real-time delivery of multi-channel studio-quality audio have been achieved through innovative choices of WLAN system coordination functions, communication methods and control parameters together with special upper-level firmware.

According to a preferred embodiment of the present invention, audio data formed from samples is organized in audio frames and sent to receivers within consecutive beacon intervals using multicast messaging. According to the WLAN standard, two coexisting modes of operation are possible. The usual mode is called contention-based service, which is widely used in commercial data communication products. Another mode is called contention-free service, rarely used but precisely specified in the IEEE 802.11 standard, which is the basis of the present invention. Beacon frames are used to control switching between these two modes of operation. The length of the beacon interval is a programmable parameter, and in the present invention it is adjusted so that an optimal amount of isochronous audio data can be sent to the receiver with minimal coordination delay. In a preferred embodiment of the invention, the optimal amount is the amount required for high quality audio reproduction of the different audio data.

According to another preferred embodiment of the present invention, an enhanced error control system optimized for different audio delivery is used, which minimizes or completely eliminates the need for retransmissions, in which system the received The signal contains error control data blocks and is used for error correction purposes.

According to a third preferred embodiment of the invention the receivers are synchronized at the exact same instance within each beacon interval in a multicast system with the help of an end-of-frame interrupt generated by a control frame terminating each beacon interval ,

In particular, the method according to the invention is characterized by what is stated in the characterizing part of claim 1 .

Furthermore, the system according to the invention is characterized by what is stated in the characterizing part of claim 7 .

Significant benefits can be obtained with the help of the present invention. With per-speaker reception of studio-quality digital audio, all error factors associated with traditional speaker signal paths can be avoided. According to one embodiment of the present invention, performing digital-to-analog conversion at the active loudspeaker itself maximizes sound quality by localizing the propagation path of the analog signal over a fixed and optimized active loudspeaker circuit.

Signal cables, their connectors and differential signal transmitters/receivers as well as the associated material and installation work can be completely avoided. This eliminates all the expense, failure and installation issues associated with them. Since the mass-produced standard WLAN technology is the basis of the invention, its production costs can be very low according to an embodiment of the invention.

According to an embodiment of the present invention, since the multicast transmission mode and frequency multicast synchronization are used, phase errors between channels can be effectively eliminated.

According to one embodiment of the present invention, system-level delays can be minimized to insignificant levels due to the use of optimized transport frame sizes.

With the help of an error control method according to an embodiment of the present invention, a simple and fast best-effort error correction scheme can be obtained.

Description of drawings

In the following, the invention will be described in more detail with reference to exemplary embodiments illustrated in the accompanying drawings, wherein

Fig. 1 shows a block diagram of a general system configuration of the present invention.

Fig. 2 shows a block diagram of an example transmitter base station according to the present invention.

Fig. 3 shows a block diagram of another example transmitter base station according to the present invention.

Figure 4 shows a block diagram of an example receiver according to the present invention.

Fig. 5 shows an audio data structure representing a multi-channel audio sample according to the present invention.

Figure 6 shows a data structure representing an 8-tuple of audio samples with additional error control modules according to the present invention.

Fig. 7 shows an error correction algorithm according to the invention with the help of the data structure of Fig. 6 .

Figure 8 shows a block diagram of a Medium Access Control (MAC) architecture that can be used by the present invention.

Figure 9 shows a general MAC frame structure that can be used by the present invention.

FIG. 10 shows the data structure of the WLAN frame control field that can be used by the present invention.

Figure 11 shows a block diagram of a possible Media Access Control (MAC) address, multicast version that can be used by the present invention.

Figure 12 shows the data structure of a generic beacon frame that can be used by the present invention.

FIG. 13 shows the data structure of a beacon frame according to the present invention.

Fig. 14 shows the data structure of capacity information that can be used by the present invention.

Figure 15 shows the data structure of information elements that can be used by the present invention.

Fig. 16 shows the data structure of the traffic indication table (Traffic IndicationMap, TIM) element format that can be used by the present invention.

FIG. 17 shows a data structure of an Extended Rate PHY (Extended RatePHY, ERP) information element that can be used by the present invention.

Fig. 18 shows the data structure of the extended supported rates element (supported rates element) that can be used by the present invention.

Fig. 19 shows the data structure of a Contention-Free (CF) parameter setting element that can be used by the present invention.

FIG. 20 shows the data structure of a CF-End frame that can be used by the present invention.

Figure 21 shows the data structure of the ERP-OFDM PHY frame structure that can be used by the present invention.

Figure 22 shows a graph of the bandwidth requirements of the present invention.

Figure 23 shows a table according to the invention of the number of 8x32 bit sample records in consecutive data blocks, with respect to proper sequencing of digital audio for transmission.

Figure 24 shows a graph of dithering behavior according to the present invention.

Figure 25 shows a block diagram of a general data structure related to worst case transmission timing according to the present invention.

Fig. 26 shows a flowchart of audio input processing according to the present invention.

Herein, the following terms will be used in connection with the present invention.

1 WLAN, wireless local area network

2 audio sources

3 digital audio transmitter

4 base stations

5 remote controller

6 receivers

7 speakers, typically active speakers with built-in power amplifiers

8 Sound (eg, surround) systems

9 Audio data format/sampling

10 Transport-level audio data format

11 error correction code

12 MAC sublayer, MAC = media access control

13 Distributed coordination function

14 point coordination function

15 Competition-free communication services

16 Contention-based communication services

17 General MAC frame structure

18 frame control

19 Duration/ID

20 Address 1

21 Address 2

22 address 3

23 sequence control

24 Address 4

25 Frame body

26 FCS, Frame Control Sequence (Frame Control Sequence)

27 MAC head

28 MAC frames

29 WLAN frame control field

30 protocol version

31 types

32 subtypes

33 to DS, DS = distributed system

34 from DS

35 More Fragments (more Frag)

36 Retry

37 More data

38 Pwr Mgt

39 WEP, Wired Equivalent Privacy (Wired Equivalent Privacy)

40 Order (order)

41 Separate address

42 group address

43 Unicast address

44 Multicast address

45 broadcast address

46 class beacon frame (generic beacon frame)

47 frame control

48 Duration

49 destination address

50 source address

51 BSS ID

52 sequence control

53 frame body

54 FCS

55 Timestamp

56 beacon interval

57 Capacity information (Capability info)

58 SSID, Service Set Identity (Service Set Identity)

59 optional fields

60 Beacon frame used in the present invention

61 frame control

62 duration

63 destination address

64 source address

65 BSSID, basic service setting identification

66 sequence control

67 frame body

68 FCS

69 Timestamp

70 beacon interval

71 capacity information

72 SSID

73 CF parameter set

74 TIM, Traffic Indication Map (Traffic Indication Map)

75 ERP, extended rate PHY (Extended Rate PHY)

76 Extended rate

77 element format

78 Element ID

79 length

80 information

81 TIM elements

82 Element ID

83 length

84 DTM count, DTM = delivery traffic indication map (Delivery Traffic IndicationMap)

85 DTIM cycles

86 bitmap control (bitmap control)

87 partial virtual bitmap

89 ERP Information Elements

90 Element ID

91 length

92 Non ERP-present (Non ERP-present)

93 User Protection

94 Barker Preamble mode (Barker Preamble mode)

95 r3-r7

96 Extended Supported Rate Element Formats

97 Element ID

98 length

99 extended supported rates

100 The CF parameter sets the element format, CF=no competition

101 Element ID

102 length

103 CFP count

104 CFP, contention-free cycles

105 CFP max duration

106 CFP Dur Remaining

107 CF-End frame

108 MAC-head

109 CF-end MAC frame

110 frame control

111 Duration

112 RA, receiver address

113 BSSID

114 FCS

115 ERP-OFDM PHY frame structure

OFDM = Orthogonal Frequency Division Multiplexing

116 coded/OFDM

117 PSDU, Protocol Service Data Unit

118 PLCP preamble

119 SINGAL

120 speed

121 Reserved

122 LENGTH

123 tails

124 parity

125 services

127 frame control

128 Duration/ID

129 Address 1

130 Address 2

131 address 3

132 sequence control

133 Address 4

134 frame body

135 FCS

136 domain of interest

137 WLAN repetition period/beacon interval (N*TU)

138 Abbreviated (Foreshortened) no contention cycle

139 Multiplexers and Receivers

140 serial to parallel converter

141 buffer

142 USB host controller, USB = common serial bus

143 USB input

144 S/PDIF input, S/PDIF=Sony/Philips digital interface

145 analog inputs

146 Analog buffers and multiplexers

147 A/D converter, A/D = analog to digital

148 MAC/Baseband Processors

149 Microcontrollers

150 D/A converters and filters, D/A = digital to analog

151 Select analog input

152 A/D conversion

153 Select digital input

154 24-bit reformatting

155 Select audio input

156 Channels 8

157 No

158 yes

159 Produce FEC and write to the buffer, FEC = Forward Error Correction (Forward ErrorCorrection)

160 Calculate lost channels

161 Generate FEC and write to the buffer

162 samples i-1

163 samples i

164 samples i+1

165 Corrected samples i

166 ESS, Extended Service Set

167 IBSS

168 CP pollable

169 CF polling request

170 Confidential

171 Reserved

172 antenna

173 most significant bits

174 audio MAC frames

175 Control MAC frame

Detailed ways

system

According to FIG. 1 , the system comprises one or several audio sources 2 , which may be digital or analog sources, said audio sources 2 also comprising the possibility of a direct computer connection, for example via a USB interface. The source 2 is connected to an audio transmitter 3 which is further connected to a base station 4 comprising antenna means 172 for wireless transmission. Typically, the transmitter 3 and base station 4 are controlled by a remote controller 5 or computer. The signal from the base station 4 is sent via the WLAN network 1 using an isochronous multicast message to a receiver 6 of a multi-channel (eg surround) sound system 8 comprising several loudspeakers 7 . In other words, audio data from source 2 is converted to digital data by devices 3 and 4 and passed to the speaker receiver as standard WLAN digital data.

transmitter base station

FIG. 2 shows a simple example version of the transmitter 4 that does not include an analog input, and FIG. 3 shows the device with an analog input 145 . Typically, the transmitter base station 4 is a 108 Mbit/s extended IEEE 802.11g WLAN MIMO access point station, which receives a specified number of monaural, stereo or multi-channel analog and AES3, S/PDIF or USB digital audio signals. 108 Mbit/s is practically the lowest possible standard bit rate for the system of the present invention. In the future, higher WLAN transmission speeds are expected. They will make it possible to use retransmissions to improve error correction methods. The analog input 145 is electronically buffered 146 and fed to an analog selector capable of selecting stereo 24 One of the stereo inputs of the bit. DSP processing makes it possible to perform any type of matrixing to generate derived channel signals from right and left channel stereo signals. Alternatively, if a digital AES3 or S/PDIF input 144 is selected instead, it is converted to a parallel format with a shift register 140 and then fed to a 32-bit digital signal processor 4 via a memory buffer 141, the digital signal Processor 4 processes the signal into an 8-channel 24-bit 192kSample/s parallel format used in WLAN transmission. The use of the USB2.0 interface 143 allows the direct input of an 8-channel 24-bit 192kSample/s audio stream which is then stored directly into a parallel 32-bit memory format. For example a DSP with a precision extension from a possibly shorter sample size to 24 bits can be used together with a four-sample third-order polynomial curve fitting method from a lower sample rate to 192kSample/s resampling. After these processing steps, a forward error correction code octet 11 (Figs. 6 and 7) is prepended to each 24-bit audio sample 9. The resulting 32-bit samples are buffered into memory for transmission in the form of a record representing 8 samples 9 of simultaneous samples from the eight audio channels.

Within the transmitter station 3 there is a 16KB memory ring buffer 141 or FIFO buffer for immediately storing incoming data waiting to be transmitted. After initialization, the sender station 3 uses a contention-free beacon interval setting of one TU with a CFPMaxDuration parameter setting of 512 μs. This setting is established when the incoming audio stream ends and minimizes the reaction time to the incoming audio stream, thus also sizing the memory buffer. Alternatively, it is possible to stuff the audio stream by filling the buffer with all zero data when there is no incoming audio. When audio data streaming starts and there is no more than 6.6KB of space available in the ring or FIFO buffer for incoming data, the contention-free interval is reprogrammed to 3 TUs with the CFPMaxDuration parameter set to 2782 μs. WLAN digital audio transmission also starts at this point. Depending on WLAN traffic conditions, it may take 459 to 1483 μs for the first audio data to be transmitted, during which time from 2.8KB to 9.2KB more audio data has entered the buffer. Therefore, depending on the timing relationship between the audio source and the WLAN, the input buffer size requirement limit is from 9.4KB to 15.8KB. For this, a 16KB ring buffer 141 or FIFO is sufficient for all cases. The buffer 141 is filled with incoming audio data through the signal processor and is transferred from the buffer 141 with the help of a timer controlled DMA controller or another processor maintaining the correct WLAN data transfer rate. There is a high speed parallel or serial standard interface between the processor and the WLAN chip, the software application program interface (API) of the WLAN chip is used for flow control as described above. The maximum delay caused by transmission is approximately 1500 μs. In addition, there is a receive buffer delay of approximately 500 μs. So the total delay is less than 2000μs. This corresponds to the propagation of sound of less than 1 meter, making this delay negligible for the human hearing experience.

The WLAN portion of the base station (and receiver) complies with the IEEE 802.11g standard with range and transmission rate extensions proposed by Atheros Inc. and Airgo Inc. A MIMO antenna arrangement 172 is also typically used. The normal bit rate is 108Mbit/s. These implementations of the extended IEEE 802.11g WLAN also incorporate a robust transmission error correction mechanism that efficiently distributes eventual transmission path burst errors to single-bit reception errors at the receive and can be corrected at the octet level All of them. This feature is exploited in the specified application layer forward error correction method. Contention-based, individually addressed messaging between the base station 4 and receiver stations is used for configuration, status monitoring and control of the receivers and audio equipment attached to them. In the transmitter base station 4 there is a handheld remote controller receiver for infrared, a USB 2.0 universal receiver/transmitter for Bluetooth and WLAN handheld remote controller adapters and a USB 2.0 computer communication receiver/transmitter.

System configuration, monitoring and control

System configuration, monitoring and control is performed via the handheld remote controller(s) or via the (personal) computer application(s), as described above.

receiver

According to Figure 4, the receiver 6 typically consists of a MIMO antenna subsystem 172 and IEEE 802g conformant WLAN circuitry with Atheros or Airgo range and transfer rate extensions. For configuration, status monitoring and control operations there are typically software controlled multi-color LEDs to aid in the identification and status of individual speakers 7 . The WLAN works at a nominal speed of 108Mbit/s. The received audio data stream is buffered into a 16KB input ring or FIFO memory buffer, and audio output from the buffer begins when the buffer reaches the 6.6KB limit. Data for the receiver's currently selected channel is selected by the 32-bit processor 149 and fed to the 24-bit D/A converter 150, followed by high quality volume control circuitry and buffer amplifiers for that particular audio reception select audio channels of the amplifier unit to generate studio-quality analog audio signals. Channel selection is performed by configuration and software control on competing communication services of the WLAN.

The eight-channel receivers 6 each have an internal crystal-derived clock used to generate a 192000 Sample/s clock. These clocks are restarted by an end-of-frame interrupt generated by a CF-End control message every 3072 μs receive slot. The differential analog output signal from the D/A converter is the differential input for the configured channel's active loudspeaker. In addition, full eight channels of digital audio are available on standard AES3 (S/PDIF) coaxial and optical outputs. This facilitates the implementation of a wireless studio-quality 8-channel digital audio link. Another pair of AES3 (S/PDIF) coaxial and optical outputs is provided for powered speakers with digital inputs.

remote control terminal

There are two approaches to the control of the system, the battery powered handheld control terminal 5 and the software application available for multiple platforms including Linux, MS Windows, Apple and Symbian operating systems.

Handheld Remote Controller

The handheld remote controller 5 includes a keypad, a small display, a processor and a communication link to a base station. The keypad function allows selection of audio source 2, audio speaker bank 8 and individual speaker 7 configuration and control. Speaker groups 8 as well as individual speakers 7 can be smoothly activated and deactivated, and their general and individual volume levels can be adjusted. The handheld remote controller communicates with the transmitter base station 4 via an infrared, Bluetooth or WLAN link. The base station 4 takes over the control of the loudspeakers by using contention mode communication and group addressing or individually addressed individual receivers. In the remote controller 5 there is a panic key and function which causes smooth immediate muting of all speakers 7.

remote control software

The system described above can be fully controlled by a computer running configuration, monitoring and control application software. The transmitter base station transmits commands and responses using Bluetooth, IrDA, LAN, WLAN or USB2.0 links.

method

According to Figures 5 and 6, for each individual sample, the device of the present invention transmits in real time and synchronously 24-bit 192000Sample/s digital audio streams of 8 completely independent and uniform normalized uncompressed channels together with high-efficiency sound transmission specific 8 bit Forward Error Correction (FEC) code 11. In the remainder of this document, the eight discrete 24-bit sample groups 10 and their associated 8-bit FEC octets 11 (for a total of 32 octets or 256 bits) will be referred to as the transport-grade audio data format. The supported application-level digital audio data bandwidth requirement is therefore 36.864Mbit/s. Together with the FEC octet (octet) 11, the application data bandwidth requirement becomes 49.152 Mbit/s. In addition, there is overhead caused by the effects of MAC framing, WEP encapsulation, PHY framing, and competing traffic time allocations. These are such that even a WLAN bit rate of 54 Mbit/s is insufficient for this application. With today's standard WLAN technology, the required performance cannot be achieved. The novel transmission method described below is based on the innovative use of contention-free multicast transmission with Point Coordination Function (PCF), as specified in the IEEE 802.11 standard. Through careful parameter tuning, the bandwidth optimization of WLAN can be divided between the PCF contention-free media access mode and the usual decentralized control function (Decentralized Control Function, DCF) contention access mode, making synchronous multi-channel digital audio transmission possible. . Under the 108Mbit/s extension of IEEE 802.11gWLAN network, through the use of ERP-OFDM PHY layer framing, the target eight (8) independent 24+8-bit, 192kSample/s digital audio streams are transmitted synchronously together with normal contention-based WLAN data Flow is possible. High channel counts, high resolution and high sampling rates guarantee wireless distribution of the best sound quality commercially available today.

data structure

According to Fig. 25, the purpose of the present invention is to deliver enough audio blocks (transport grade audio data format) 10 in order to produce high quality audio sound. First, the beacon interval 137, defined by software settings, must be chosen correctly for this purpose. Beacon signals defining the length of the beacon interval 137 are sent at intervals defined by integers in the IEEE 802.11g WLAN standard. The value of the integer may have a value from 1-N. In other words, the beacon interval 137 is the product of the beacon integer and the time unit (TU). In the IEEE 802.11g WLAN standard, the length of a TU is 1024 μs, so the beacon interval 137 is a multiple of TU (1024 μs). However, the standard defines that in each beacon interval 137, enough time should be reserved for competing traffic, more precisely and enough time for maximum size frames, ACKs, 2 slot times, and . According to the present invention, for the beacon interval 137, the optimal value of the number of discovery time units TU is three. In a preferred embodiment of the invention the optimal amount can also be defined as a sufficient amount. This gives enough time to send 12 audio MAC frames 174 and 2 control MAC frames 175 within one beacon interval 137 . Each audio MAC frame 174 includes 49-51 transport level audio data format blocks 10, the number of which is defined according to the table of FIG. In this figure, one row represents the content of the audio MAC frame 174 in one contention-free period 138 of the beacon interval 137 . As can be seen from Figure 23, after each 125 beacon interval, the predetermined sequence is repeated. With the help of this detailed sequence, the average streaming rates of the audio input and WLAN output match, so jitter can be kept to a minimum, as shown in Figure 24. This also results in a minimal requirement for buffer memory in both the transmitter and receiver 6 .

bandwidth division

According to FIG. 25, in order to guarantee timely delivery of audio data, the highest possible repetition rate of the contention-free period 138 must be achieved. At the same time, the maximum part of the network capacity must be reserved for audio traffic. The IEEE 802.11 standard requires that sufficient contention traffic time must exist within each repeated contention-free interval for the transmission of a maximum size data frame together with its acknowledgment frame plus two SIFS periods and two slot times. Under 108MBbit/s bit rate and ERP-OFDM PHY framing, this requirement is equal to 212+40+2×10+2×9=290μs. As described in the IEEE 802.11 standard, at the beginning of the contention-free period 138, contention traffic can shorten the contention period by a maximum of RTS control frame, CTS control frame, a maximum size data frame, ACK control frame plus four SIFS value. Under 108Mbit/s bit rate and ERP-OFDM PHY framing, this requirement is equal to 40+40+212+40+4×10=372μs. The contention-free period begins with a beacon frame 67 (FIG. 13), followed by SIFS. Under 108Mbit/s bit rate and ERP-OFDM PHY framing, this requirement is equal to 76+10=86μs. The contention-free period ends with CF-End frame 109 (FIG. 20). At 108 Mbit/s bit rate and ERP-OFDM PHY framing, this requirement equals 40 μs. For contention-free data traffic, the remainder of the contention-free repetition interval is available. Since the granularity of the contention-free interval is a time unit (TU) of 1024 μs, when the contention-free interval is set to one TU, the available time for contention-free traffic is 1024-290-372-86-40=236 μs. Considering the maximum data frame size stipulated by IEEE 802.11, PIFS, PHY encapsulation overhead, WEP and MAC between subsequent data frames, only the maximum value of 72 samples of 8-channel 24+8 bits, 192kSample/s data can delivered within that time. Real time requires slightly more than 196 samples during each TU due to the audio sampling rate. So this is not a working solution. With the contention-free interval set to two TUs, the time available for contention-free data becomes 1260 μs, which allows transmission of up to five large-sized 72-sample frames in 1155 μs plus an additional 27-sample frame in 104 μs . At a minimum real-time requirement slightly greater than 2x196=392 samples, a total of 387 samples in two TUs. So the capacity is only slightly insufficient for real-time audio delivery. Therefore, it is evident that the minimum and most desirable contention-free repetition interval is 3 TUs, 3 x 1024 = 3072 μs.

In order to optimize the smooth flow of data and to minimize buffering needs, the average rate of samples per TU should be kept as close as possible to 1024/1000×192=196.61 by varying the size of the data frame in an appropriate manner according to Figure 23. Introducing the frame size algorithm below, which is one of the key innovations in this invention. The contention-free time is first split into 12 variable-sized buffers. These are presented as columns in Figure 23. The buffer size varies between 49 samples and 51 sample records, each sample record is set according to the sizing rule described below. The first, second, fourth, fifth and sixth buffers of the adjacent buffer set of 12 buffers each contain 51 samples, while the first buffer of the third set contains 50 sample records. All other buffers contain 49 samples. This sequence is repeated at every seventh buffer set up to the 125th buffer set, after which the entire cycle is repeated from the beginning. One exception to the above rule is that the first buffer of the 60th buffer set contains 50 samples instead of 51 as stated by the internal repetition rules. The algorithm according to Figure 24 ensures that the buffer jitter stays below +/- one sample within all buffer sets and goes to zero at the end of every 125th sample buffer set. Under this adjustment algorithm, there is a worst-case margin of 180 μs within the contention-free data delivery time. This time is sufficient for one additional 1632 data octet frame or two 450 octet user data transfers, which can be used for additional real-time control purposes as desired. The device also makes it possible to support efficient user contention traffic up to 12Mbit/s together with real-time audio transmission. This competing traffic is available for system configuration and control and other independent data exchanges.

As shown above, at least three TUs need to be selected for the duration of the beacon repetition interval to reserve enough bandwidth for contention-free isochronous audio traffic. Choosing a minimum of three TUs also minimizes system latency and buffering requirements. Furthermore, choosing a minimum value of three TUs ensures maximum bandwidth for contention-free based traffic, except for contention-free isochronous audio traffic.

In this application, the combination of these three limiting factors is called optimal.

error control

According to Fig. 7, the error control method is optimized for simplicity and speed under the assumption of a human listening to multi-channel studio quality speech and music audio sounds. This implies a simple and fast best-effort error correction scheme that either completely corrects the remaining application-level transmission errors of the erroneous frame or reduces their audible effects to an unnoticeable level. This method utilizes the inherent properties of long 24-bit audio data samples and high 192kSample/s sampling rate and extended IEEE802.11g implementation to transform burst errors originating in the transmission path into single bit errors in reception. However, this error correction scheme is not suitable for applications that cannot tolerate errors.

According to FIG. 7, an eight-bit forward error control (FEC) octet 11 is appended to the front of the most significant octet 173 of each sample 162-165 prior to the transmission stage. The FEC octet 11 is formed by duplicating the most significant octet 173 of the data sample. Due to the WLAN transmission error correction method, almost all residual reception errors are single bit errors. It is therefore sufficient to correct the effects of single bit errors. This allows the received most significant data octet to be compared with the corresponding received FEC octet 11, and if they differ the most significant octet of the first 162 and last 164 error-free samples is used The group average is used to replace the most significant sample octet 173. This effectively removes almost all remaining errors in the most significant data octet. Since the lower order bits represent only less than 4% of the dynamic range of the audio, their significance is negligible. At the high sampling rates used, residual errors are inaudible to the human ear.

It is worth remembering that multicast mode transmission simultaneously delivers all eight channels of audio data to all receivers in the designated multicast group simultaneously. This makes it possible to implement any intelligent error correction method that relays on data from adjacent audio channels, if such an approach feels appealing. Furthermore, alternatively this powerful standard 8/24 error correction method can be used for more stringent error correction if special error correction hardware or very high computational power is available on the transmitter and receiver.

Synchronize

According to FIG. 20, synchronization within the system is based on the repeated occurrence of the end-of-frame interrupt generated by the CF-End frame 109 at exactly 2782 μs after the start of each repeated 3072 μs contention-free repetition interval. The end-of-frame interrupt of this control message 109 synchronizes all receivers 6 within the inaccuracy of the difference in interrupt latency of the receivers. Since all receivers are programmed to wait for that particular interrupt, the system-level synchronization jitter caused by interrupt latency is on the order of one instruction execution cycle (adding very small processor-to-processor crystal oscillator phase jitter ). In practice, this total jitter is of the order of 100ns and may not be noticed by a human listener. For comparison, the 192kSample/s audio sample cycle is 5.21μs.

Detailed description of the WLAN transmission cycle

According to Figure 25, in the idle state, when no audio signal is present, the transmitter is programmed to run a beacon interval of one time unit (1 TU). In this way, the system responds to a delay in the start of a new incoming audio stream, thus minimizing the size of the transmitter buffer. When the start of audio streaming is detected, the transmitter transitions to the active transmission state. During the active transmission phase, the beacon interval is reprogrammed to three time units and each WLAN audio transmission slot is thus programmed to last three time units (TU) (1024 μs each), for a total of 3072 μs. A Point Coordination Function (PCF) is implemented in the transmitter of the WLAN access point station. The beacon repetition interval, and therefore the contention-free repetition interval, is set to three time units and each such period contains a contention-free and contention part. The length of the allocated contention-free period is set to 2782 μs using the CFPMaxDuration parameter in the beacon frame 67, and this setting leaves a guaranteed 290 μs for the decentralized control function (DCF). This time is large enough to transmit a maximum length data frame along with its acknowledgment and associated interframe elements during the contention period, as required by the IEEE 802.11 standard. It also means that a minimum bandwidth of 6.16Mbit/s (when using maximum size data frames) is always available for competing traffic. Under heavy traffic of large frames, the allocated contention free period is abbreviated from when frames are being transmitted during the expected start of the contention free period. Since this contention exchange can include RTS, CTS and ACK control frames and their associated inter-frame elements in addition to maximum size data frames, a busy medium can consume up to a maximum of 372 μs from the contention-free period onwards.

The worst-case transfer timing scenarios for audio data are as follows. The expected contention period begins to occur, but the maximum length contention transmission sequence has just begun. This will result in an abridgement of the 372μs contention-free period. Only after this abbreviated delay, the 40 μs beacon message setting the NAV condition can be transmitted. The first audio data block transmission starts after another 10 μs SIFS time has elapsed. This is a total of 459 μs after the expected start of the contention free period. In the case of minor abridgements, a quiet filler period is inserted by the transmitter software to achieve a 459 μs tick. This arrangement ensures that the first audio bit is always sent on the same relative tick of the 3072 μs contention-free repetition interval. The available delivery time for contention-free audio data is thus 3072-458-290-40-10 = 2274 μs. In the worst case scenario, the first audio buffer contains fifty one 24+8 bit sample records. The next eleven audio data blocks each contain 49 sample records. When the MAC and ERP-OFDM PHY framing domains are included, the corresponding frame transfer times for 49, 50 and 51 sample records become 156 μs, 160 μs and 160 μs, respectively. In the largest size 12-tuple, there are 51+11×49 audio records. Twelve blocks are transferred end-to-end between them with 19 μs PIFS cycles. The worst case total time from the first audio bit to the last is thus 160+11*156+11*19=2085 μs. Effective time compression on WLAN is thus 2085/3072 = 68 percent. This twelfth audio block is again followed by a 19 μs PIFS period and followed by a programmed idle delay of 180 μs after which the 40 μs CF-End broadcast frame 109 will terminate the contention-free period and also reset the initial pass start beacon frame And set the NAV condition. This happens at exactly the same time as the end of the contention-free period according to the timer set via CFPMaxDuration. A time margin of 180 μs for a contention free period out of the minimum achievable time of 2284 μs represents only an 8 percent contention free time margin. At this point, the contention period begins to allow the transmission of a single maximum size frame and ACK response plus the associated two interframe SIFS times and two slot times, as specified in the IEEE 802.11 standard.

Transmitter and Base Operation

According to Fig. 26, in phase 151 or 153 the system selects between n analog input 145 or mAES (S/PDIF) digital input 144/143. If the analog input is selected, the buffered analog stereo signal is then converted in stage 152 to 2x24 bit form by a 192 kS/s dual A/D converter. If digital input is selected, then in stage 154 the incoming digital stream is also reformatted to 24 bit 192 kS/s format. In stage 156 the number of audio channels is checked, if some channels are missing 157, the missing channels are counted in step 160 and FEC is generated in step 161 and written to the buffer with the audio data. If the number of lanes is eight at step 156, the data with FEC through steps 158 and 159 is directed directly into the buffer.

Multicast in this application means a process in which all receivers receive the same data packet from which they extract data suitable for themselves. In other words, all eight speaker receivers receive data for all audio channels, but only extract data for the channels assigned to them for further processing.

The invention is applicable to a variety of isochronous data transmission systems, but as described herein it is particularly suitable for multi-channel audio purposes.

Video solutions are therefore also suitable for some embodiments of the invention.

In addition to the WLAN transmission medium, the invention is also suitable for ultra-wideband radio transmission technology, or HomePlug AV type transmission technology, in which the main power cable is also used for data transmission. In the latter case, the transmission system is not literally cableless, but since active speakers always require an external power feed through the cable, no additional cables are required for data transmission.

Claims (24)

1. A synchronous transmission method for streaming multi-channel digital synchronous data such as audio data in a standard wireless local area network transmission system, wherein bandwidth is reserved for contention-based traffic and contention-free traffic, - said audio data (10) formed by samples (9) are organized in audio frames (174) and sent to receivers (6) using multicast in consecutive beacon intervals (137), It is characterized by: - adjusting the contention-free traffic (138) of said beacon interval (137) to an optimum value such that: o reserve sufficient bandwidth for contention-free isochronous audio traffic, o Minimize system latency and buffering requirements and o Ensure maximum bandwidth for contention-based traffic in addition to contention-free isochronous audio traffic as described. 2. The method according to claim 1, characterized in that the beacon interval (137) is set to 3 time units (TU) in the IEEE 802.11 standard. 3. The method according to claim 1 or 2, characterized in that the number of samples (9, 10) in the audio frame (174) is changed so that the transmitter (3, 4) and receiver (6) Minimize buffer size in . 4. The method of claim 3, wherein the number of samples (9, 10) in the audio frame (174) is varied in a cycle of 125 consecutive beacon intervals (137). 5. The method according to claim 4, characterized in that, according to the table of Fig. 23, the number of samples (9, 10) is in the cycle of 125 consecutive beacon intervals (137) in the audio frame (174) In the table, each row represents the content of a beacon interval (137). 6. A synchronous transmission method for streaming multi-channel digital synchronous data such as audio data in a standard wireless local area network transmission system, wherein bandwidth is reserved for contention-based traffic and contention-free traffic, - said audio data (10) formed by samples (9) are organized in audio frames (174) and sent to receivers (6) using multicast in consecutive beacon intervals (137), It is characterized by: - adjusting the contention-free traffic (138) of said beacon interval (137) to an optimal value, and The length of the beacon interval (137) is adjusted such that the required amount of audio data (9) can be sent to the receiver (6) with minimal system delay. 7. A wireless transmission system for streaming digital serial audio data (9, 10), wherein system bandwidth is reserved for contention and non-contention traffic, said system comprising means for: - organizing audio data (10) formed by samples (9) in audio data frames (174) and control frames (175), and - means for sending said frames (174, 175) to the receiver (6) in successive beacon intervals (137), It is characterized in that the system includes means for: - adjusting the contention-free traffic (138) of said beacon interval (137) to an optimum value such that: o reserve sufficient bandwidth for contention-free isochronous audio traffic, o minimize system latency and buffering requirements, and o Ensure maximum bandwidth for said contention-based traffic in addition to said contention-free isochronous audio traffic. 8. The system according to claim 7, characterized in that it comprises means for setting the beacon interval (137) to 3 time units (TU) in the IEEE 802.11 standard. 9. The system according to claim 7 or 8, characterized in that the system comprises means for changing the number of samples (9, 10) in the audio frame (174) so that the transmitters (3, 4) and means for minimizing the buffer size in the receiver (6). 10. The system according to claim 9, characterized in that the system comprises means for changing the samples (9, 10) in the audio frame (174) in a cycle of 125 consecutive beacon intervals (137) number of devices. 11. The system according to claim 10, characterized in that the system comprises means for changing the samples ( 9, 10), wherein each row in said table represents the contents of a beacon interval (137). 12. A wireless transmission system for streaming digital serial audio data (9, 10), wherein system bandwidth is reserved for contention and non-contention traffic, said system comprising means for: - organizing audio data (10) formed by samples (9) in audio data frames (174) and control frames (175), and - means for sending said frames (174, 175) to the receiver (6) in successive beacon intervals (137), It is characterized in that the system includes means for: - adjusting the contention-free traffic (138) of said beacon interval (137) to an optimal value, and - adjusting the length of the beacon interval (137) such that the required amount of audio data (9, 174) can be sent to the receiver (6) with minimum system delay 13. A method of error correction in a system for streaming digital serial audio data for a real-time solution (6, 7) according to any preceding method or system claim, in said method, - said audio data is divided into data blocks (10) of predetermined length, It is characterized by: - said data block (10) comprises an error control block (11), the content of which block (11) is a copy of the most significant bits (173) of the corresponding audio data (9), - comparing the content of said error control block (11) with the received content of the corresponding content of said audio data (10, 173), - if a difference between said audio data (10, 173) and said error control block (11) is detected, use the average of the correctly received preceding (162) and following (164) audio data (10) to replace the corresponding audio data. 14. A method of error correction in a system for streaming digital serial audio data for a real-time solution (6, 7), in which method, - said audio data is divided into data blocks (10) of predetermined length, It is characterized by: - said data block (10) comprises an error control block (11), the content of which block (11) is a copy of the most significant bits (173) of the corresponding audio data (9), - comparing the content of said error control block (11) with the received content of the corresponding content of said audio data (10, 173), - if a difference between said audio data (10, 173) and said error control block (11) is detected, use the average of the correctly received preceding (162) and following (164) audio data (10) to replace the corresponding audio data. 15. The method according to claim 14, characterized in that said block (10) is 32 bits long. 16. A method according to claim 14 or 15, characterized in that the error control is further enhanced by comparing the data with information having other channels. 17. An error correction system for streaming digital serial audio data for a real-time solution (6, 7), wherein said system comprises means for: - means for dividing said audio data into data blocks (10) of predetermined length, It is characterized in that the device is used for: - inclusion of an error control block (11) into said data block (10), the content of said block (11) being a copy of the most significant bits (173) of the corresponding audio data (9), - comparing the content of said error control block (11) with the received content of the corresponding content of said audio data (10, 173), - if a difference between said audio data (10, 173) and said error control block (11) is detected, use the average of the correctly received preceding (162) and following (164) audio data (10) instead of the corresponding audio data. 18. A system according to claim 17, characterized in that said block (10) is 32 bits long. 19. System according to claim 17 or 18, characterized in that said error control is further enhanced by comparing said data with information having other channels. 20. A method of synchronization in a multicast system according to any preceding method or system claim, which wirelessly streams digital serial audio or video data for real-time solutions (6, 7), in which In the above method: - data is organized in frames (17, 174, 175) comprising control frames (175) and audio or video frames (174), and - sending organized audio or video data (9, 10) to multiple receivers (6) by multicasting in consecutive beacon intervals (137), It is characterized by: - Synchronization of said audio or video data between multiple receivers (6) by an interrupt command included in each beacon interval (137). 21. Synchronization method in a multicast system that wirelessly streams digital serial audio or video data for real-time solutions (6, 7), in said method: - data is organized in frames (17, 174, 175) comprising control frames (175) and audio or video frames (174), and - sending organized audio or video data (9, 10) to multiple receivers (6) by multicasting in consecutive beacon intervals (137), It is characterized by: - Synchronization of said audio or video data between multiple receivers (6) by including in each beacon interval (137) an end-of-frame interrupt generated by a well-timed CF-End control frame. 22. The method according to claim 21, wherein the interrupt command is an end-of-frame interrupt command. 23. A system using a multicast method for wirelessly streaming digital serial audio or video data for real-time solutions (6, 7), said system comprising: - means for organizing said data in beacon intervals (137), said beacon intervals comprising audio or video data frames (174) and control frames (175), and - means for sending organization frames (174, 175) to a plurality of receivers (6) by multicast within consecutive beacon intervals (137), It is characterized in that the system includes: - Means for synchronizing said audio or video data (9, 10, 174) among said plurality of receivers by means of an interrupt command included in each beacon interval (137). 24. The system of claim 23, wherein the end-of-frame interrupt of the accurately transmitted CF-End control frame is used for frequent low-jitter resynchronization.

Images