Wireless conference system that detects packet loss early

Abstract

The wireless conference system (100) comprises an access point (101; 300) and a plurality of conference units (111-115; 200), and is configured for bidirectional TDMA-based wireless communication of delay-sensitive audio data packets. The access point (101; 300) and the conference units (111-115; 200) each comprise an actively synchronized clock (203; 303) that generates a local audio clock signal used for processing the audio data packets and a local synchronization clock signal used for TDMA-based wireless communication. The conference units (111-115; 200) and the access point (101; 300) each comprise a transceiver (201; 301). The receiver (220; 320) therein comprises a packet loss detection section (222; 322) configured to detect loss of an audio data packet, which comprises means (223; 323) for determining an expected arrival time of the audio data packet from a local synchronization clock signal and a predetermined expected transmission delay, and means (224; 324) for detecting loss of an audio data packet if the audio data packet does not arrive by said expected arrival time. The receiver (220; 320) further comprises a packet loss concealment unit (225; 325) configured to generate a replacement packet for the lost audio data packet.

Description

[Technical field]

[0001] The present invention relates generally to wireless conference systems used in conference rooms for multiple users to participate in a conference. Such conference systems typically include a central access point managed by a conference organizer or chairperson, and multiple user conference units installed or integrated in user desks present in the conference room. The present invention particularly relates to latency sensitive two-way audio transmission between the conference units and the access point in such conference systems. [Background technology]

[0002] Conference rooms are often equipped with conference systems that allow a large number of users, from tens to hundreds, to participate in a single conference. A wireless conference system typically comprises a central access point and multiple conference units coupled to the central access point. For example, a separate conference unit may be provided for each user, or one conference unit may be shared by two adjacent users in a conference room. Each conference unit typically has a microphone connector for connecting a microphone (e.g., a gooseneck microphone), a built-in speaker, and one or more headphone connectors for connecting headphones. Each conference unit further comprises a controller, a data processor, and a transceiver configured for wireless bidirectional data transfer with a similar transceiver in the central access point. The transceiver of the conference unit comprises a transmitter capable of transmitting digital data packets upstream to a receiver that forms part of the transceiver of the central access point. Similarly, the transceiver of the central access point comprises a transmitter capable of transmitting digital data packets downstream to a receiver that forms part of the transceiver of the conference unit. In a typical situation where one participant is speaking and others are listening, the audio captured by the microphone of a conference unit as the person speaks is digitized, packetized, and transmitted upstream to a central access point. The access point processes the received audio packets, including mixing them with other audio input sources, and distributes the processed audio downstream to all other conference units in the conference room, where a receiver receives the audio and a processor processes the received audio packets for playback on headphones. This includes transmitting the audio downstream to the conference unit that originally received the audio packets so that the speaker can hear themselves on the headphones.

[0003] The wireless two-way communication of audio between the conference units and the access points has low-latency requirements, also called real-time requirements. Depending on the quality standards applied, the end-to-end delay in the round-trip time (RTT) in the conference system, i.e. the maximum delay for an audio packet to travel round trip between the conference unit and the central access point, including all processing in the conference unit and the central access point, may be limited to a value ranging from 10 ms to 30 ms, e.g. 15 ms. Audio packets arriving later than the maximum allowed delay time are not used for playback. In case an audio packet does not arrive in time (packet loss), the lost packet is replaced by a packet determined by a packet loss concealment algorithm to avoid audible artifacts that would be noticed by the conference participants of the conference units that did not receive the audio packet in time. It is important to detect the packet loss as early as possible in order for such a packet loss concealment algorithm to be able to generate a replacement packet in a timely manner.

[0004] In an exemplary conferencing system, voice packets are transferred in time slots of a TDMA frame over a Wi-Fi connection between the conference units and a central access point. The TDMA frame may have a length of, for example, 5 milliseconds and be subdivided into 10 time slots, each with a length of 500 microseconds. The conference unit captures 5 milliseconds of voice, typically the speech of the conference participants, digitizes the 5 milliseconds of voice into a digital voice packet, and transmits the digital voice packet in a single time slot of the TDMA frame. A 5 millisecond delay has already occurred in the packet generation in this example. When soft TDMA is applied to avoid interference with other users of the channel, i.e., a TDMA scheme in which a time slot is fixedly assigned to a transceiver, but the transceiver listens to the channel for interfering traffic in the time slot and determines whether the channel is free before transmitting data in that time slot, soft TDMA introduces a variable delay or jitter, which is called Listen-Before-Talk jitter or LBT jitter. The interrupt mechanisms that control data packet processing on the sender and receiver, packet and voice processing times, data packet propagation times over the radio channel, and poor synchronization between the transmitter and receiver can combine with LBT jitter to cause additional delays of up to 2 ms in one-way transmission. In a conferencing system implementing a soft-TDMA scheme with a TDMA frame length of 5 ms, the one-way delay for voice packets can reach up to 7 ms.

[0005] A simple way to detect packet loss in order to initiate packet loss concealment relies on the sequence number of the digital voice packet. If the receiver receives a voice packet with a later sequence number, it considers the packet lost. However, if you rely on the sequence number, the earliest that you can detect the loss of a data packet is when you receive a later data packet. To generate and send the next voice packet, at least another 5 ms has passed in the above example, a 5 ms TDMA frame, which is not enough time for the receiver to apply packet loss concealment and generate a replacement packet for the previous lost voice packet.

[0006] A solution to improve packet loss detection, which relies on the sequence number of voice packets, is to shorten the TDMA frame length and therefore the length of the voice segments embedded in a single voice packet. However, shortening the packet length has a negative impact on the capacity of the wireless channel. Since the overhead per packet is assumed to be constant, the overhead / payload ratio is increased, which would result in a decrease in the effective capacity of the wireless link. Furthermore, shortening the packets requires an increase in the packet processing speed, which would require more expensive processors.

[0007] US patent application US2015 / 0201289A1 "Method and Apparatus for Rendering Audio in Wireless Hearing Instruments" recognizes the problem of audio packet loss over a wireless communication link in another application, more precisely in the application of hearing aids worn on one or both sides of a person's head to assist patients suffering from hearing loss. In paragraph

[0030] of US2015 / 0201289A1, packets are retransmitted from the audio source device to improve the overall packet error rate performance of the wireless link. Retransmissions may be performed if the audio source device fails to receive an acknowledgement from the audio sink device, or may be sent unconditionally with a set number of retransmissions. The proposal to send acknowledgements and retransmissions indicates that the delay constraints in the hearing aid application are less stringent than in a conferencing system. As shown in the above example, the quality constraints of the conferencing system do not allow retransmissions. Paragraphs

[0029] and

[0031] of US2015 / 0201289A1 further state that the packet concealment strategy can be determined using link layer information shared between the radio of the wireless communication circuit and the DSP (Digital Signal Processor) of the processing circuit. The TDMA mechanism deployed in the radio has an inherently good timing mechanism that can schedule packet arrivals. If a scheduled reception event occurs without a received packet, the radio notifies the DSP of a packet loss, allowing the DSP to insert a packet loss concealment frame of information.

[0008] US patent application US2019 / 0104423A1 "Ultra-Low Latency Audio over Bluetooth" recognizes the problem of low latency requirements for audio packets transferred over a wireless connection in another application, more precisely in an Audio-over-Bluetooth transfer application between a device and a wireless headset or wireless earbud. US2019 / 0104423A1 also teaches reliance on acknowledgements and retransmissions, enhancing acknowledgements by combining BTC (Bluetooth Classic) packets and BLTE (Bluetooth Low Energy) packets within a single Bluetooth frame, limiting the upper limit of the number of retransmissions and packet concealment per frame cycle, and reducing the latency for the transmission of wireless audio packets using time-efficient audio coding and decoding implementing FEC (Forward Error Correction) such as RS (Reed-Solomon). This is described, for example, in paragraph

[0005] of US2019 / 0104423A1.

[0009] The object of the present invention is to disclose an embodiment of a conference system that overcomes or mitigates one or some of the above-mentioned drawbacks of existing solutions. In particular, the object of the present invention is to disclose an embodiment of a conference system in which audio packet loss occurring in the wireless transmission between the conference units and the central access point is detected faster, thereby enabling faster initiation of packet concealment, so that the conference system can meet the delay requirements and quality criteria applied to the conference system. A further object of the present invention is to disclose an embodiment of such a conference system without adversely affecting the effective bandwidth of the wireless link. Summary of the Invention

[0010] According to an embodiment of the present invention, the above object is achieved by a wireless conference system adapted for allowing multiple users to participate in a conference in a conference room, as defined by claim 1, the wireless conference system comprising an access point and a plurality of conference units, one or more of the access point and the conference unit comprises a transceiver configured for bidirectional time division multiple access-based or TDMA-based wireless communication of delay-sensitive voice data packets between the one or more conference units and the access point, the transceiver comprising a transmitter and a receiver; the access point and the one or more conference units have respective clocks that are actively synchronized, one of the respective clocks being configured to generate a local audio clock signal used locally in processing the audio data packets and a local synchronization clock signal used in the TDMA-based wireless communication; The receiver comprises a packet loss detection unit configured to detect loss of audio data packets transmitted from the conference unit to the access point or vice versa, the packet loss detection unit comprising: means configured to determine an expected arrival time of said voice data packets from said local synchronous clock signal and a predetermined expected transmission delay; means configured to detect loss of the voice data packet if the voice data packet does not arrive by the expected arrival time; The receiver comprises a packet loss concealment unit configured to generate replacement packets for voice data packets whose loss has been detected by the packet loss detection unit.

[0011] Thus, according to an embodiment of the invention, the access point and one or more conference units are equipped with clocks or timers that are actively synchronized with high precision. As a result, the receiver knows when a data packet is to be transmitted and the receiver can determine the expected arrival time of the data packet. The actively synchronized clocks allow all transceivers to know the start and end times of a TDMA frame and the time slots within the TDMA frame within the clock synchronization tolerance. As a result, the receiver can derive the expected arrival time of a voice data packet from the local clock signal, i.e., the transmission time of that voice data packet plus the expected transmission delay. The transmission time is known from the synchronized clocks used by the transceivers in TDMA wireless communication. In effect, each conference unit knows the TDMA schedule and can derive where it is within the TDMA frame from its synchronized clock. The expected transmission delay takes into account the total time required for interrupt processing at the sender and receiver, propagation of the voice data packet over the radio link, jitter of various natures, and imprecision of various natures, and as a result indicates an upper bound on the total time from when the data packet processor at the sender sends out a voice data packet for transmission until the data packet processor at the receiver receives the same voice data packet for processing. If a data packet is not received by its expected arrival time, packet loss concealment is invoked at the receiver and a substitute packet is generated.

[0012] The transmission delay, i.e. the overall delay due to propagation, jitter, interrupt handling, processing and synchronization imprecision in the conferencing system, is in the range of a few milliseconds, typically 1-3 milliseconds for one direction. Packet loss concealment in embodiments according to the invention can be started much faster than conventional conferencing systems that rely on acknowledgements, packet retransmissions and timeouts at the receiver for packet loss detection, or on packet loss detection based on the numbering of voice packets and the number of missing voice packets at the receiver. The earlier the packet loss concealment can be started, the lower the end-to-end overall system delay and the lower the risk that the packet loss concealment cannot generate a replacement for the insertion and playback of the voice stream in a timely manner, resulting in a lower number of audible artifacts. Furthermore, embodiments of the conferencing system according to the invention do not require a shorter TDMA frame length and / or voice data packet length. Such measures allow faster detection of packet loss and earlier start of packet loss concealment, at the expense of an increased overhead / payload ratio and therefore a reduced effective bandwidth of the wireless link (the amount of useful data that can be transferred per unit time, i.e. voice samples).

[0013] A conference unit in the present invention refers to various units installed or integrated on users' desks in a conference room. Such conference units typically include an audio input connector, such as a built-in microphone array or a connector for a gooseneck microphone, which can be used by a single user or shared between two users if the conference unit is installed between the seats of two adjacent users. The conference unit also typically includes one or more audio output connectors, such as a connector for headphones for one or more users, and typically also includes a built-in speaker. The audio input and audio output connectors may be combined into one connector. The conference unit may include connectors for a camera or other sensors, may include a display, and may include physical or virtual (i.e. displayed) buttons for controlling the audio input (e.g., muting the microphone), controlling the audio output (e.g., controlling the headphone volume), controlling other sensors, interacting with the chairperson (e.g., requesting to speak), and / or serving as voting buttons. The conference unit may further include indicators, such as colored LEDs, that indicate the status of the conference unit to the chairperson, the user of the conference unit, and / or other conference participants. The conference unit further comprises a processor for digitizing and packetizing the audio captured by a microphone connected to its audio input connector, and a radio transmitter for transmitting the audio data packets to a central unit, the so-called access point. The conference unit also comprises a receiver for receiving the audio data packets from the central unit, and a processor for depacketizing the received audio data packets and generating an audio stream (analog or digital) that is provided via the audio output connector. The transmitter and receiver together form a transceiver for two-way communication with the central unit. The processor for generating the audio data packets for transmission and the processor for processing the received audio data packets can be integrated into a single physical processor.

[0014] In the present invention, an access point refers to a central unit that is managed and controlled by a chairperson or a conference organization. The access point performs two-way wireless communication with all conference units in the conference room by TDMA method. The access point includes a transmitter and a receiver that respectively transmit and receive audio packets in time slots of a TDMA frame. Audio data packets received in different time slots are from different conference units. The access point typically has a processor that processes the audio data packets received from the different conference units and selects or combines (generally processes) the audio packets from one or more conference units into a single audio stream for transmission to the conference unit by its transmitter. The access point can receive input from a chairperson who manages the conference and decides which conference participants are allowed to speak at any given time. In addition to the conference units, audio streams may also be provided to interpretation units to allow an interpreter or translator to upload an interpretation or translation of the audio stream that is further distributed to the conference units by the central access point.

[0015] A wireless conference system according to the present invention comprises a set of conference units installed in a conference room, and an access point to which the conference units are connected in a multipoint-to-point manner via a wireless bidirectional link shared in a TDMA manner.

[0016] Packet loss concealment in the context of the present invention includes any algorithm or technique that generates replacement voice samples for lost voice packets, i.e. voice data packets that do not arrive at the receiver or arrive at the receiver too late and therefore can no longer be processed and integrated into the voice stream in a timely manner. Packet loss concealment techniques typically use recently received voice packets to generate replacement packets for lost voice data packets. Packet loss concealment techniques generally seek to avoid or minimize audible effects as a result of the substitution and can be based, for example, on the frequency or tone of the most recent voice samples.

[0017] An embodiment of the wireless conference system according to the invention, as defined in claim 2, is arranged such that the access point and the one or more conference units do not acknowledge receipt of audio data packets.

[0018] In fact, a preferred embodiment of the wireless conferencing system according to the present invention implements a protocol that does not involve or disables acknowledgments for audio data packets. Although acknowledgments or their absence trigger retransmissions, attempts to recover lost audio data packets by retransmissions would delay the initiation of packet loss concealment and increase the overall latency of the wireless conferencing system.

[0019] An embodiment of the wireless conference system according to the invention, as defined in claim 3, is arranged such that the access point and the one or more conference units do not retransmit lost audio data packets.

[0020] In fact, preferred embodiments of the wireless conferencing system according to the present invention implement a protocol that does not involve retransmission of upstream and downstream audio data packets, since as explained above, retransmission attempts would delay the initiation of packet loss concealment and therefore increase the overall latency of the wireless conferencing system.

[0021] An embodiment of the wireless conferencing system according to the invention, as defined in claim 4, has a round-trip delay limit of 25 milliseconds for wireless transmission from the conference unit to the access point and from the access point to the conference unit.

[0022] Thus, an embodiment of the wireless conferencing system can set a 25 millisecond limit on the round trip time of audio data packets. In other words, this limit sets the maximum allowable delay between when a conference participant speaks into a microphone connected to their conference unit and when they are heard on a headphone connected to the same conference unit. Those skilled in the art will appreciate that alternative embodiments of the wireless conferencing system according to the present invention may implement any other round trip time delay limit less than 25 milliseconds. Such alternative embodiments set a higher quality standard for audio, as long as the underlying technology, such as a packet loss concealment algorithm, can meet this lower round trip time limit.

[0023] In an embodiment of the wireless conference system according to the invention, as defined in claim 5, delay-sensitive audio data packets have a round-trip delay limit of 15 milliseconds for wireless transmission from the conference unit to the access point and from the access point to the conference unit.

[0024] In fact, the preferred embodiment of the wireless conferencing system limits the round trip time of an audio data packet between the conference unit and the central access point to 15 milliseconds. Those skilled in the art will appreciate that it is possible to implement a round trip time limit of less than 15 milliseconds, but at the risk of having to introduce more expensive high capacity data packet processors and / or having to implement shorter TDMA frames and shorter audio packet lengths that adversely affect the effective bandwidth of the wireless link to minimize audible artifacts.

[0025] In an embodiment of the wireless conference system according to the invention, as defined in claim 6, the TDMA based wireless communication uses TDMA frames of 5 ms.

[0026] Thus, the preferred embodiment implements a TDMA frame having a length of 5 milliseconds. This means that the audio data packets also contain 5 milliseconds of audio samples. With an audio sampling rate of 48 kHz, each audio data packet contains 240 audio samples. The 240 audio samples constitute the payload portion of the audio data packet. Furthermore, the audio data packets contain overhead. Those skilled in the art will appreciate that short TDMA frames and short audio data packets have a negative impact on the effective bandwidth of the wireless link. The payload portion of the audio data packet decreases, while its overhead portion remains constant. Those skilled in the art will further appreciate that long TDMA frames and long audio data packets can complicate the task of packet processors and packet loss concealment techniques in generating an audio stream in a timely manner without audible artifacts. In other words, the preferred TDMA frame length of 5 milliseconds is the result of a trade-off between the effective bandwidth of the wireless link between the conference unit and the access point and the quality criteria that the conference system must meet.

[0027] An embodiment of the wireless conference system according to the invention, as defined by claim 7, is arranged such that the transmitter listens for interfering traffic in an assigned time slot in the TDMA frame before transmitting an audio data packet. In fact, the preferred embodiment of the present invention implements a so-called soft TDMA scheme, in which time slots constituting part of a TDMA frame are fixedly assigned to the conference units according to a predetermined scheme, but the transmission time within each time slot of the TDMA frame is flexibly determined by the conference units based on a Listen Before Talk (LBT) mechanism. The LBT mechanism has the advantage of more effectively using the entire available bandwidth of the wireless channel, since the wireless channel can be used simultaneously by different wireless systems. On the other hand, the LBT mechanism causes jitter or uncertainty in the arrival time of the audio data packets at the receiver, since the transmitter transmits the audio data packets only when it is established that the time slot is not used by other transmitters inside or outside the wireless conference system. The LBT jitter in a conference system is usually present on the order of 1-2 milliseconds.

[0028] In an embodiment of the wireless conference system according to the present invention, as defined in claim 8, the wireless communication uses Wi-Fi (IEEE802.11).

[0029] In fact, the Wi-Fi protocol is available as a wireless technology in conferencing systems to connect conference units and a central access point: Wi-Fi has the reach to cover the area of ​​a conference room with a single access point (or at most a few access points), and multiple channels or frequency bands to deal with interference.

[0030] In an embodiment of the wireless conference system according to the invention, as defined in claim 9, one or more conference units comprise a clock synchronization unit and are configured to actively synchronize their respective clocks with a clock in the access point based on timestamps inserted in beacon messages periodically broadcast by the access point.

[0031] Thus, a preferred method for realizing active synchronization of the access point clock and the conference unit clock relies on beacon messages broadcast periodically by the access point. If the access point relies on the Wi-Fi protocol, for example, the beacon messages may correspond to messages in which the access point periodically broadcasts its SSID so that devices can detect the presence of the access point and establish a connection. Compared to other messages, beacon messages are generally transmitted with a low modulation scheme, i.e., using a simpler constellation scheme and increased redundancy, making the beacon messages more robust, with a larger bandwidth and reduced risk of being lost before reaching the receiver. Thus, by using beacon messages to convey time values ​​or timestamps between the access points and the conference units of the conference system, the active clock synchronization essential to the invention is made more robust. A further advantage in the case of Wi-Fi is that the beacon messages transferred therein are backward compatible with previous flavors or versions of Wi-Fi technology.

[0032] In an embodiment of the wireless conference system according to the invention, as defined in claim 10, the predetermined expected transmission delay is determined as the sum of the propagation delay, the jitter, the interrupt handling delay, the processing delay and the clock synchronization imprecision.

[0033] Indeed, the overall uncertainty in the reception time of a voice data packet in the preferred embodiment includes a first contribution due to effective propagation through the air, which is typically on the order of 100-200 microseconds. The overall uncertainty in the reception time of a voice packet may further include a second contribution due to LBT jitter in embodiments where a soft TDMA scheme is implemented. This second contribution is large and typically ranges from 1-1.5 milliseconds for a 5 millisecond TDMA frame. The overall uncertainty in the reception time of a voice packet further includes third, fourth and fifth contributions arising respectively from interrupt handling at the transmit and receive sides, i.e. interrupt handling indicating that an event has occurred, e.g. packet reception, packet and voice processing times, and synchronization uncertainty between the transmit and receive clocks, which may depend on the active synchronization mechanism deployed.

[0034] In an embodiment of the wireless conference system according to the invention, as defined in claim 11, the jitter delay includes a listen-before-talk jitter contribution.

[0035] As mentioned above, jitter of various natures can contribute to the overall inaccuracy regarding the reception time of a voice packet. When a soft TDMA scheme is applied, where the transmitter waits until the channel is free before transmitting in the scheduled time slot, a significant jitter contribution comes from the listen-before-talk behavior of the transmitter.

[0036] In an embodiment of the wireless conference system according to the invention, as defined in claim 12, the predetermined expected transmission delay is set to a value between 1.5 ms and 2 ms.

[0037] Testing has shown that in a conferencing system using soft-TDMA Wi-Fi with a 5 ms time frame for communication between the conference units and the central access point of the conferencing system, and using Wi-Fi beacon timestamps for active synchronization, with an expected overall transmission delay of 1.5 ms, it is possible to implement the invention with substantial benefits in packet loss concealment detection and early start-up, without sacrificing the effective bandwidth of the wireless link.

[0038] According to a second aspect, as defined in claim 13, the present invention relates to a method for transferring delay-sensitive audio data packets between one or more conference units and an access point in a conference system adapted to allow multiple users to participate in a conference in a conference room, the method comprising bidirectional time division multiple access or TDMA-based wireless communication of the audio data packets, the method further comprising: actively synchronizing clocks of each of the one or more conference units and the access point, one of the respective clocks being configured to generate a local audio clock signal used locally in processing the audio data packets and a local synchronization clock signal used in the TDMA-based wireless communication; detecting loss of voice data packets transmitted from the conference unit to the access point or vice versa, determining an expected arrival time of said voice data packet from said locally synchronous clock signal and a predetermined expected transmission delay; detecting a loss of the voice data packet if the voice data packet does not arrive by the expected arrival time; and generating a replacement packet for the voice data packet whose loss has been detected through packet loss concealment. [Brief description of the drawings]

[0039] [Figure 1] FIG. 1 is a schematic diagram of one embodiment of a wireless conferencing system 100 in accordance with the present invention. [Diagram 2] FIG. 2 is a functional block diagram of an embodiment of a conference unit 200 forming part of an embodiment of a wireless conference system 100 in accordance with the present invention. [Diagram 3] FIG. 3 is a functional block diagram of an embodiment of an access point 300 forming part of an embodiment of a wireless conferencing system 100 in accordance with the present invention. [Figure 4] FIG. 4 is a pseudo flow diagram illustrating one embodiment of a method for forwarding delay-sensitive voice data packets in accordance with the present invention. [Diagram 5] FIG. 5 illustrates an exemplary embodiment of a computing system 500 suitable for performing one or more steps in accordance with embodiments of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0040] FIG. 1 shows a wireless conference system comprising an access point AP or 101 and five conference units CU1 to CU5 or 111 to 115. The conference units 111 to 115 capture audio signals from each conference participant with a microphone, digitize the audio signals into audio data packets, and transmit the audio data packets to the access point 101. The access point 101 receives audio packets from one or more conference units, processes the audio packets, and generates new audio packets for distribution to all conference units 111 to 112 such that the conference units 111 to 115 receive audio streams corresponding to an audio signal generated by a single selected conference participant (the selection can be made by a chairperson controlling the access point 101), a composite audio signal generated by multiple selected conference participants (again, the selection can be made by a chairperson controlling the access point 101), or a processed audio signal that has been, for example, interpreted or translated, an interpretation or translation superimposed on the original audio signal, etc. In the embodiment of FIG. 1, the conference units 111 to 115 communicate bidirectionally wirelessly with the access point 101 using Wi-Fi. The Wi-Fi channel selected for the upstream direction, i.e., from the conference units 111-115 to the access point 101, is shared using a soft time division multiple access (soft TDMA) protocol. In the soft TDMA protocol, the transmitter of the conference unit must first listen for a free channel before transmitting in the assigned time slot. The length of a TDMA frame in the conference system 100 of FIG. 1 is assumed to be 5 ms. The TDMA frame contains 10 time slots of 0.5 ms. If the channel is occupied, the transmitter waits until the channel is free. This Listen-Before-Talk or LBT mechanism introduces jitter, i.e., uncertainty in the arrival time of packets at the access point 101 within the assigned time slot. In the downstream direction, i.e., from the access point 101 to the conference units 111-115, the voice data packets are also transmitted in the time slots of the TDMA frame.Generally, the same TDMA frame is used for both uplink and downlink, and the same Wi-Fi channel is used for both uplink and downlink. The transmitter in the access point 101 also implements an LBT mechanism to avoid interference with other Wi-Fi networks using the same channel.

[0041] FIG. 2 is a diagram showing functional building blocks of a conference unit CU or 200 relevant from the viewpoint of the present invention. The conference unit 200 in FIG. 2 shows a possible implementation example of each of the conference units 111-115 shown in FIG. 1. The conference unit 200 comprises a transceiver 201, an antenna 202 for wireless transmission, a clock 203, a clock synchronization logic 204, an audio input connector or microphone connector 205, and an audio output connector or headphone connector 206. FIG. 2 further shows, in a schematic manner, a microphone 250 connected to the microphone connector 205 for capturing audio signals, e.g., speech of the conference participants, and a headphone 260 or built-in speaker connected to the audio output connector 206 for allowing the conference participants to follow the conference conversation. The transceiver 201 comprises a transmitter 210 and a receiver 220. The transmitter 210 comprises a processor 211, which is configured to receive an audio signal captured by the microphone 250, digitize the audio signal, and packetize the digitized audio signal into audio data packets, each containing 5 milliseconds of the captured audio. The processor 211 transfers the generated audio data packets to the antenna 202, i.e. the radio interface, which transmits the audio data packets in the upstream direction to the access point of the conference system using the Wi-Fi protocol with the soft TDMA and LBT mechanisms described above. In the downstream direction, the antenna 202 periodically receives Wi-Fi beacon messages from the associated access point. The beacon messages contain timestamps or time values ​​that are used by the clock synchronization unit 204 to synchronize the clock 203 with the clock in the access point. The clock signal is generated locally by the clock 203 in the conference unit 200, but is actively synchronized with the clock in the access point through the beacon messages and is used by the arrival time estimation unit 223, which estimates the arrival time of the next audio data packet. Since the clock 203 is actively synchronized with the clock used to transmit the audio data packets in the access point, the ETA unit 223 can accurately determine the time when the next audio data packet will be available for transmission in the access point.The ETA unit 223 increases the time by a predefined transmission delay, i.e. the allowance required for interrupt handling at the sender and receiver, for air propagation over the Wi-Fi channel, and for jitter that may be caused, for example, by the LBT mechanism implemented at the sender. This transmission delay may furthermore take into account the inaccuracy of the clock synchronization between the clock 203 and the clock in the access point to which it is actively synchronized. In the example of FIG. 2, the transmission delay is predefined to correspond to 1.5 ms. The locally generated clock signal and the predefined transmission delay allow the ETA unit 223 to defensively estimate the arrival time of the next audio data packet. The loss detection unit 224, which is interrupted by the processor 221 upon the arrival of the audio data packet, verifies whether the audio data packet has been received by the next expected arrival time of the audio data packet. If the audio data packet is received in time, the processor 221 processes the received audio data packet and allows the contained audio samples to be output as part of the audio stream via the audio connector 206. Each time an audio data packet is not received by its expected arrival time, the loss detection unit 224 triggers a packet concealment unit 225, which is part of the receiver 220, to generate a substitute packet and feed the substitute packet to the processor 221 for insertion into the audio stream output via the audio connector 206. The packet concealment unit 225 applies a processing-intensive concealment algorithm, typically using a previously received audio data packet, to generate a substitute audio data packet to replace the missing audio data packet, which does not produce audible artifacts when inserted into the received audio stream. Active clock synchronization and arrival time estimation of the audio data packets allows early triggering of concealment of lost packets. This allows the conference unit 200 to meet the delay requirements set for the conference system, typically 15 ms to 25 ms for the round trip time of an audio packet between the conference unit 200 and the access point.

[0042] Fig. 3 shows functional building blocks of an access point AP or 300 relevant from the point of view of the present invention. The access point 300 of Fig. 3 shows a possible implementation of the access point 101 shown in Fig. 1. The access point 300 comprises a transceiver 301, an antenna 302 for wireless transmission, a clock 303 and an interface 304 for the chairperson or organization of the conference. The transceiver 301 comprises a transmitter 310 and a receiver 320. The transmitter 310 comprises a processor 311 configured to generate audio data packets, each containing 5 milliseconds of audio. The content of these audio data packets is controlled by the chairperson of the conference via the interface 304. The content of these audio data packets can be, for example, audio received from a single conference unit, audio received from multiple conference units, audio received from an interpreter or translator, etc. The processor 311 transfers the generated audio data packets to the antenna 302, i.e. the wireless interface, which transmits the audio data packets downstream to the conference units of the conference system using the Wi-Fi protocol with the soft TDMA and LBT mechanisms described above. In the downstream direction, the antenna 302 also periodically transmits Wi-Fi beacon messages, including for example an access point identifier, so that the conference units can associate with the access point 300. These beacon messages also contain a timestamp or a time value from the clock 303, which allows the conference units to actively synchronize their clocks with the clock 303 of the access point 300. The clock signal generated locally by the clock 303 of the access point 300 is used by the arrival time estimation unit 323, which estimates the arrival time of the next audio data packet coming from the conference unit. Since the clock 303 is actively synchronized with the clock used to transmit audio data packets in the conference unit, the ETA unit 323 can accurately determine the time when the next audio data packet will be available for transmission in the conference unit.The ETA unit 323 increases the time with a predefined transmission delay, i.e. the allowance required for interrupt handling at the sender and receiver, propagation in the air on the Wi-Fi channel, and possible jitter as a result of, for example, the LBT mechanism implemented at the sender. This transmission delay may furthermore take into account the inaccuracy of the clock synchronization between the clock 303 and the clock in the conference unit with which it is actively synchronized. In the example of FIG. 3, the transmission delay is predefined to correspond to 1.5 ms. The locally generated clock signal and the predefined transmission delay allow the ETA unit 323 to defensively estimate the arrival time of the next audio data packet at the conference unit. The loss detection unit 324, which is interrupted by the processor 321 upon the arrival of the audio data packet, verifies whether the audio data packet has been received by the next expected arrival time of the audio data packet. If the audio data packet is received in time, the processor 321 processes the received audio data packet and outputs the contained audio samples for downstream transmission. Each time an audio data packet is not received by its expected arrival time, the loss detection unit 324 triggers a packet concealment unit 325, which is part of the receiver 320, to generate a replacement packet and feed the replacement packet to the processor 321 for insertion into the outgoing audio stream. The packet concealment unit 325 applies a processing-intensive concealment algorithm, typically using a previously received audio data packet, to generate a replacement audio data packet to replace the missing audio data packet, which does not produce audible artifacts when inserted into the received audio stream. Active clock synchronization and arrival time estimation of the audio data packets allows early triggering of concealment of lost packets. This allows the access point 300 to meet the delay requirements set for the conference system, typically 15 ms to 25 ms for the round trip time of an audio packet between the conference unit and the access point 300.

[0043] Fig. 4 is a pseudo flow diagram illustrating an embodiment of a method for forwarding delay-sensitive voice data packets according to the present invention. The pseudo flow diagram of Fig. 4 shows the steps performed at the receiving side. In a first step 401, the expected arrival time of the voice data packet is determined from a time signal received from an actively synchronized clock 411 and an expected transmission delay 412, which is pre-set and recorded, for example in a computer memory or register. In a second step 402, it is verified whether the voice data packet is received by the expected arrival time determined in the first step 401. If the voice data packet is received by the expected arrival time, in step 403, the received voice data packet is processed as usual. If the voice data packet is not received by the expected arrival time, in step 404, packet loss concealment is activated in order to generate a replacement voice packet for the lost voice data packet. The packet loss concealment in step 404 is activated as soon as the expected arrival time is reached. Finally, in step 405, the processed speech from the received speech data packets or the speech generated by concealment is streamed to form a continuous speech stream without audible artifacts.

[0044] 5 illustrates a suitable computing system 500 for implementing one or more steps of an embodiment of the method for transferring delay-sensitive audio data packets according to the present invention. The computing system 500 is generally formed as a suitable general-purpose computer and may include a bus 510, a processor 502, a local memory 504, one or more optional input interfaces 514, one or more optional output interfaces 516, a communication interface 512, a storage element interface 506, and one or more storage elements 508. The bus 510 may include one or more conductors that allow communication between the components of the computing system 500. The processor 502 may include any type of conventional processor or microprocessor that interprets and executes programming instructions. The local memory 504 may include a random access memory (RAM) or other type of dynamic storage device that stores information and instructions for execution by the processor 502, and / or a read-only memory (ROM) or other type of static storage device that stores static information and instructions for use by the processor 502. The input interface 514 may include one or more conventional mechanisms that allow an operator or user to input information into the computing device 500, such as a keyboard 520, a mouse 530, a pen, a voice recognition and / or biometric mechanism, a camera, etc. The output interface 516 may include one or more conventional mechanisms that output information to an operator or user, such as a display 540. The communication interface 512 may include any transceiver-like mechanism, such as, for example, one or more Ethernet interfaces, that allows the computing system 500 to communicate with other devices and / or systems, such as other computing devices 581, 582, 583. The communication interface 512 of the computing system 500 may be connected to another computing system by a local area network (LAN) or a wide area network (WAN), such as, for example, the Internet.The storage element interface 506 may include a storage interface, such as a Serial Advanced Technology Attachment (SATA) interface or a Small Computer System Interface (SCSI), for connecting the bus 510 to one or more storage elements 508, e.g., one or more local disks, such as SATA disk drives, and for controlling the reading and writing of data to and / or from these storage elements 508. Although the storage element 508 above is described as a local disk, in general any other suitable computer-readable medium may be used, such as a removable magnetic disk, an optical storage medium such as a CD or DVD-ROM disk, a solid-state drive, a flash memory card, etc. The computing system 500 may thus correspond to the processing circuits 211, 221 in the embodiment illustrated in FIG. 2, or the processing circuits 311, 321 in the embodiment illustrated in FIG. 3. The processing circuits used in the transmitter and receiver, respectively, of these embodiments may obviously be part of a single processor or computer, as indicated by the dashed rectangles in these figures.

[0045] Although the present invention has been described with reference to specific embodiments, it will be apparent to those skilled in the art that the present invention is not limited to the details of the aforementioned exemplary embodiments, and that the present invention can be embodied with various changes and modifications without departing from the scope thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, and the scope of the present invention is indicated by the appended claims, rather than by the foregoing description, and therefore all changes that come within the meaning and range of equivalence of the claims are intended to be embraced therein. In other words, this patent application is intended to cover any modifications, variations, or equivalents that are within the scope of the basic underlying principles and whose essential attributes are claimed. Furthermore, readers of this patent application will understand that the words "comprise" or "include" do not exclude other elements or steps, that the words "a" or "an" do not exclude a plurality, and that a single element, such as a computer system, processor, or other integrated unit, may perform the functions of several means recited in the claims. Reference signs in the claims should not be construed as limiting the respective claims. Terms such as "first," "second," "third," "a," "b," "c," and the like, when used in the description or claims, are introduced to distinguish between similar elements or steps and do not necessarily describe them sequentially or chronologically. Similarly, terms such as "top," "bottom," "upper," "lower," and the like, are introduced for purposes of explanation and do not necessarily indicate relative positions. The terms so used are interchangeable under appropriate circumstances, and it will be understood that embodiments of the invention are capable of operating in accordance with the invention in other sequences or with orientations different from those described or illustrated above.

Claims

1. A wireless conference system (100) adapted to allow a plurality of users to participate in a conference in a conference room, the wireless conference system (100) comprising an access point (101; 300) and a plurality of conference units (111-115; 200); one or more of the access point (101; 300) and the conference units (111-115; 200) comprises a transceiver (201; 301) configured for bidirectional time division multiple access (TDMA) based wireless communication of delay sensitive voice data packets between the one or more conference units (111-115; 200) and the access point (101; 300), the transceiver (201; 301) comprising a transmitter (210; 310) and a receiver (220; 320); said access point (101; 300) and said one or more conference units (111-115; 200) comprise respective clocks (203; 303) that are actively synchronized, one of said respective clocks (203; 303) being configured to generate a local audio clock signal that is used locally in processing said audio data packets and a local synchronization clock signal that is used for said TDMA based wireless communication, The receiver (220; 320) comprises a packet loss detection unit (222; 322) configured to detect loss of voice data packets transmitted from a conference unit (111-115; 200) to the access point (101; 300) or vice versa, the packet loss detection unit (222; 322) comprising: means (223; 323) adapted to determine an expected arrival time of said voice data packets from said local synchronous clock signal and a predetermined expected transmission delay; means (224; 324) adapted to detect loss of said voice data packet if said voice data packet does not arrive by said expected arrival time, The wireless conference system (100) is characterized in that the receiver (220; 320) comprises a packet loss concealment unit (225; 325) configured to generate a substitute packet for the voice data packet whose loss has been detected by the packet loss detection unit (222; 322).

2. 2. The wireless conference system (100) of claim 1, wherein the access point (101; 300) and the one or more conference units (111-115; 200) are configured not to acknowledge receipt of voice data packets.

3. 2. The wireless conference system (100) of claim 1, wherein the access point (101; 300) and the one or more conference units (101-105; 200) are configured not to retransmit lost voice data packets.

4. 4. The wireless conference system (100) according to claim 1, wherein the delay-sensitive voice data packets have a round-trip time delay upper limit of 25 milliseconds for wireless transmission from the conference unit (111-115; 200) to the access point (101; 300) and from the access point (101; 300) to the conference unit (111-115; 200).

5. 5. The wireless conference system (100) according to claim 1, wherein the delay-sensitive voice data packets have a round-trip time delay upper limit of 15 milliseconds for wireless transmission from the conference unit (111-115; 200) to the access point (101; 300) and from the access point (101; 300) to the conference unit (111-115; 200).

6. 6. The wireless conference system (100) of claim 1, wherein the TDMA-based wireless communication uses 5 ms TDMA frames.

7. 7. The wireless conference system of claim 1, wherein the transmitter is configured to listen for interfering traffic in an assigned time slot in a TDMA frame before transmitting a voice data packet.

8. The wireless conference system (100) according to any one of claims 1 to 7, wherein the wireless communication uses Wi-Fi.

9. 9. The wireless conference system (100) of claim 1, wherein the one or more conference units (111-115; 200) comprise a clock synchronization unit (204) configured to actively synchronize their respective clocks (203) with a clock (303) in the access point (101; 300) based on timestamps inserted in beacon messages periodically broadcast by the access point (101; 300).

10. 10. The wireless conference system (100) of claim 1, wherein the predetermined expected transmission delay is determined as the sum of propagation delay, jitter, interrupt handling delay, processing delay, and clock synchronization inaccuracy.

11. The wireless conferencing system (100) of claim 7, wherein the predetermined expected transmission delay is determined as the sum of propagation delay, jitter, interrupt processing delay, processing delay, and clock synchronization inaccuracy, the jitter delay including contributions of listen-before-talk jitter.

12. The wireless conference system (100) of any of claims 1 to 11, wherein the predetermined expected transmission delay is set to a value between 1.5 ms and 2 ms.

13. 1. A method for transporting delay-sensitive voice data packets between one or more conference units (111-115; 200) and an access point (101; 300) in a wireless conference system (100) adapted to allow multiple users to participate in a conference in a conference room, the transport method comprising bidirectional, time division multiple access (TDMA) based wireless communication of the voice data packets, the method further comprising: - actively synchronising respective clocks (203; 303) of said one or more conference units (111-115; 200) and said access point (101; 300), wherein one of said respective clocks (203; 303) is configured to generate a local audio clock signal used locally in processing said audio data packets and a local synchronization clock signal used in said TDMA based wireless communication; detecting loss of voice data packets transmitted from said conference unit (111-115; 200) to said access point (101; 300) or vice versa, determining an expected arrival time of said voice data packet from said locally synchronous clock signal and a predetermined expected transmission delay; detecting a loss of the voice data packet if the voice data packet does not arrive by the expected arrival time; generating a replacement packet for the voice data packet whose loss was detected through packet loss concealment.

Images