Brady, P.T., "Effects of Transmission Delay on Conversational Behavior on Echo-free Telephone Circuits," BellSys.Tech.J.50(1971), 115-134.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....G.114 [7] gives guidelines on the tolerable delay for a normal telephone conversation. The maximum one way end to end delay acceptable by most users is only 150 ms. A telephony system with longer end to end delays will cause users to engage in collided talks and mutual silences more often [8]. To balance conversation quality and network efficiency, a compromise must be made between packet delay and header overhead. Such a tradeoff is illustrated by Fig. 1, in which a G.723.1 codec is assumed. In this paper, we present a novel voice multiplexing scheme which makes it possible to ....

P. T. Brady, "Effects of transmission delay on conversational behavior on echo-free telephone circuits," Bell Labs Tech. J., vol. 50, pp. 115--134, 1971.

....5 9 13 2 6 10 14 3 7 11 15 4 8 12 16 1 Packet loss 5 9 13 2 6 10 14 4 8 12 16 1 Reconstructed stream 2 4 5 6 8 9 10 12 13 14 16 IEEE Network . September October 1998 44 Retransmission Interactive audio applications have tight latency bounds, and end to end delays need to be less than 250 ms [36]. For this reason such applications do not typically employ retransmission based recovery for lost packets. If larger end to end delays can be tolerated, the use of retransmission to recover from loss becomes a possibility. A widely deployed reliable multicast scheme based on the retransmission ....

P. T. Brady, "Effects of transmission delay on conversational behavior on echofree telephone circuits," Bell Sys. Tech. J., vol. 50, Jan. 1971, pp. 115--34.

....comes in bursts, as seen in Figure 1, separated by periods of silence. Generally speaking, the duration of burst and silence periods are random in nature and can be considered as random variables. Work characterizing the alternation of these periods has been done by Brady and can be found in [19, 20]. If speech is transmitted as a series of packets on the network, a decision must be made as to the rate at which such packets are transmitted during a talk burst. A high rate of transmission implies better quality speech, but a greater demand is placed on the network because of the overhead ....

....choose a relatively rapid rate of packetization. Both the lengths of bursts of speech and the periods of silence between bursts were assumed to be exponentially distributed, with a mean length of 4 and one seconds, respectively. This assumption is consistent with that suggested by several others [18, 19, 20]. 4.5 Discussion of results The model constructed was used to determine the utilization of the control network for various numbers of network nodes and numbers of conversations. More specifically, while holding the arrival rate for control data requests per node constant (mean time between ....

P.T. Brady, "Effects of transmission delay on conversational behaviour on echo-free telephone circuits," in Bell Syst. Tech. J., Vol. 50, 1971.

.... 50 of the duration of the following silence period without affecting the quality of the playback [62] The tolerance of packet voice for playback adjustment assumes that talkspurts are generally isolated from each other by relatively long silence periods, a property reported in empirical studies [6, 8]. To ensure this isolation, voice protocols enforce a minimum intertalkspurt time, or hangover time, when marking talkspurt boundaries [24] 1.1.4 Error Control In a packet switched network end to end protocols may need to protect the client from the misordering, duplication, corruption, and ....

....Note, however, that the the control time cannot be arbitrarily large due to constraints on the end to end delay. Since voice data consists of an alternating series of talkspurts and silence periods and since talkspurts are 26 generally isolated from each other by relatively long silence periods [6, 8], voice protocols typically impose the control time on the first packet of each talkspurt. We refer to the playback time of a packet as the point in time at which playback of the packet must begin at the receiver in order to achieve a zero gap playback schedule for the talkspurt. We refer to the ....

[Article contains additional citation context not shown here]

P. T. Brady. Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits. Bell System Technical Journal, 50(1):115--134, January 1971.

....replacement, but this results in degradation of speech quality, which increases rapidly with loss rate and packet size. Mbone audio packet sizes range from 20 to 160ms, with 40ms packets being the default. For interactive conferencing the coding delay and propagation time should be small ( 250ms[4]) however smaller packets require more work by the routers and have a larger relative packet header overhead [5] The loss of larger packets has a more pronounced effect at the receiver. Our experience shows that a 10 loss rate for 40ms packets is the highest that can be tolerated with silence ....

Paul T. Brady. Effects of transmission delay on conversational behavior on echo-free telephone circuits. Bell System Technical Journal, 50:115-- 134, January 1971.

....deal with the loss To achieve the first objective, we observe that studies of human interaction have shown that people can tolerate a certain amount of latency in their conversations. In other words, we perceive a response from another person to be almost instant if it occurs within about 200ms [10]. We gain leverage from this observation thereby increasing the interval available for recovery with our first technique: adding a limited amount of playout buffering (about 100ms) at each side to gain some additional time for retransmission. Our second technique, which aims to maximize the ....

....user) deterioration of quality, by introducing limited buffering at the receiver. This is called playout buffering and the buffering delay is called playout or control delay. In interactive applications the playout delay is limited by the perceptual tolerance of the user, which is around 200 ms [10]. In other words, a human user can tolerate a maximum channel round trip delay of 200 ms in an interactive conversation 1 . In interactive applications the delay must be allocated to both endpoints, meaning that each may use up to 100 ms of playout delay. An example of using playout buffering is ....

Brady, P., "Effects of Transmission Delay on Conversational Behavior on Echo-free Telephone Circuits," Bell System Technical Journal, Vol. 50, No. 1, pp.115-134, January 1971.

....reduce the network bandwidth consumed by a voice channel is the suppression of transmissions during silence periods. The bandwidth savings are significant because the activity periods for interactive speech constitute only around 40 of the total time [4] 2. 2 Roundtrip Delay Behavioral studies [5, 15] have shown that roundtrip delays above a certain threshold degrade the interactive nature of the conversation. High quality voice applications require less than 200 ms roundtrip delays, but delays of up to 600 ms have been shown to be acceptable [15] Recent CCITT guidelines [6] suggest that even ....

P. T. Brady. Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits. Bell System Technical Journal, 50(1):115--134, January 1971.

....buffering at the receiver, with no perceptible deterioration in quality by the user. This is called playout buffering 2 and the buffering delay is called playout delay. In interactive applications the playout delay is limited by the perceptual tolerance of the user, which is around 200 mSec [3]. Therefore, for NTSC video a playout delay of 100 mSec allows three consecutive frames to be buffered at the receiver before delivered to the application (see Figure 2) Buffering increases the time available for retransmission by 100 mSec, which is sufficient for several retransmission attempts ....

Brady, P., "Effects of Transmission Delay on Conversational Behavior on Echo-free Telephone Circuits," Bell System Technical Journal, Vol. 50, No. 1, pp.115 - 134, January 1971. Page 24

....an important criterion if user interactivity is to be maintained. The value of the end to end delay should be kept below 600 ms in the absence of echoes, but is recommended to be 3.2. Robust Audio Tool Solutions 37 restricted to less than 400 ms if conversation patterns are not to break down [Brady, 1971] . End to end delay is directly impacted by the size of packets transmitted over the network. A delay equal to the size of one packet is incurred at the transmitter, since samples in the packet have to be collected before a packet can be sent. A rough estimate of the reconstruction delay required ....

....audio applications. In real time conferencing audio systems interactivity is known to be substantially reduced for round trip delays larger than 600 ms. Large round trip delays in conversational situations increase the frequency of confusions and amount of both double talking and mutual silence [Brady, 1971] . The components of this delay in a real time Internet audio tool are: ffl Processing and packetisation delay at transmitter. ffl Variable delay experienced by audio packets traversing the network. ffl Audio reconstruction buffering at the receiver necessary for smoothing out network delay ....

Paul T. Brady. Effects of transmission delay on conversational behavior on echo-free telephone circuits. Bell System Technical Journal, 50:115--134, January 1971.

....media, where real time signals like audio and video are delivered from a server somewhere in the network to a human user that interactively views the material. Unlike human to human communication, which requires relatively tight and consistent end to end delays for good interactive performance [6], server to human communication can afford a certain level of artificial delay. As a result, streaming media applications often have sufficient time to recover from lost packets through retransmission and thereby avoid unnecessary degradation in reconstructed signal quality. We refer to this ....

....do not estimate L. 5 Related Work There has been a significant amount of work done on streaming multimedia, and much of this is oriented towards improving the performance of interactive multimedia streams. Because delay requirements of interactive multimedia (less than 200 ms by some measures [6]) typically preclude soft ARQ for error recovery, interactive multimedia research has explored alternative ways to improving signal quality. Before discussing related work on soft ARQ, we briefly review these alternative techniques. One such technique for interactive media is to adjust the ....

P. Brady. Effects of transmission delay on conversational behavior on echo-free telephone circuits. Bell Syst. Tech. J., 50(1):115--134, January 1971.

....applications, although it performs well for non interactive use. The major advantage of interleaving is that it does not increase the bandwidth requirements of a stream. 3. 3 Retransmission Interactive audio applications have tight latency bounds and end to end delays need to be less than 250ms [5]. For this reason such applications do not typically employ retransmission based recovery for lost packets. If larger end to end delays can be tolerated the use of retransmission to recover from loss becomes a possibility. A widely deployed reliable multicast scheme based on the retransmission of ....

P. T. Brady. Effects of transmission delay on conversational behavior on echo-free telephone circuits. Bell System Technical Journal, 50:115--134, January 1971.

....audio applications. In real time conferencing audio systems interactivity is known to be substantially reduced for round trip delays larger than 600ms. Large round trip delays in conversational situations increase the frequency of confusions and amount of both double talking and mutual silence [18]. The components of this delay in a real time Internet audio tool are: ffl Processing and packetisation delay at transmitter. ffl Variable delay experienced by audio packets traversing the network. ffl Audio reconstruction buffering at the receiver necessary for smoothing out network delay ....

Paul T. Brady. Effects of transmission delay on conversational behavior on echo-free telephone circuits. Bell System Technical Journal, 50:115--134, January 1971.

....of the constraints on telecommunications traditionally imposed by human communications will prove valuable in simplifying the multi service networks. The tolerable round trip delay for echoless voice communication is quoted as being between 600 and 1200 milliseconds, based on research at Bell Labs [152]. But if visual cues are present indicating that a speaker is about speak, such as in VAT, then much longer delays are tolerable. Whilst there is some evidence for the minimum quality threshold, it remains to be proven that it will be acceptable to all users. Technically aware users may accept ....

P. T. Brady, "Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits," Bell System Technical Journal 50 (Jan 1971), 115--134.

.... that the the control time cannot be arbitrarily large due to constraints on the end to end delay (see Table 1) Since voice data consists of an alternating series of talkspurts and silence periods and since talkspurts are generally isolated from each other by relatively long silence periods [4, 5], voice protocols typically impose the control time on the first packet of each talkspurt. We refer to the playback time of a packet as the point in time at which playback of the packet must begin at the receiver in order to achieve a zero gap playback schedule for the talkspurt. We refer to the ....

....queue, we model a single unidirectional voice channel. The voice traffic stream is modeled as alternating talkspurts and silence periods whose lengths are exponentially distributed with means 350 ms and 650 ms, respectively. These figures represent a talk activity of 35 percent as suggested in [5]. The packetization interval, T p , determines the duration of speech (in milliseconds) captured by each network packet. The number of packets in a talkspurt will be the length of the talkspurt in milliseconds divided by T p . The one way delay of the network, TNet , consists of three ....

P. T. Brady. Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits. Bell System Technical Journal, pages 115--134, January 1971.

....uses this approach. Although buffering is suitable for most playback applications, this solution is not viable for many interesting classes of applications. For example, conferencing applications need low end to end latency and can only buffer a maximum of about 800 milliseconds worth of data [64, 41, 10]. This implies that the network delay can not exceed 800 milliseconds. Furthermore, in some playback applications the receiver may have extremely limited buffering capabilities due to cost constraints. In both these situations, delay and jitter must be bounded. The Tenet protocol suite allows ....

P. Brady, Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits, The Bell System Technical Journal, Vol. 50, Num.1, January, 1971, pp. 115-134.

....in quality to the user, by introducing limited buffering at the receiver. This is called playout buffering and the buffering delay is called playout or control delay. In interactive applications the playout delay is limited by the perceptual tolerance of the user, which is around 200 ms [4]. In other words, a human user can tolerate a maximum channel round trip delay of 200 ms in an interactive conversation 3 . In interactive applications the delay must be allocated to both endpoints, meaning that each may use up to 100 ms of playout delay. An example of using playout buffering is ....

Brady, P., "Effects of Transmission Delay on Conversational Behavior on Echo-free Telephone Circuits," Bell System Technical Journal, Vol. 50, No. 1, pp.115-134, January 1971.

....encodes only the difference between consecutive samples, reducing the number of bits per sample to 2 Gamma 5 bits. Coding techniques with even lower bit rates, e.g. Linear Predictive Coding (LPC) exist, though speech fidelity is frequently poor [11] 2. 2 Roundtrip Delay Behavioral studies [5, 13] have shown that roundtrip delays above a certain threshold degrade the interactive nature of the conversation. Quantifying this factor is difficult since individual human users have different tolerances for delay and these tolerances vary with the application. High quality voice applications ....

P. T. Brady. Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits. Bell System Technical Journal, 50(1):115--134, January 1971.

....encodes only the difference between consecutive samples, reducing the number of bits per sample to 2 Gamma 5 bits. Coding techniques with even lower bit rates, e.g. Linear Predictive Coding (LPC) exist, though speech fidelity is frequently poor [11] 2. 2 Roundtrip Delay Behavioral studies [5, 13] have shown that roundtrip delays above a certain threshold degrade the interactive nature of the conversation. Quantifying this factor is difficult since individual human users have different tolerances for delay and these tolerances vary with the application. High quality voice applications ....

P. T. Brady. Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits. Bell System Technical Journal, 50(1):115--134, January 1971.

....the network effects; this enables sample play out to be smoothed. In a packet speech system, the end to end delay is always a critical factor in the usability of a real time voice system, and should be kept below 600ms in the absence of echoes (The figure may be in fact be less than this 400ms) [6], if conversation patterns are not to break down. The size of the packets (in ms) chosen for a packet speech system directly impacts the endto end delay. A delay equal to the size of one packet is incurred at the transmitter, since the samples in the packet have to be collected before a packet ....

Brady P.T. 'Effects of Transmission delay on Conversational Behaviour on Echo-Free Telephone Circuits' Bell System Technical Journal, pp 115-134, January 1971.

....recovery. Dempsey et al. 5] applied retransmission for the recovery of audio packets. They showed that by adding some delay before the playout of each received audio packet, retransmission can be used to protect audio data from packet loss. Their work hinges on the earlier behavior study by Brady [3] showing that although less than 200 ms round trip delay is required for high quality voice applications, delays up to 600 ms can be tolerable by human ears. Ramamurthy and Raychaudhuri [21] applied a similar technique to video transmission over ATM. They analyzed the performance of video ....

P. Brady, "Effects of Transmission Delay on Conversational Behavior on Echo-Free Telephone Circuits, " Bell System Technical Journal, vol. 50, no. 1, pp. 115 -- 134, Jan. 1993.

No context found.

Brady, P.T., "Effects of Transmission Delay on Conversational Behavior on Echo-free Telephone Circuits," BellSys.Tech.J.50(1971), 115-134.

No context found.

Brady, P., "Effects of Transmission Delay on Conversational Behavior on Echo-free Telephone Circuits," Bell System Technical Journal, Vol. 50, No. 1, pp.115-134, January 1971.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC