jackson%sdcsvax@ucsd.edu> (12/08/89)
As I understand it, work is afoot to implement a standard for the ISDN defined 7kHz voice service, wherein audio sampled (presumably) at 16 ksamples per second is encoded (using cunning modern techniques) at the ISDN bearer channel rate (64 kbps). I envisage the appearance of "hi-fi" telephones capable of using this service. Voice would be clearer and music could be carried (with fidelity equivalent to that of a.m. radio). Further, digital technology could enable superior echo cancelling allowing speakerphone use without the "in-a-tomb" effect. Clearly, the new phones would have to be compatible with POTS phones, but Q.931 and SS7 know enough for the service to be negotiated automatically on call set-up. Such phones might become the next great consumer electronics fad, following compact discs and cellular phones. Once people heard the higher quality, they might feel they had to have one, to keep up with their yuppie friends. Any comments? From those who know how the technology is progressing? From potential owners? Oh yes, if these things caught on, they would drive the market for ISDN lines to residential as well as business premises. Just what the local carriers need! Dick Jackson
goldstein@delni.enet.dec.com (12/09/89)
In article <1933@accuvax.nwu.edu>, ttidca.TTI.COM!jackson%sdcsvax@ucsd.edu (Dick Jackson) writes... >As I understand it, work is afoot to implement a standard for the ISDN >defined 7kHz voice service, wherein audio sampled (presumably) at 16 >ksamples per second is encoded (using cunning modern techniques) at >the ISDN bearer channel rate (64 kbps). >Any comments? From those who know how the technology is progressing? >From potential owners? Funny you should ask. Yes, there's a new ISDN 7 kHz audio bearer service. It makes use of 64 kbps ADPCM encoding. (Digression: Standard PCM uses 64 kbps to do 3.1 kHz audio. ADPCM is more efficient, so 32 kbps is essentially adequate for 3.1 kHz audio, with only minimal distortion (modems might complain, humans won't). So if you use the ADPCM principle on the usual 64 kbps bandwidth, you can get better audio.) The network uses PCM to generate tones and announcements for ISDN telephones in the telephony bearer service. The 7 kHz standard says that you begin all calls in standard 3.1 PCM mode, specifying that it's really a 7 kHz call. Once the two ends are connected to each other, they do a handshake to confirm that they're ready to switch into 7 kHz mode. That way the terminals are in 3.1 PCM mode when doing call setup (talking to the network) and in 7 kHz ADPCM mode when actually communicating with each other. This hack makes it essentially transparent to the network, which will speed implementation. You just need the chips in your telephones. I don't personally see much use for it in "handsets", given their cruddy mic/speaker combos, but it could be very nice for speakerphones, audio dial-up program services, remote broadcast feeds, etc. fred (member, ANSI T1S1, speaking for himself)
stodol@diku.dk (David Stodolsky) (12/20/89)
goldstein@delni.enet.dec.com in <2002@accuvax.nwu.edu> writes >Funny you should ask. Yes, there's a new ISDN 7 kHz audio bearer >service. It makes use of 64 kbps ADPCM encoding. Mermelstein, P., (1988). G.722, A New CCITT Coding Standard for Digital Transmission of Wideband Audio Signals (IEEE Communications Magazine, v. 26, n. 1) describes a way to split audio input into two 4 khz bands using ADPCM coders. Audio data can be transmitted at 64, 56, or 48 kbits, thus allowing simultaneous transmission of other data. The system is targeted toward "audio- visual conferencing applications where one would like to approach the quality of face-to-face communication (p. 8)." My interest, is not the improvement in audio quality, but the use of data-speech multiplexing. This is projected in the article, for speaker identification or fax on the established connection. One of the major problems in teleconferencing is speaker selection, how to decide on the next speaker without using the normal cues one has when face-to-face. The Danish Telecommunication Research Labs. produced a pre-ISDN prototype with separate lines for audio, and speaker id and queuing data via modem, some years back. It turned out to be too complex for practical use. A version of my equal-time resolution rule was programmed into that system (Stodolsky, D. (1987). Dialogue management program for the Apple II computer. _Behavior Research Methods, Instruments, & Computers_, _19_, 483484.). This rule has been show to yield benefits in both emotional tone and group performance in controlled experiments. I would like to see the rule applied in one of these new ISDN conferencing systems, but its hard to get the attention of the equipment suppliers on this point. They typically resort to centralized control by a chairmen, without even the ability to run on "auto pilot", where people queue themselves up by pressing a "request" button or just by starting to talk with a voice-operated switch "pressing" the button for them. Central control of speakers was strongly disliked in the prototype system. In fact, all units were eventually rebuilt, so each one could be the "master" in a multi-unit conference. Chairmen management seemed a bit clumsy, even when the queuing was automatic and the chair just announced the name of the next speaker. From a psychological standpoint, fully distributed control is the only way to go, and it is quite feasible with ISDN, any takers? David S. Stodolsky, PhD Routing: <@uunet.uu.net:stodol@diku.dk> Department of Psychology Internet: <stodol@diku.dk> Copenhagen Univ., Njalsg. 88 Voice + 45 31 58 48 86 DK-2300 Copenhagen S, Denmark Fax. + 45 31 54 32 11