V.61: talking and sending data on one carrier (ASVD)¶

V.61 (ITU-T V.61, "ASVD", analogue simultaneous voice and data) lets two modems carry a telephone conversation and a data connection at the same time on one call. It was approved in 1996, alongside V.70's DSVD, and works quite differently. DSVD compresses the voice and sends it as data. ASVD sends the voice as an analogue signal, added straight onto the modem's own carrier, around each data point. The data runs at 4800 bit/s while someone speaks. When both people are quiet, the modems switch to data alone at up to 14400 bit/s. It grew out of AT&T's VoiceSpan modems.

D-Modem has a V.61 data pump of its own, written from the Recommendation. Two D-Modems agree on ASVD in V.8bis, run V.61's start-up, and connect. The computer's data goes over V.42 as on any call. The speech comes from the same place as V.70's: a SmartUSB56's handset, or a program of your own with dmodemd's --dsvd-audio.

For the full technical description, see the technical details.

Turning it on¶

Both ends need V.8bis and ASVD on:

AT+A8E=1,1,,1+ASVD=1

+ASVD is D-Modem's own command: 1 offers ASVD in V.8bis and chooses it when the other end offers it too, 0 (the default) does not. AT+ASVD? shows the setting. With ASVD at one end only, the call is an ordinary data call.

When both ends offer V.70's DSVD as well (+DSVD=1), DSVD is chosen: it carries more data beside the voice on a good line.

AT+MS=V61 (or AT+MS=61) runs V.61 straight away, without V.8bis. Both ends need it, since no other modem standard falls back to V.61.

What the user sees¶

ATD and ATA work as usual. CONNECT reports 4800 bit/s when both ends have a voice path, as the call can always carry that much. While the line is quiet the data goes faster than that. Without a voice path at either end, CONNECT reports the full data-alone speed.

Where the voice goes¶

As for V.70:

• A SmartUSB56 with a handset (--usb-voice-codec): the phone on the modem's handset jack.
• --dsvd-audio <command> (dmodemd, not on Windows): a program that plays the far end's speech and records ours, 16 bit samples at 8000 a second.

With neither, the modem tells the other end it has no voice and every frame carries data alone.

Speeds in simulation¶

Two D-Modems back to back on a simulated phone line (pump_test, dsp_loopback), without speech:

With a tone on each side's microphone, the data runs at 4800 bit/s without a bit error down to 20 dB, and each side hears more than 90% of the other's tone. All of these hold at both of V.61's symbol rates (3000 and 2800 a second), at 8000 and 9600 samples a second, with clock offsets of 300 ppm, and with near and far echoes.

Limits¶

• Only between two D-Modems so far. No other maker's V.61 modem has been tried.
• The voice has not yet been tried on a SmartUSB56's handset itself, only in tests.
• The audio pre-emphasis filter is fixed at its flat setting. V.61 lets a modem choose another per frame to suit the speech.
• The rate renegotiation of clause 11 is not done. A call whose line gets worse retrains instead.
• The auxiliary control channel (up to 342 bit/s in the control segments) is not used.
• ASVD is only agreed at the start of a call, not in the middle of a telephone conversation.