Why do you need PVDMs?

Question

Hi all,
Dumb question!
Why do you need PVDMs in a NM-HDV2-1T1/E1? The voice from the IP phones is in digital form, and so surely is the voice coming in over the T1/E1 - why do you need these DSPs?
thanks,
alec

Answer 1

You could almost call them transcoders in this case.
The default g711 codec for a given country (g711ulaw for North America, g711alaw for most everywhere else including I think the UK) is identical or nearly identical to that used on PRIs or clear-channel channelized circuits. If the g711 variant you're using for your VoIP traffic isn't the same as the g711 variant for PSTN circuit you're using, the DSPs handle that transcoding. If you're using robbed-bit channelized circuits, the DSPs have to drop a little bit of information. If you're using g729 or another low-speed codec, the DSPs have to "expand" the audio up to the PSTN channel size. And of course, the reverse of all this stuff takes place when audio comes in from the PSTN and has to be packetized and sent out on your IP network.
The DSPs are also in charge of getting the right data to the PSTN channel at the right time. T1/E1 lines have strict timing requirements and the DSPs handle transmitting bits at just the right times. Part of that is handling the de-jitter buffers that turn possibly jittery RTP traffic into a nice isochronous bitstream for the PSTN circuit, and doing predictive insertion if we don't get voice data in time to put it on the circuit.
One more duty the DSPs handle is echo cancellation from the PSTN. They "memorize" outgoing voice traffic, and then they listen for it coming back from the PSTN. If it's present, they mathematically "subtract" it from the incoming voice stream, which eliminates it.
This is not a complete list of what the DSPs do for you, but I think I hit the highlights.