What does Format Conversion do to digital audio?

[Jeremy H]

Hi Guys,

If you maintain sampling frequency and bit rate - is converting a digital format (say spdif to aes/ebu or vice versa) a critical audio process as is ADA or dithering?


Or to put it another way, should I expect a format conversion to potentially degrade the audio as much as a second full ADA process might?


cheers,
Jeremy

[otto ruiter]

Format conversion when the bit depth and sample rate are not changed will not alter the audio quality at all, however the format you convert to may not be as robust as the original, depending on what you convert. E.G. an optical link (toslink, ADAT) may have higher jitter at the receiving end than a copper link such as AES/EBU or SPDIF.

[Jeremy H]

A belated "thanks" for that Otto. Exactly what I needed to confirm.
Jeremy

[chris p]

Hi Otto, could I please seek a clarification?

[quote]the format you convert to may not be as robust as the original, depending on what you convert. E.G. an optical link (toslink, ADAT) may have higher jitter at the receiving end than a copper link such as AES/EBU or SPDIF.[/quote]

My limited understanding is that the optical cable is simply sending a stream of 1's and 0's, just like the copper cable. The receiving unit then assembles the bit stream into an actual audio "word" of 16, 20 or 24 bits as appropriate, and then when it receives a clock impulse (generated internally, from an external word clock or from the timing signal within the ADAT stream) bangs it out to whatever (usually either a DA chip and then forth as an analog signal, or perhaps through firewire to some poor overloaded CPU).

Jitter arises when the clock impulses are not evenly spaced, smearing the audio signal in time (and we all know that timing is very important for spacial imaging, so even sub-millisecond smearing is a BAD THING

[Thirteen]

It's generally held that clocking via ADAT optical should be a second choice if the two units can be clocked together via BNC wordclock, and the optical path is left for data only.

[smash]

From what I understand, the basic problem with an Optical transfer, as opposed to an Electrical one, is "Echo".

Basically, the light goes into the cable - reaches the other end, and say 99%
of the light goes out, the other 1% is reflected back into the cable, which can the re reflect back at the opposite end and end up as a second, delayed pulse.

Becuase we are dealing with light here, it is not a "delay" but it comes across as Jitter in the signal.

The higher quality the Optical Cable, the higher the transmission.

Also check out http://www.monstercable.com/productdisplay.asp?pin=3863

NB - not a plug for Monster, just backing up my post.

[mal stanley]

Optical is subject to jitter much more than other formats (eg: AES) if you rely on it to also carry the word clock....as long as you word clock separately you are fine with audio data only on optical..

[JulienG]

Optical is subject to jitter much more than other formats (eg: AES) if you rely on it to also carry the word clock....as long as you word clock separately you are fine with audio data only on optical..

My guess would be that if you used proper multimode glass fibre cables jitter would be reduced, however glass fibre has to be matched fairly well or else you just end up with a bigger problem.

Yes, I know I've just brought up the whole digital cable quality thing, but having done the plastic fibre / glass fibre thing at the day job (in telecoms) I'd go the glass every day.