Port bit perfect?

Lol. That made my day, seeing this in the midst of the firestorm. At this time no one cares; but my guess is that it is no different from it's immediate predecessor, the Connect. And not the zp90, that one preceded Connect.

Actually, a little known fact is the Connect bug that caused the non-bit perfect thing was fixed a while ago.  As expected, nobody noticed.  

Yeah, good to see the usual suspects - concerns I mean, not people. Hi res is missing will hopefully surface soon, so we can think business as usual.

You are a goldmine of interesting information! If I didn’t know differently, I’d simply assume you have access to the Sonos codebase.

I captured the S/PDIF data and analysed it. 

I care. That makes one of me.

And is it identical to what is going in? The original question was about whether the Port is bit perfect. The Connect had volume limiting and a one-bit offset.

Cheers, Peter.

Interesting. Where is this documented?

It’s not.  A little bird told me.  That’s all I can say. 

That I can’t confirm absolutely. However any modification to the signal in Fixed Volume ought to show up in the lowest byte being non-zero.

With Variable Volume there does appear to be some soft knee limiting as the samples approach full scale.

Is this still true if only the high volumes are being modified (as in a volume limiter)?

I’ll hazard an answer. If the volume limiter has a non-zero attack or release time, it could reduce very low amplitude signals adjacent to the peaks, thus pushing them into the lowest byte. Otherwise, no.

I did some tests with single frequency tones at full scale and a few steps below. 

With Variable Volume there appeared to be evidence of limiting above about 85% volume when fed an input containing full scale samples. Some odd harmonics began to appear in the frequency domain. This looks akin to the ‘soft clipping’ technique which, with real music, is a lot less damaging than allowing hard clipping. I guess EQ could in theory push samples above full scale otherwise.

I wonder: does this mean *both* generations of the Connect are bit-perfect (again)? I have been reading the forum sporadically over the years and recall reading that the Connect Gen2 had a change in hardware that does not allow bit-perfect output.

To provide some context: I currently use a ZP90 (Connect Gen1) connected to a simple DAC for my main system, and a number of Play:1 / One SL units in other rooms. With the S1->S2 upgrade I have been looking into my options to keep the system up to date (i.e. run S2).

The Port is very expensive for what it is, and I’m not sure what Sonos were thinking by omitting any control buttons. I like the Amp, but am concerned about its sound signature, and I have no immediate desire to replace my amplifier. That leaves the option of getting a Connect Gen2 while I can, and hoping it will remain current for a good number of years to come. But before buying what is essentially the same device for the second time, I’d like to at least know that I’m not taking a step back..

Thanks!

Simon

Yes, all generations of the Connect are bit-perfect when set to fixed volume.

Thanks! That’s good to know. I just ordered a 2nd gen Connect in the UK for £200 (half of the RRP of the Port).

I'm not sure to understand everything. Simple question : Is Sonos Port bit perfect on digit out yes or not ?

See the tagged answer at the top.

Question: does bit perfect in fixed mode mean that it is as transparent as any other source out there, when supplying digital out to an external DAC? And that would mean that in that mode it will sound exactly the same as the Connect? Or, indeed, as any other streamer, never mind the price?

Or is bit perfect a different attribute?

Bit-perfect means it delivers the stream unmodified from the source.  For every bit in, the exact same bit comes out.

More specifically, as per the tagged answer, it means that what comes out of the decoder is passed straight through without modification. No volume adjustment, no limiting, no EQ. 

So where the encoding is lossless it should indeed mean that what comes out as PCM matches what went in as PCM. 

So there should be no difference heard between Port and Connect where both are in fixed volume mode supplying a digital stream to an external DAC. Or compared to any other streamer at any price point that is also bit perfect, operating in the same manner. Barring defective units of course.

I can understand that there could be audible differences, however unlikely these days, where the internal DAC is in use to supply analog signals, where EQ is set to flat in both cases. And specifically, it seems very unlikely that Sonos would have done anything to have the Port sound different(worse) than the Connect to extent of the claims that are flying around. But that is for another thread.

Hence the incredulity when people claim gross subjective differences between the Connect and Port when using an external DAC.