Question

Connect no longer bit-perfect?


It looks like the Connect is no longer bit-perfect. Here's my evidence: let's discuss this.

First, I constructed a wav file of pink noise with amplitude ramping up from zero to digital max and back to zero.
I play this through my Connect and record the SPDIF output from the coax output into my PC.
The recording uses a Scarlett 8i6 audio interface set to use the Connect as master clock.
I record into a DAW (Sonar) multiple times - all instances are identical.
However, this recorded signal is not quite the same as the original wav file - it can be up to -21dB different.
See https://www.dropbox.com/s/t8od479xo9hi5el/connect_diff.PNG?dl=1
Note the expanded scale on the difference (third) track.

It looks like the difference gets larger when the signal is larger. To confirm this, I import the
original and difference files into Matlab and plot the raw data (difference vs original). There is clearly audio compression
happening here. See https://www.dropbox.com/s/p1yq6wcqafvnhaj/diff_vs_orig.png?dl=1
The scale is such that digital maximum is 1.

There also appears to be a slight bias when the waveform is negative and the signal is below the
compression threshold. See an expanded version of the previous plot
https://www.dropbox.com/s/9001tl9mkle4wly/diff_vs_orig_zoom.png?dl=1

Happy to answer questions about the method and conclusions.

Cheers, Peter.

p.s. Volume is set to fixed - I haven't tried variable.
In a loopback test (8i6 out from DAW to 8i6 in, no Sonos gear involved), I get bit-perfect cancellation.

This topic has been closed for further comments. You can use the search bar to find a similar topic, or create a new one by clicking Create Topic at the top of the page.

453 replies

It is my understanding that DACs also need an analog filter (called a reconstruction filter) on their output.
There is normally a simple external RC analogue filter as part of the design of the output stages coming from the DAC. Most DACs will give a recommended design in their data sheet. There is the potential to mess the implementation of this filter up, but this sort of filter is well understood and is neither complex nor expensive to implement.

Cheers,

Keith
I wonder where Sonos implement their brickwall limiter? Older Connects don't have it, but software versions prior to 6.0 don't either. So it would seem to need a mixture of hardware and software. If software can activate it, presumably it comes before the DAC, so it is in the digital domain. In which case, why is additional hardware needed?

I presume that the brickwall limiter applies to all Sonos units, not just the Connect. Does this mean it is just software? Or are older Play units immune to it because they lack the right hardware?


I'm making educated guesses here. The brickwall sounds, to me, like it is implemented in the digital domain as part of the compression algorithm. In other words it's DSP based. This could be entirely software based (using the general purpose processor) or hardware accelerated (using dedicated audio DSP hardware). I suspect the latter.

In this case, the implementation later Connect units could point to better DSP capability which, given it is highly probable that the components used in various models of ZP80/ZP90/Connect have changed over the years, shouldn't be surprising.

Cheers,

Keith
Userlevel 7
Badge +21
In this case, the implementation later Connect units could point to better DSP capability which, given it is highly probable that the components used in various models of ZP80/ZP90/Connect have changed over the years, shouldn't be surprising.

So the question that comes now is... how do I tell which one mine is? Maybe we should all dig through the /support/review page and look at Connect details to see if we can find something to identify?
So the question that comes now is... how do I tell which one mine is? Maybe we should all dig through the /support/review page and look at Connect details to see if we can find something to identify?
No need. Look in About My Sonos System. Ryan said that this affected hardware version 1.16.5 or higher.
Userlevel 1
Badge
Well, I said I was finished with this, but I still own a Sonos product, a PlayBar, so as a customer I will chime in again.

This thread has gone way off subject. Here's the thing, I went to my Audio dealer and told the salesperson I wanted something that would transport my lossless 44.1/16 audio files iTunes or FLAC, which I was switching to and from Deezer or similar to my Peachtree amp without modification - bit perfect. I was directed to the Sonos Connect, which, at the time did just that. I was not interested in the Connect's analog output. I was having nightmares with Apple's Airport Express, having to reset it frequently, though it did transport the files bit perfect when it worked.

Now this is key - it doesn't matter if anyone thinks I should be happy with a Connect and a house full of Sonos speakers. It doesn't matter that the Connect has a good interface and can do global searches. That is not what I was sold, that is not what I had and most certainly not what I wanted and paid for.

The Connect worked fine for me until I moved in September 2015 and did not get my system reconnected until the end of 2015. It was in a smaller room with hard walls instead of the log walls of my house. I tried a lot of things to figure out why it didn't sound "right", quilt on the wall behind the sofa and more. Then I see Peter's post that started this thread. Seeing Sonos staff member Ryan's post saying he had submitted a ticket, I had hope. It never went anywhere.

I decided to try something else. I liked the idea of having the files stored on the transport device, so I tried the Auralic Aeries Mini with a 240GB SSD in it that I was not using. It was dramatic. Now "dramatic" is just my opinion, but that's my two cents. My system was back!

I called the audio dealer where I bough the Connect and they were stunned about the firmware changes made in late 2015. They were still selling and shipping them as bit perfect transports. They only carried Bluesound and I had some issues with it so bought the Aeries Mini from Audio Advisors. I told them about this thread and they too were surprised because they were selling them as always - bit perfect transports.

I don't think this hi res, MQA, Sonos speaker, audiophile thing is relavent to this discussion. Maybe Sonos didn't specify the Connects and its predecessors were bit perfect, but they were and dealers did sell them as such and still do. There is at least a breakdown in communication between Sonos and some of their dealers. To me it's almost as if Samsung updated my TV firmware and decided green pixels were no longer important.
There is at least a breakdown in communication between Sonos and some of their dealers. .
Not necessarily; not if - as you also say may be the case - Sonos did not specify the Connects as bit perfect, and if they did not also say this in any way to their dealers in any communication. Your dealer may have a better handle on where he got the bit perfect thing from and it may not be from Sonos.
As far as I know, the bit-perfect reputation of the Connect relies on a single test of its predecessor, the ZP80, found here. That test was published in Sep/Oct 2006.

It would seem the Connect's reputation as a bit-perfect transport is no longer deserved.

My preliminary tests indicate that the analog output of the Connect likely has the same brickwall limiter applied, and by extension, this probably applies to every bit of Sonos gear uniformly. Ryan from Sonos can chip in if this is not the case.
As far as I know, the bit-perfect reputation of the Connect relies on a single test of it's predecessor,
The Atkinson review in Stereophile - is that the only place where this was stated and then relied on in word of mouth?

By the way, I had picked up a very interesting "tell" in that review in this statement of Atkinson:
"It's just a shame, I guess, that these groundbreaking audio products didn't come from an established high-end audio company."
As far as I know, the bit-perfect reputation of the Connect relies on a single test of it's predecessor,
The Atkinson review in Stereophile - is that the only place where this was stated and then relied on in word of mouth?

The tests here provide corroborative evidence. The S/PDIF lowest byte would not be zeroes if the samples had been adulterated.
Arguably, if there is any misselling involved it was from the dealer. However, the dealer was operating on what was, at the time, correct information so I don't think they missold.

As has already been stated, the information on the Connect being bit-perfect doesn't seem to come from Sonos. So no misselling there either.

However, what I think is a bit disappointing is there was an implicit view in the industry that the Connect was a bit-perfect transport and, whilst they didn't specifically make this claim themselves, Sonos were aware of it and did nothing to correct that view when they made the change.

This all makes me think this was a misstep rather than a deliberate strategy and I'm hopeful that, now it has come to light, that Sonos will be doing something to correct the issue.

I'm normally the first person to point out that software developments are not as straightforward as many people claim, but in this case, reverting to a bit-perfect output should be an easy thing to do, as the code is already there and has been for most of the last decade.

Implementing a user-selection may take a little extra effort and testing if they want to retain the ability of Connect's to support volume normalisation under user-control.

Cheers,

Keith
The tests here provide corroborative evidence. The lowest byte would not be zeroes if the samples had been adulterated.
Can you explain a bit more what you mean here Ratty? Corroborative evidence of what? What is potentially doing the adulteration? Sorry but I'm having trouble cluing in to what you mean - perhaps the G&T with dinner 🙂
I have a sneaking suspicion that this may have crept in around the time of Trueplay.

In the case of PLAY:1 the effect of Trueplay's EQ could, when added to the existing EQ to make it sound 'big', on rare occasions result in bass clipping upstream of the volume control. (This was often dismissed as 'hardware fault'. It wasn't: I witnessed this on multiple PLAY:1s requiring the same, possibly pathological, room compensation.)

A quick solution may well have been to implement a soft knee limiter. Certainly something has largely resolved PLAY:1s earlier difficulties. If true, it's a tragic irony that this was carried over into CONNECT, when the unit doesn't even support Trueplay.
The tests here provide corroborative evidence. The lowest byte would not be zeroes if the samples had been adulterated.
Can you explain a bit more what you mean here Ratty? Corroborative evidence of what? What is potentially doing the adulteration? Sorry but I'm having trouble cluing in to what you mean - perhaps the G&T with dinner :)

My point was simply that in those 10-year-old ZP80 tests the samples must have passed through unsullied, otherwise the lowest byte of the 24-bit S/PDIF would have been populated -- as it indeed was when the tester dropped the volume below 100%.
My point was simply that in those 10-year-old ZP80 tests the samples must have passed through unsullied, otherwise the lowest byte of the 24-bit S/PDIF would have been populated -- as it indeed was when the tester dropped the volume below 100%.
OK, I get it now. Took me a while 😛
Userlevel 5
Badge +4
Not sure what models were specifically affected. Sonos would have to answer that.

As to your trying to get this thread back on topic; I have been chuckling for the past couple of days at the outrage directed at me for "derailing" this thread. Outrage which came from a poster who ignored everything about the Connect being bit-perfect and instead shot us off the rails into Hires/MQA land!


I said I didn't understand why the same few people come into every thread where people are discussing getting more from their system and call those people fools.

I said that we could do with Sonos offering an "audiophile" product (Note audiophile for their marketing reasons) that offered Bit Perfect output and with MQA now coming to Tidal and talk of Spotify getting on board it would be a nice option that I am sure they would do incredibly well with.


The only reason I came back on was because I searched for "Sonos Connect Modification" on google to look for something, and what was the first hit?

You, 4 years ago calling any Sonos Connect mods snake oil. :D

https://en.community.sonos.com/advanced-setups-229000/zp90connect-modification-comparison-27830/index1.html#post177195

At least you're consistent.

The whole gang is there too, Kumar et al.



I agree with much of what you say, but there is also loads I don't agree on, but hey ho, such is life, it would be boring if we all thought the same.
Why wouldn't I be consistent? Has something changed about the biology of the human ear, the laws of acoustics, or the Nyquist-Shannon Sampling Theorem in the last 89 years?
Purveyors of Snake Oil haven't changed much either, lol. Heck, we even have a few in the White House now.
Userlevel 5
Badge +4
Why wouldn't I be consistent? Has something changed about the biology of the human ear, the laws of acoustics, or the Nyquist-Shannon Sampling Theorem in the last 89 years?

I didn't say you shouldn't.


You have an incredibly aggressive nature, every post of yours makes you seem like a very angry chap.


I didn't say you shouldn't.


You have an incredibly aggressive nature, every post of yours makes you seem like a very angry chap.


Attacking the messenger is the last bastion of defence for a failure.
Userlevel 5
Badge +4


I didn't say you shouldn't.


You have an incredibly aggressive nature, every post of yours makes you seem like a very angry chap.


Attacking the messenger is the last bastion of defence for a failure.


Jesus wept! 😃

Jesus wept! :D

I agree; even he would have:D

Since you have deigned to return: any comment on the hydrogen audio discussion on MQA that I provided a link to?
Userlevel 5
Badge +4
No. I read through it last night and it is just 'same old, same old'.

I don't enjoy arguing with people, and all you lot want to do is argue. It doesn't matter what I say, I will get jagitate tell me that I'm wrong or I'm a fool so don't see the point.
No. I read through it last night and it is just 'same old, same old'.

I don't enjoy arguing with people, and all you lot want to do is argue. It doesn't matter what I say, I will get jagitate tell me that I'm wrong or I'm a fool so don't see the point.


Wow, I didn't think we'd be seeing the victim card so early from you. I honestly had higher hopes.

And the childish respelling of my username? Nice touch. I'm sure you can find a large following on gaming console sites. 😃
Userlevel 5
Badge +4
No. I read through it last night and it is just 'same old, same old'.

I don't enjoy arguing with people, and all you lot want to do is argue. It doesn't matter what I say, I will get jagitate tell me that I'm wrong or I'm a fool so don't see the point.


Wow, I didn't think we'd be seeing the victim card so early from you. I honestly had higher hopes.

And the childish respelling of my username? Nice touch. I'm sure you can find a large following on gaming console sites. :D



If you weren't such an opinionated bigot I would bother, but what's the point?

I know for a fact I'm not the only one who thinks this because of the PMs I have had from other members telling me they have just learnt to ignore you or don't post in the threads you post in, although pretty hard with such a post count as you do seem to be in almost every thread.

Find a large following? Is that what you think people want? Is that why you have nearly 13000 posts on here? To get a following.
Oh dear.
Wow. Do you even know the meaning of the word you're so casually using? Highly offensive, and far from true in this case.