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.
Page 10 / 19
You are guessing. ZP80 (from which CONNECT is descended) was the second ever 'ZonePlayer' after the ZP100. Specifically designed to link to existing amplification, with or without an outboard DAC. It's been around since 2006. Hardly an afterthought.
Can someone explain the hardware difference between the ZP90 and the Connect? Might there be a simple modification on the connect to bring it back to the ZP90 standards?
You are guessing. ZP80 (from which CONNECT is descended) was the second ever 'ZonePlayer' after the ZP100. Specifically designed to link to existing amplification, with or without an outboard DAC. It's been around since 2006. Hardly an afterthought.
Okay, bad guess, but the issue remains.
If you truly believe the latter part, sell the kit and move on if you are an "audiophile" and wish to remain so.
For other readers here, this might be of interest:
"Audiophile" is a very fuzzy word. One audiophile could find that the Connect is not worthy of being audiophile kit unless it was attached to an external DAC costing a lot of money. Another could find that even then, it is not worthy of being audiophile kit unless it was modified by changing its internal componentry - there are people that offer this service and audiophiles that buy it from them even today, at the cost of loss of reliability.
And some ex audiophiles - I was one for over a decade, dabbling around with heavy valve amplifiers, external DACs, expensive cables and monster speakers - find that a play 1 pair + Sub can beat any audiophile kit for sound quality at up to 2-3 times the price point of the former. With kit like the various play units, I find that the Connect is now obsolete from a features point of view - but not from that of sound quality.
Depending therefore on what kind of audiophile you are, Sonos kit is/isn't meant for you.
That the Connect has more capable hardware than the zp90 ought to be a given. Why then can the Connect not do the bit perfect thing and the volume normalisation thing is a mystery to me. Also a mystery is why, if both aren't possible, Sonos prefers to offer volume normalisation over bit perfect.
Unless, not having ever specified "bit perfect", there was no need for them to stay with that if they think that normalisation is of more value than bit perfect to their target market.
I can't hear any differences between my zp90 and my play 1 units on how effective they are at normalisation, so the point is moot to me. I also prefer the sound quality from my 1 pair + Sub to my zp90 zone, but that is subjective preference, I admit.
The other thing that is easy enough to do, but I doubt anyone will, is even a reduced protocol blind test between zp90 and Connect, to see if there are audible difference between the two once the usual biases that afflict all humans are ruled out. Wiring each to different analog input jacks of the same amp, and supplying the same source to both Sonos units makes it easy to achieve the two conditions that are not always easy to achieve at home: ruling out variability in source and sound levels. Toggling back and forth is also easy to do via the controller app. Perhaps we will see this rare event here soon!
Unless, not having ever specified "bit perfect", there was no need for them to stay with that if they think that normalisation is of more value than bit perfect to their target market.
I can't hear any differences between my zp90 and my play 1 units on how effective they are at normalisation, so the point is moot to me. I also prefer the sound quality from my 1 pair + Sub to my zp90 zone, but that is subjective preference, I admit.
The other thing that is easy enough to do, but I doubt anyone will, is even a reduced protocol blind test between zp90 and Connect, to see if there are audible difference between the two once the usual biases that afflict all humans are ruled out. Wiring each to different analog input jacks of the same amp, and supplying the same source to both Sonos units makes it easy to achieve the two conditions that are not always easy to achieve at home: ruling out variability in source and sound levels. Toggling back and forth is also easy to do via the controller app. Perhaps we will see this rare event here soon!
The other thing that is easy enough to do, but I doubt anyone will, is even a reduced protocol blind test between zp90 and Connect, to see if there are audible difference between the two once the usual biases that afflict all humans are ruled out. Wiring each to different analog input jacks of the same amp, and supplying the same source to both Sonos units makes it easy to achieve the two conditions that are not always easy to achieve at home: ruling out variability in source and sound levels. Toggling back and forth is also easy to do via the controller app. Perhaps we will see this rare event here soon!
My active monitors only have one analog input... But maybe I'll be able to pull off such a test with a rca switch :)
(and also cable splitters from the source etc)
That being said, I tend to trust objective measurements more than listening tests, which can be fickle. The best ways of identifying differences in blind tests are typically with either speech or pink noise, but we typically tend to listen to music and not pink noise on our stereos. And when listening to music, the brain usually fills in things that might be missing objectively, or that we expect to be there. When comparing two tracks of audio next to each other, the brain can "smear" the listening impression of one track over to the other for example. The very reason that sighted listening is so unreliable also makes blind listening somewhat unreliable. There's no easy way out, and no getting around that our brain and our hearing apparatus are very unreliable instruments.
But if there can be shown to be a measurable degradation within a range that we know is audible to human beings (and not only to bats or dogs for example), then I'll just take that at face value. I just want my system to perform as good as it can within the human hearing range (so I don't care for hi-rez files for example). That's why I employ active monitors without passive crossovers and with very good drivers, have a large subwoofer (bigger bass drivers have less distortion), avoid vinyl like the plague, listen in the near-field, and buy the cheapest cables I can find on eBay. For me it's all about objective performance, and getting as much bang for the bucks as I can. This improves my subjective appreciation of the music. I think its partly because of the objective quality of the sound, and partly related to my smug self-gratulatory satisfaction over knowing that I have good sound ;)
But I'll see if I'll be able to arrange a blind test with the two connects!
Lol. I prefer the short route these days; if it sounds good to me and if I enjoy the music/experience, I don't bother about objective measurements to confirm or in any way validate my experience! Perhaps this is the result of a decade of equipment obsession that I am happy to have left behind.
The blind test would be of interest, one way, or the other to those that are hurting by the loss of bit perfect. But how would they get over the knowing of the loss of bits even if the blind test proves that the loss isn't audible?:D
And if you do get around to doing the blind test, remember to also see what, if any, are the differences in normalisation effectiveness between the two units.
No one said it couldn't. It just doesn't because it's not been programmed this way. There's no technical reason I can think of that it couldn't (other than the normal issues of resourcing and timescales).
I have always presumed this was likely to be some sort of error/oversight. It may, of course, have been deliberate, but the ramifications weren't considered. Other possibilities exist.
I don't think that both aren't possible. However, not being possible is not the same as not being possible within the current development window, or roadmap, etc.
Consider that vastly more people will benefit from volume normalisation than from the Connect being bit-perfect. This is especially true on streaming services where the source is far from "bit perfect" in the first place. I would hazard a guess that less than 5% of the music played on Sonos systems across the world is from lossless sources.
I'm not saying getting rid of the option for bit-perfect output is the right thing. Personally I would have kept it as an option (and, maybe, they will in the future).
All assumption/speculation on my part, but offered as food for thought.
Cheers,
Keith
Btw, I agree that two play:1s plus sub, equalized with trueplay, is almost laughably good. I have that setup in my kitchen. I do prefer the other setup I have in the salon by some degree, but the sonos setup really is superb.
Something to think about - could that preference also be caused just by the fact that you don't have an objective measurement set for the 1 pair + Sub?! Extending that question further, a 5 pair + Sub, that is physically closer to your AVI set up, also is something that you will never prefer to the AVI only because objective measurements for this also are not available?
Sonos never publishes much by way of detailed specs and there is little out there by way of objective measurements for their kit either, a state of affairs that Sonos and its target market seems to be quite comfortable with. I get that you need these, the questions are just something to ponder:).
No one said it couldn't. It just doesn't because it's not been programmed this way. There's no technical reason I can think of that it couldn't (other than the normal issues of resourcing and timescales).
I have always presumed this was likely to be some sort of error/oversight. It may, of course, have been deliberate, but the ramifications weren't considered. Other possibilities exist.
I might have misunderstood, but I thought that the bit perfect change was also applied to Connects already sold, via a software upgrade that everyone gets from time to time. If that was the case, how come it was not applied to zp90 as well, if, as you say, that is just as capable as a Connect in being able to accommodate the changed programming? Are you suggesting that not applying the change to the zp90 was also an error, in addition to the error of not considering the ramifications of the bit perfect thing?
And based on what some people here have reported that this change has caused night and day changes in sound quality, it should be very easily audibly noticeable in a zp90 v Connect stand off. Somehow, I doubt that to be the case...something that only a good blind test can decide.
Of course, confirming that there isn't a day/night difference won't need that sophisticated a test - the null hypothesis thing.
It's not a night and day difference. That would be a huge exaggeration. Maybe sunny and partly cloudy. I might have even used the term night and day somewhere way back in this thread.
I'd did compare Apple lossless connected digitally to my amp by my Connect, an Auralic Aries Mini and the original CD on an Oppo DV981 also connected digitally. My Peachtree amp has two optical inputs and an AUX input. I could not tell a difference between the Aries and the Oppo, but the clarity of the Connect was obviously degraded. I had to adjust the volume up for the Connect and this made instantaneous comparison impossible. Must have something to do with the volume normalization. Was this THE reason for my perception of clarity? I can't be sure, but I don't think so because my perception of something being wrong started before I started investigating why. I should have removed the normalization tag and tried that, but didn't think of it at the time. So, those who might compare a Connect to a Z90 might want to rip an Apple lossless using something other than iTunes or remove the tag using MP3TAG and see if that makes a difference. iTunes will put the tag in the file even if you have it turned off. The iTunes setting only applies to when you are playing. Apple assumes you might change your mind about volume normalization even if you have it off.
I'd did compare Apple lossless connected digitally to my amp by my Connect, an Auralic Aries Mini and the original CD on an Oppo DV981 also connected digitally. My Peachtree amp has two optical inputs and an AUX input. I could not tell a difference between the Aries and the Oppo, but the clarity of the Connect was obviously degraded. I had to adjust the volume up for the Connect and this made instantaneous comparison impossible. Must have something to do with the volume normalization. Was this THE reason for my perception of clarity? I can't be sure, but I don't think so because my perception of something being wrong started before I started investigating why. I should have removed the normalization tag and tried that, but didn't think of it at the time. So, those who might compare a Connect to a Z90 might want to rip an Apple lossless using something other than iTunes or remove the tag using MP3TAG and see if that makes a difference. iTunes will put the tag in the file even if you have it turned off. The iTunes setting only applies to when you are playing. Apple assumes you might change your mind about volume normalization even if you have it off.
Yours are, so why concern yourself with mine?
Because it isn't just you that will read what I have to suggest, others read this forum as well, and some may well find it making sense to them.
I'd did compare Apple lossless connected digitally to my amp by my Connect, an Auralic Aries Mini and the original CD on an Oppo DV981 also connected digitally. My Peachtree amp has two optical inputs and an AUX input. I could not tell a difference between the Aries and the Oppo, but the clarity of the Connect was obviously degraded.
Is this a general statement on the Connect, or is it that the degraded clarity from the Connect you refer to arose after Sonos did their thing to it, and it wasn't the case till then, with the sound on par with the other two sources? And that you noticed this before you came upon this thread - I ask this because some have admitted that to be the case.
I'd did compare Apple lossless connected digitally to my amp by my Connect, an Auralic Aries Mini and the original CD on an Oppo DV981 also connected digitally. My Peachtree amp has two optical inputs and an AUX input. I could not tell a difference between the Aries and the Oppo, but the clarity of the Connect was obviously degraded.
Is this a general statement on the Connect, or is it that the degraded clarity from the Connect you refer to arose after Sonos did their thing to it, and it wasn't the case till then, with the sound on par with the other two sources? And that you noticed this before you came upon this thread - I ask this because some have admitted that to be the case.
From 12 days ago when I said:
"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. "
It still hasn't. I have. Connect is gone now, after comparing it to the Aries.
People use lots of words to describe how one speaker, CD, or any source of sound differs when compared to another. I used clarity, I could use crisp, maybe a lot more, but that's me. To say the Connect was like tossing a blanket over the speakers would be a bit much, but along those lines. My imperfect ears, but even imperfect ears can detect differences.
I was searching on the site, and everywhere, trying to find out if others had issues and found this thread. The above comparison I made was after giving up that Ryan S would have any influence for an option to bypass the changes that had been made that Peter Mc explained in the post that started this thread.
The only comparison made before was with a CD and my own Apple Lossless file through a Connect optically connected to a system costing about a half million at the Kettering, Ohio dealer that sold me the Connect. I couldn't tell a difference. I think it qualified as a true audiophile system. The silver ribbon cables cost more than I spend on most cars. Took it home and with a Marantz amp and my Belle Klipches I had owed since 1976 (reluctantly sold for about what I paid for them due to the upcoming condo move and replaced with a pair of Dynaudio 1.8 Mk II speakers sold to me by friend) it sounded great, and with the Dynaudio speakers too, but I couldn't "entertain" the entire neighborhood.
Lol, good idea.
I believe that my post is relevant to this thread because regardless of what may have happened to it since 2011, my Connect which I also have, with its analog outputs in use to drive a third party amp + speakers, has caused no change in the way the music from my non Sonos speakers sounds. I agree that my post may not be relevant to those that are not concerned about how music sounds, but not everyone that is on this forum thinks that way, many of us are here to listen to the music without dissecting or otherwise parsing the equipment that delivers it.
On the other hand I also understand that those that want transparency for its sake, for that being the way they want the Connect to work, may want to change brands if Sonos does not restore it to them.
I believe that my post is relevant to this thread because regardless of what may have happened to it since 2011, my Connect which I also have, with its analog outputs in use to drive a third party amp + speakers, has caused no change in the way the music from my non Sonos speakers sounds. I agree that my post may not be relevant to those that are not concerned about how music sounds, but not everyone that is on this forum thinks that way, many of us are here to listen to the music without dissecting or otherwise parsing the equipment that delivers it.
On the other hand I also understand that those that want transparency for its sake, for that being the way they want the Connect to work, may want to change brands if Sonos does not restore it to them.
The "tubes (valves) sound better" audiophiles should love this soft knee limiting, as it's the same thing tubes do when approaching maximum power output. I remember when NAD marketed their "soft clipping", which is also the same thing. The audiophile press loved it. I had one, the circuit failed. Simple to fix, as it was just a resistor and 4 diodes, lol.
http://sound.whsites.net/articles/soft-clip.htm
http://sound.whsites.net/articles/soft-clip.htm
http://sound.whsites.net/articles/soft-clip.htm
No one was ever accused of inaccuracy when overestimating the hypocrisy of the audiophile press. Now excuse me while I try to get the "orange peel" sound out of my amplifier. :8
I might have misunderstood, but I thought that the bit perfect change was also applied to Connects already sold, via a software upgrade that everyone gets from time to time.
Yes, it would have been, subject to the hardware being capable.
If that was the case, how come it was not applied to zp90 as well,
Because the ZP90 clearly doesn't have the capability.
if, as you say, that is just as capable as a Connect in being able to accommodate the changed programming?
I didn't say that at all.
I only said I saw no reason why recent Connect units could not support both "bit-perfect" and the non-bit-perfect normalisation mode. We know for a fact that the Connect is capable of supporting both modes as both modes have been seen in the field, just not at the same time.
Are you suggesting that not applying the change to the zp90 was also an error, in addition to the error of not considering the ramifications of the bit perfect thing?
No, I simply believe the ZP90 lacks the hardware resources to support this change.
Cheers,
Keith
From 12 days ago when I said:
"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. "
It still hasn't. I have. Connect is gone now, after comparing it to the Aries.
People use lots of words to describe how one speaker, CD, or any source of sound differs when compared to another. I used clarity, I could use crisp, maybe a lot more, but that's me. To say the Connect was like tossing a blanket over the speakers would be a bit much, but along those lines. My imperfect ears, but even imperfect ears can detect differences.
I was searching on the site, and everywhere, trying to find out if others had issues and found this thread. The above comparison I made was after giving up that Ryan S would have any influence for an option to bypass the changes that had been made that Peter Mc explained in the post that started this thread.
The only comparison made before was with a CD and my own Apple Lossless file through a Connect optically connected to a system costing about a half million at the Kettering, Ohio dealer that sold me the Connect. I couldn't tell a difference. I think it qualified as a true audiophile system. The silver ribbon cables cost more than I spend on most cars. Took it home and with a Marantz amp and my Belle Klipches I had owed since 1976 (reluctantly sold for about what I paid for them due to the upcoming condo move and replaced with a pair of Dynaudio 1.8 Mk II speakers sold to me by friend) it sounded great, and with the Dynaudio speakers too, but I couldn't "entertain" the entire neighborhood.
This is pretty much what brought me here.
I had thought for a while I had an issue with my speakers, Meridian DSP5200s, they are now 10 years old, I know them inside out.
Recently music sounded flat, music I know well. When I say flat I mean lack of dynamic range, now, this is not on every track, I could get into someone like Diana Krall or Leonard Coen for a few weeks and not really notice it, but then on some classical pieces with big dynamic ranges or on a particular dance track I was sent with a massive dynamic range it was sounding 'flat' on my speakers, where as on the iPhone and a pair of £29 Sennheiser in ears it sounded mental.
I thought my speakers were playing up, but then I borrowed a Meridian MS200 (a bit like a connect) so I could test our Roon, all of a sudden my system came back to life.
I didn't even consider that anything had changed on the connect side of things, I just thought that the connect was average at best, hence why I came on the forum, to see if anything else was in the pipeline that maybe was better quality. I then found this thread and posted in it.
Now, if it is a case that the Connect going back to being bit perfect will bring my system back to life I will be a happy bunny.
Is Volume normalisation a good thing?
Can a system that does it on the fly not reduce dynamic range, therefore not alter the actual track itself?
It would be nice if we had an option to turn it on/off, simply allowing the system to disable it when you chose 'fixed output' would be good enough.
Is Volume normalisation a good thing?
Yes, based on my understanding of it, which is that normalisation should be a change effected in output levels across the entire dynamic range of a track by moving it in its entirety to the left or right as happens when I move the volume control slider to change volume. Relative differences between low and high should not be altered for this feature to be more than just compression done such that the lows can be heard without the highs going too loud. It ought to just have different songs in a playlist be delivered at the same median sound levels relative to each other, not by just having the loud part of every song pegged at one level across. If one of the songs in the playlist has exceptionally loud highs compared to the rest, it should sound to be so; or, if it sounds the same as all other songs, the lows on that song should go a lot lower than those on other songs.
Since my units are zp90s, I have nothing to say on what route Sonos has taken and on whether their claim that it results in a very subtle change in the sound is valid or not. And of course, it is very unlikely that we will see a ABX test between a zp90 and late model Connect to know more about this.
On the subject of Connect sound quality in general, there is a parallel discussion I was engaged in on another thread with a member that finds a day and night difference between music from Connect via its DAC and when used with a Chord DAC downstream. It is of course possible that his Connect is one of the rare ones that is still bit perfect, but that is probably not the case. It is also very likely that his views will change once he comes across this thread.;)
PS: If what Sonos has done is to reduce dynamic range, how is this different from what the recording industry is accused of doing in the context of "Loudness wars"? Somehow, I doubt that this will found to have been the case when audible results are objectively evaluated.
Note that a compressor reduces the volume of the loudest samples, which then allows the mastering engineer to increase the overall volume without the loudest parts of the music clipping. A limiter just reduces the volume. Of course, then you can turn the volume up yourself :)
I hope this answers your question.
Cheers, Peter.
Yes and no! You say just before quoted that Sonos has used a limiter, not a compressor. But in your quote you say that all the limiter does is reduce volume. But if that is the case, it should do so across the board of the dynamic range of the song, if, as you also say, all it takes to counter that is turning up the volume. To then level match against a zp90 for example, I would have to turn up the volume on your Connect, or reduce it on my zp90. Then there will be no reason for an audible difference to exist. Right? Or where is my logic incorrect?
And on the quoted above, with respect to what you think has been done to every other speaker:
This should result in every speaker, from the 1 to the 5, sounding strangled at the louder end of music with high dynamic range, regardless of the volume control setting on the speaker, compared to every other speaker out there where the source has not been so dealt with.
Why is it that no one anywhere has pointed out this as a specific criticism/drawback of the Sonos play speaker line up? Not even ratty as an example of one who is much more knowledgeable than I, who uses a 5 + Sub set up now, replacing a zp80 + third party active pair/Sub.
To say nothing of those that still swear by the Connect, but only if it has an expensive DAC hanging on its digital output.
it is highly possible that a 1dB truncation of the loudest music would also influence such a test.
Or is it the case that this reduction has been applied at the loud end of the dynamic range when the volume control is at the highest end of the scale, as in fully to the right, as well?
If so, is there a volume control level below which this has not been applied? If so, what is that level?
The same question to this quote: At what volume levels on the Connect - 100%? And is this at the high end of the dynamic range? If so, this means that there is a 1dB reduction in volume that is applied to music with high dynamic range, to the loudest parts of it, when played at 100% volume. Since I never play my play units in that mode, this isn't the extent of reduction dealt out to me, and this would also be the case for almost every Connect user. And definitely to everyone using a headphone!
Or is it the case that it would be present when the Connect is in fixed output mode, with the volume levels controlled by the downstream unit, but not be applicable when the Connect is used to control volume levels via variable level mode? If this be the case, again, it would also not carry over to Sonos speakers except when running at 100% volume on the loudest parts of music with high dynamic range.
So my question boils down to whether this issue is just a notional and theoretical one for every one other than those using Connect in fixed output mode?
If yes, then the fix is also easy for the remaining set of Connect users: switch to variable level mode.
Page 10 / 19
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.