Hi-Resolution Audio and Sonos



Show first post

172 replies

I don't think anyone is disputing that there is a huge range of qualities in audio equipment. I have no problem with your desire to listen in a way that allows you to hear every element, every nuance of tone and feeling. That's your business.

But this thread is about whether it makes sense for Sonos to add HiRes to the system. And HiRes, and even 16 bit lossless, are off the radar of 99.999% of Sonos' target market, I would suggest.
So that would be a big "NO" on performing your test using foobar2000's A/B/X function @nevalti?

So tell me again, who exactly is refusing to "LISTEN"?
nevalti,

One detail about your testing methodology that I forgot to inquire about: How did you match levels between the various formats and the live performance? In a testing session humans tend to prefer the louder presentation -- even if they are otherwise identical. Tiny level differences that are very difficult to accurately determine (subjectively or analytically) will sway the listening jury.

---

With respect to attempting to send hi-res audio to multiple rooms, I recently attended a seminar laying out techniques for distributing 2k, 4k,and 8k video to multiple rooms over a LAN. There are similar, sometimes ugly, discussions about the need for 4k or 8k, display size, color space, frame rate, data compression, viewing distance vs the ability of humans to discriminate the difference.

In the seminar we were concentrating on raw data handling techniques, not subjective differences. While the quantity of data required for hi-res audio is nowhere near that required for transporting multiple high resolution video streams, it was clear that we would need better than average quality networking support for reliable hi-res multi-room audio support. In addition to the need for more sophisticated networking hardware, system layout and setup are critical too. Currently, this sort of network is beyond the capability of the typical DIY installer. (and much more expensive)

SONOS is designed for high reliability on home networks -- which are typically haphazard at best. The additional data required for multi-room hi-res would stress the typical home network -- resulting in much louder complaints of "doesn't work" than complaints of "isn't hi-res".

"Wi-Fi 6" is beginning to roll out. In a few years it will become pervasive and we could re-visit the practicality of wireless multi-room hi-res home audio. Note that every Wi-Fi gadget in the home will need to be replaced in order to fully realize the potential of Wi-Fi 6. By then, 10G networking will become more common too.
Badge
I'm pretty sure it's because the number of "I keep getting dropouts on my Sonos!" calls/complaints would shoot up due to the extra bandwidth required - especially for systems just on the edge now. Which doesn't benefit Sonos one iota.

5 years ago that would be a totally valid complaint. I just visited my girlfriend's parents who are farmers, they have fiber internet in the middle of nowhere.

I'm pretty sure it's because the number of "I keep getting dropouts on my Sonos!" calls/complaints would shoot up due to the extra bandwidth required - especially for systems just on the edge now. Which doesn't benefit Sonos one iota.5 years ago that would be a totally valid complaint. I just visited my girlfriend's parents who are farmers, they have fiber internet in the middle of nowhere.

Internet bandwidth is not the issue. Witness the fact that people routinely stream 4K TV. It's the requirement to pump multiple streams around the house over the average DIY wireless network, whilst maintaining sub-millisecond sync, which could be compromised by potentially increasing per-stream bandwidth by a factor of 6.
Your girlfriend's parents are lucky. Many rural customers are stuck with low bandwidth ADSL connections.

Internet bandwidth is not the issue. Witness the fact that people routinely stream 4K TV. It's the requirement to pump multiple streams around the house over the average DIY wireless network, whilst maintaining sub-millisecond sync, which could be compromised by potentially increasing per-stream bandwidth by a factor of 6.

I agree. I get flawless HD streams into the TV at home, but cannot stream the corresponding uncompressed - required for lip sync - audio around the home minus stuttering except where the Sonos units in the group have been wired to the home network.
Of course the problem there is the small buffer size, but I suspect even buffered Hi Res streams will run into headwinds that will negate even the psychological benefits of Hi Res. Nothing kills music enjoyment faster than stuttering.
PS: In some respects India = rural first world?!
Funny, I’ve done the same comparison, Connect vs Squeezebox, via a pair of the most transparent, “revealing” speakers available, my QUAD ESL-63s. Noticed absolutely no difference whatsoever, as expected.

Some people have adequacy issues, and need to feel “superior” by lying. It’s sad, but all too common.
Userlevel 7
Badge +21
As far as local bandwidth issues, take a look around here at the number of folks trying to add an audio device to their Connect that need to go to compressed audio to get it usable.

I'm a firm believer in Boost mode, with or without a Boost, and keeping my Sonos traffic off my LAN so I can configure my WiFi as I wish with no consideration of the Sonos restrictions/requirements. I'm a fan of wiring everything easy to connect to Ethernet too. These two solved long term aggravations here and with the addition of static/reserved IP addresses for all Sonos gear solved all my Sonos issues.
Badge
As far as local bandwidth issues, take a look around here at the number of folks trying to add an audio device to their Connect that need to go to compressed audio to get it usable.

I'm a firm believer in Boost mode, with or without a Boost, and keeping my Sonos traffic off my LAN so I can configure my WiFi as I wish with no consideration of the Sonos restrictions/requirements. I'm a fan of wiring everything easy to connect to Ethernet too. These two solved long term aggravations here and with the addition of static/reserved IP addresses for all Sonos gear solved all my Sonos issues.


I guess I am extremely fortunate. My router is a bit outdated now, Apple Airport Extreme, but on a good day with like 23 connected devices I am pulling 700 mb per second.
My impression is that Sonos is much more interested in the mass market than it is in the "specialty/hi-res" market. They'd rather sell a thousand speakers to your parents and their friends, than they would selling you and ten of your friends special speakers that can handle hi-res.

I'd guess that the margin of profit on a thousand speakers would be a larger amount of money than the margin on ten speakers.

At the end of the day, Sonos is in business to make money. I suspect that colors some of their decision making around what speakers and codecs they want to make and support.

But not being a Sonos employee, or even an investor, I don't know that for sure, it's just a logical deduction based on evidence that I can see. I have no insight into where Patrick Spence wants to take the company. For all I know, next week they'll introduce a new speaker, just to make me look silly.
I'd guess that the margin of profit on a thousand speakers would be a larger amount of money than the margin on ten speakers.

That is why price points have little correlation with speaker sound quality with today's tech. What Sonos sells for USD 200 has to be sold at many times that price by the little (in comparison to USD 2 billion) audiophile maker to get an adequate return on his investment. There is therefore little to none additional engineering value in such a speaker that may be priced at USD 1000, compared to the USD 200 Sonos product. The audiophile make also has to spend much more per unit on marketing spend to convince audiophiles that such value exists for the additional USD 800 spend. The reality is that all this additional spend that drives the USD 1000 price is there only to support low manufacturing and marketing scales.

So those that say the Sonos is not an audiophile brand are implicitly saying that Sonos offers more sound quality/dollar than the audiophile brands out there, that are severely handicapped by their minuscule scales. Of course the clever marketers will turn this situation around using an elitist pitch for their kit, to which human nature is commonly a willing victim. Everyone is a snob, some hide it better than others. And so the little audiophile makes also have their own little stall in the market, by catering to this human trait.

It wasn't so some decades ago; it took all the progress over this time in solid state and digital tech in componentry that made it possible for excellent output quality to be available at low prices to those that had the ability to achieve scale.

PS: of course, a by product of this is that there really is a lot less to discuss these days on the sound quality subject, which means that there is a lot of noise and heat on things that apparently have scope for discussion, like Hi Res. Lot of that, but no forward motion.

Another great blog post by Mark Waldrep, an expert on HD-Audio in the studio, where it actually matters.

http://www.realhd-audio.com/?p=6664

 

Another great blog post by Mark Waldrep, an expert on HD-Audio in the studio, where it actually matters.

http://www.realhd-audio.com/?p=6664

 

“Let’s push for lossless CD quality streams and better sounding masters. Forget about MQA and any company — hardware or software — that tells you to get behind “hi-res” audio or music.”

Sums it up nicely.

Badge +2

Thanks Ratty,

Just read the article. Excellent explanation. Totally agree with the conclusion regarding better 'Masters'.  Best sound I ever heard was from a remastered Mobile Fidelity LP back in the 70's. And to think I sold it for a fiver when CD's came out and I replaced it!

Another great blog post by Mark Waldrep, an expert on HD-Audio in the studio, where it actually matters.

http://www.realhd-audio.com/?p=6664

 

A follow-up post:

http://www.realhd-audio.com/?p=6725

Badge

Well, that graph is very clear yet it doesn't make the point,  the subject is to prove should oversampling and using more bits than actually needed, in few words having redundancy digital input can be actually identified by trained ears as presenting a more lifelike experience.

I am in the case as it was Sonos and only Sonos that forced our household to jump onto HD audio through Amazon Music HD, now as I am not able to play more than cd quality I cannot say is there’s any difference. I am sure that if there was any wtahsoever It would me minimal in comparison to the jump from Spotify up to CD quality

However I have read an expert opinion on that:

https://www.whathifi.com/news/sonos-says-high-res-audio-support-not-roadmap

What do you think of Mr martin opinion?

Martin is a producer.  His experience with digital sampling theory and the math behind it is probably slim to none.  His position in the realm of “experts” can be likened to the difference between an expert Photoshop user, and the guy who programs the Photoshop application.  You do not need to be an expert on low level image processing to be the former, you most certainly have to be for the latter. 

The same can be said for a guy like Martin vs a guy like Monty at Xiph.  Some of the most absurd, incorrect “facts” I’ve ever heard about digital audio have come from producers and artists.  Just ask Neil Young, he lost his shirt touting the advantages of Hi-res.  Great artist, one of my favorites, but the extent of his actual knowledge about digital sampling couldn’t fill a thimble.

Userlevel 1
Badge +1

I understand that this was banned by the original post, but I can’t resist.
Hires audio is pointless. Red book CD is as good as it gets and Sonos supports it. So who cares? Why bother investing time, effort, energy, disc space etc to something for absolutely no benefit whatsoever?

I don’t think it’s useful to troll audio-files.  (pun intended).  

Nobody thinks Sonos will ever support more than 16 bit 48kHz.  Yet here you are, trolling.

Userlevel 1
Badge +1

Martin is a producer.  His experience with digital sampling theory and the math behind it is probably slim to none.  His position in the realm of “experts” can be likened to the difference between an expert Photoshop user, and the guy who programs the Photoshop application.  You do not need to be an expert on low level image processing to be the former, you most certainly have to be for the latter. 

The same can be said for a guy like Martin vs a guy like Monty at Xiph.  Some of the most absurd, incorrect “facts” I’ve ever heard about digital audio have come from producers and artists.  Just ask Neil Young, he lost his shirt touting the advantages of Hi-res.  Great artist, one of my favorites, but the extent of his actual knowledge about digital sampling couldn’t fill a thimble.

One thing to consider is that the paper spec for Redbook is tight and complete, but the actual clean implementation of that spec was impossible at first (hardware and real world limitations).  Audible digital artifacting and distortion of the brick wall filter was obvious in the 1980’s with a digital workflow.

Sampling theory is very simple to understand, as is Nyquist.  The hard part is implementing and engineering a recording and playbvack system to meet the spec.  It’s been my opinion that the two camps are arguing past each other - one ensconced in the defense of Nyquist and the spec and or best understanding of sound perception by human ears.  The other side is on the engineering and use section of the thing - and the initial shortfalls of digital was never related to the spec, but the impossibility of meeting the spec with the technology of the time.

If the sample rate were increased to 60-80kHz, then a clean anti aliasing filter could have been implemented without a lot of digital processing (which took the better part of the generation to perfect!).  It’s not because of Nyquist and sampling theory.  It’s because you eventually have to build what you are describing.

Userlevel 1
Badge +1

Martin is a producer.  His experience with digital sampling theory and the math behind it is probably slim to none.  His position in the realm of “experts” can be likened to the difference between an expert Photoshop user, and the guy who programs the Photoshop application.  You do not need to be an expert on low level image processing to be the former, you most certainly have to be for the latter. 

The same can be said for a guy like Martin vs a guy like Monty at Xiph.  Some of the most absurd, incorrect “facts” I’ve ever heard about digital audio have come from producers and artists.  Just ask Neil Young, he lost his shirt touting the advantages of Hi-res.  Great artist, one of my favorites, but the extent of his actual knowledge about digital sampling couldn’t fill a thimble.


If the sample rate were increased to 60-80kHz, then a clean anti aliasing filter could have been implemented without a lot of digital processing (which took the better part of the generation to perfect!).  It’s not because of Nyquist and sampling theory.  It’s because you eventually have to build what you are describing.

I think it is also worth noting that the AES spec for recording audio is currently 24/96, so they have sufficient bit depth and sampling rate to make the recording to be limited only by the microphones, and have enough margin to your output file (usually 24/44.1 if it is going to be ripped to a compressed format) or decimated to 16/44.1.  

This places the main burden on playback equipment.  It would be trivial today to hve 24/96 playback so you can just take the master file and not anything that’s been decimated to CD quality at this point.  Whether or not you feel there is a big difference or not.  

And these days most of the differences in the sound of recordings has more to do with the mastering direction given by the producer or artist than anything particularly technical.

 

 It would be trivial today to hve 24/96 playback so you can just take the master file and not anything that’s been decimated to CD quality at this point.  Whether or not you feel there is a big difference or not.  

And these days most of the differences in the sound of recordings has more to do with the mastering direction given by the producer or artist than anything particularly technical.

What would not be trivial is this playback not working on older players that are an inevitable part of multi room set ups - which are Sonos target markets. There is also the question of the higher density streams playing in a stable way wirelessly in groups even where all players are able to decode the bitstreams.

Given the above, that there isn't an audible difference by going down that road other than for the reasons in the last sentence quoted, means that unless the proposed Sonos HD players can downsample to older player compatible formats on the fly, allowing 24/96 playback isn't a trivial matter even if present Sonos products already have the hardware to do this.

Userlevel 6
Badge +15

Well, it’s a start:  https://www.stereophile.com/content/hi-rez-audio-distinguished-blind-testing

 

Audiophiles finally trying ABX testing, but of course at the end of the article they throw in a jab calling us objectivists.

Well, it’s a start:  https://www.stereophile.com/content/hi-rez-audio-distinguished-blind-testing

 

Audiophiles finally trying ABX testing, but of course at the end of the article they throw in a jab calling us objectivists.

I saw that. Not music, but white noise and impulse signals played through “modest” setups to listeners with an average age of 22. For all we know the tiny difference detected could have been due to intermodulation distortion in the ultrasonic range appearing in the audible band. See here

I saw that. Not music, but white noise and impulse signals played through “modest” setups to listeners with an average age of 22. For all we know the tiny difference detected could have been due to intermodulation distortion in the ultrasonic range appearing in the audible band. See here

 

Also, yet another distortion of Dr. Reiss’ meta-analysis.  The participants were not able to “discriminate high resolution content”, they were only able to discriminate a difference between the two tracks.  There was no “this one is the high resolution” or “this one sounds better”.  In fact, there were some studies in that meta-analysis where the participants found the 16/44.1 to be better, presumably due to the same intermodulation distortion. 

 

Reply