David Chesky on Streaming Convenience vs Sound Quality

[theREALdotnet]

The real mystery to me is that when we capture both files in a Digital Audio Workstation, they null out. So what causes the difference in sound?​

Only one possibility left: the mind.

Even though it’s a digital signal, could it be some triboelectric effect over long wires? Could all the complex connections across the web be picking up electrical noise along the way, degrading the sound during playback?

No. Any effect (triboelectric or otherwise) would prevent the files from nulling out.

 

[Old_School_Brad]

He wrote: "To clarify, we’re not conducting formal double-blind testing here; one person plays the files while the other listens, drawing on our experiences as listeners. With over 50 years as a pianist, composer, orchestrator, and producer of nearly a thousand albums, I trust my musical memory, which I believe extends to tonality as well."

He is 68 years old now and trusts his ears and his "musical memory"? One blind test conducted by someone trustworthy (clearly a job for amir ) could cure his arrogance.....

Eventually all tests these older golden ears do will naturally be blind..

 

[Sal1950]

His conclusions are most likely technically incorrect, but it does not make him a charlatan or a fool... simply wrong. I know I have been wrong many times in many different ways over the years. Give the guy a break.

Humm, not a fool for sure.. he's making a fortune selling the belief in magic. He's much more than just "wrong" as in mistaken.
He's firmly ensconced in the "high end boys club" selling everything from cables for $1k a foot, to all sort of valueless widgets and DXD rate recordings.
What would you call him in a society that valued integrity in business matters?

 

[KSTR]

No. Any effect (triboelectric or otherwise) would prevent the files from nulling out.

Chesky wrote: "The real mystery to me is that when we capture both files in a Digital Audio Workstation, they null out."
I take that for they compared the digital source content, not the actual analog playbacks, as a basic and mandatory check whether the source files are the same, bit-identical.
Captured analog playback never nulls out deeply, especially not when recorded asynchronously (which of course we don't know)... unless special tricks are used (read; use software like DeltaWave, not plain simple subtraction in DAW).

Anyway, bit-identical playback by definition can only show differences from secondary effects during the two actual playback processes. And as noted, the only reasonable way how the DAC output can be affected while playing the same data is different levels/structure of RF and other (common-mode) noise making it into the analog section of the DAC or devices further downstream. Highly unlikely but not impossible.... and when it happens to reach audible levels (in a proper blind test, of course) it can be measured as there must have been a difference.

 

[amirm]

Chesky wrote: "The real mystery to me is that when we capture both files in a Digital Audio Workstation, they null out."
I take that for they compared the digital source content, not the actual analog playbacks, as a basic and mandatory check whether the source files are the same, bit-identical.

Presumably the same hardware is used to play local and streaming files. Therefore, in his book, the difference in the two chains must be upstream. That is why he performed the null test. Once he got a true null, that concluded the story that his sighted testing is wrong.

 

[BeeKay]

Always handy to carry Hitchens’s razor along:

"What can be asserted without evidence can also be dismissed without evidence"

 

[Mr. Widget]

Presumably the same hardware is used to play local and streaming files. Therefore, in his book, the difference in the two chains must be upstream. That is why he performed the null test. Once he got a true null, that concluded the story that his sighted testing is wrong.

100%

Unfortunately the conclusion that many here seem to take away from that is that he is being disingenuous because of a profit motive.
I certainly that as possible, but I think it is more likely that he is being swayed by his decades of experience and relying on his perception. If his goal was to scam others, I doubt he would ever admit to the null test or its results.

 

[Galliardist]

100%

Unfortunately the conclusion that many here seem to take away from that is that he is being disingenuous because of a profit motive.
I certainly that as possible, but I think it is more likely that he is being swayed by his decades of experience and relying on his perception. If his goal was to scam others, I doubt he would ever admit to the null test or its results.

I think you're missing a point in turn - which is that the more unlikely an improbable sound difference is, and if it's one that the storyteller has investigated "scientifically" in some way, it's more believable to the audiophile majority.

They WANT the "unmeasureable difference" to reinforce their - should I use the word delusion here?

 

[antcollinet]

100%

Unfortunately the conclusion that many here seem to take away from that is that he is being disingenuous because of a profit motive.
I certainly that as possible, but I think it is more likely that he is being swayed by his decades of experience and relying on his perception. If his goal was to scam others, I doubt he would ever admit to the null test or its results.

For me ethical sales require more than "ignorance is a defense". If you are manufacturing stuff in a field outside of your technical competence, then there is a duty to either put it inside your technical competence, or get other people who are competent to check for you that you are not selling lies.

In particular - if you are making technical claims about your product, being able to demonstrate those technical claims are real. Which requires measurements.

Bear in mind the person we are talking about here actually did a Null test. Why did he do that if he didn't believe the null test should show soemething real. And yet when it didn't back up his perceptions - suddenly it was wrong. There must be something mysterious causing the sound difference.

 

[olieb]

Why did he do that if he didn't believe the null test should show soemething real.

Well it does in some twisted way when "real" is interpreted in a different way. The null test shows that there is no technical established explanation, therefore it becomes some kind of "proof for the mysterious" [and that might be one reason to mention it]. The existence of a difference in sound is just taken for granted, no doubt possible.

 

[voodooless]

There must be something mysterious causing the sound difference.

Jingling gold coins in his pockets most likely…

 

[Old_School_Brad]

Well it does in some twisted way when "real" is interpreted in a different way. The null test shows that there is no technical established explanation, therefore it becomes some kind of "proof for the mysterious" [and that might be one reason to mention it]. The existence of a difference in sound is just taken for granted, no doubt possible.

Exactly. It gets used as a "strong" argument against measurements, claiming that "they don't reflect what we actually hear."
It's a backward perspective, but unfortunately, not uncommon to see in other Hi-Fi forums.

 

[VientoB]

The only issues I have with streaming are - it's more power hungry if you listen to the same albums regularly. And you don't have offline access to your music.

 

[Galliardist]

The only issues I have with streaming are - it's more power hungry if you listen to the same albums regularly. And you don't have offline access to your music.

Most streaming services allow you to import or download albums to devices that you can play from there, either as long as you are a subscriber, or sometimes as long as the albums concerned are still on the service (occasionally that can be a new remix of the album replacing the one you've got. I believe you have to go online to the service at intervals to keep access to the downloaded files - that's the case with Qobuz. Of course, you need to play from a device with software that allows you to store locally and play back- that may require a phone or computer in the loop.

 

[BeeKay]

The only issues I have with streaming are - it's more power hungry if you listen to the same albums regularly. And you don't have offline access to your music.

Plus, what happened to me the other week, the service you subscribed to suddenly ditches a contract with a certain artist or label. Not great.
I literally went on eBay to buy a CD copy.

 

[Sal1950]

The only issues I have with streaming are - it's more power hungry if you listen to the same albums regularly. And you don't have offline access to your music.

My only issue is that in most cases I can't choose the particular mastering/mix I'm interested in hearing. Many of my most loved albums have been reissued half a dozen times over the decades. Some are beautiful and some have had their dynamic range mashed to a near flatline. Mostly I use streaming simply to preview new music before purchasing disc or download, or to enjoy Atmos mixes on Apple that haven't been released in a lossless version. I'm really hoping that like happened with 2ch, public interest/demand in lossless will push some streamer to offer it. It took us decades to get lossless 2ch streaming.

 

[JustJones]

Same old stuff, different decade.

It wasn't until March 1998 that the DVD Audio spec was settled on: PCM audio with 16, 20, or 24 bits per sample with sample rates ranging from 44.1 to 192 kHz. In time, the DVD Forum selected Meridian's lossless MLP codec for DVD Audio to reduce file size, as well. "With 96 kHz and 24 bits, you don't hear digital artifacts," David Chesky of audiophile label Chesky Records told Stereo Review. "At 44.1 kHz, there's hardness of timbres and lack of space."

Read More: https://www.slashgear.com/1712301/formats-failed-replace-audio-cd-history/

 

[GXAlan]

Presumably the same hardware is used to play local and streaming files. Therefore, in his book, the difference in the two chains must be upstream. That is why he performed the null test. Once he got a true null, that concluded the story that his sighted testing is wrong.

“The real mystery to me is that when we capture both files in a Digital Audio Workstation, they null out.”

This suggests that they are just looking at the files and not the recording. I guess it would be super easy to take a premium ADC, the sorts of which they have at their studio, and compare the two recording. Though they may not null, software like DeltaWave will identify anything that reaches audible thresholds.

There was a post a month or two ago, where the complaint was the CD vs the file ripped from the same CD. Digging into the user manual of the playback device (from Cary Audio), when used as a USB DAC, there is no processing but when used with the physical CD, there was a different upsampling algorithm which also results in different published SNR/dynamic range.

So the assumption that a playback device actual matches the volume and signal processing between streaming and as a file player is not guaranteed.

I will try to find an email for David Chesky and ask him… (edit: …and done)

 

[SSS]

Tried several times, David Chesky does not answer on emails.

 

[usersky]

I think some of us here take artists more seriously than they take themselves.