How can DAC's have a SOUND SIGNATURE if they measure as transparent?

[Wanman]

A well randomized test...

Well randomized is the key. If you only use college students, or only use volunteers at an AES conference, or even if you only use people from the US, it is unlikely they are sufficiently well randomized to represent the population of the world.

Actually the problem is more difficult that it may first seem. A famous and well regarded paper on medical research, although not without some criticism too (and medical research is what this would be): https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1182327/

 

[Wanman]

Whenever Watts' -300 to -400db claim comes up, I'm reminded of this scene

You do understand that Watts is talking about numerical precision as measured in the digital domain only, not anything analog?

 

[voodooless]

You do understand that Watts is talking about numerical precision as measured in the digital domain only, not anything analog?

Does that matter? He claims it’s audible, so whatever it does should then also be embedded in the analog signal somehow.

 

[Wanman]

Does that matter? He claims it’s audible, so whatever it does should then also be embedded in the analog signal somehow.

It might matter because he said he doesn't know why its audible. IOW, it could be that whatever is audible is correlated with the calculational precision. Watts likes to use very long digital filters. In that case, or for other calculations Watts may be doing (perhaps for his 17th order modulator) numerical errors may accumulate merely because of the large number of calculations. If using 64-bit math actually does turn out to sound better than using 32-bit math then it could be due to accumulated errors.

Its not unheard of for such problems to occur in some cases. One randomly selected paper in the general subject area: https://arxiv.org/html/physics/9807003

 

[voodooless]

It might matter because he said he doesn't know why its audible.

He didn’t even really prove that it’s audible at all.

 

[DSJR]

I remember him claiming it was the *effects* of the -300dB that were audible, but of course we eejits listening to the lecture have no means of proving it. I was worried that one or two fellow audients were lapping it all up and nodding sagely as if they know what the eff he was talking about. he makes it so convincing though, yet all this better reproduction of reverb tails, 'room boundaries' and such-like are removed by their amps it increasingly appears after hearing some dire sounds at the weekend again...

 

[IAtaman]

It might matter because he said he doesn't know why its audible. IOW, it could be that whatever is audible is correlated with the calculational precision. Watts likes to use very long digital filters. In that case, or for other calculations Watts may be doing (perhaps for his 17th order modulator) numerical errors may accumulate merely because of the large number of calculations. If using 64-bit math actually does turn out to sound better than using 32-bit math then it could be due to accumulated errors.

Its not unheard of for such problems to occur in some cases. One randomly selected paper in the general subject area: https://arxiv.org/html/physics/9807003

Just to be sure, Rob the atomic clock claims there is no limit to how accurate smalls signals needs to be, and that this explains why pure copper cables produce better depth perception and presumably better soundstage. Here is a screen shot of his slide. You are claiming there might be some truth to this? I would love to understand better.

 

[IAtaman]

It might matter because he said he doesn't know why its audible. IOW, it could be that whatever is audible is correlated with the calculational precision. Watts likes to use very long digital filters. In that case, or for other calculations Watts may be doing (perhaps for his 17th order modulator) numerical errors may accumulate merely because of the large number of calculations. If using 64-bit math actually does turn out to sound better than using 32-bit math then it could be due to accumulated errors.

Its not unheard of for such problems to occur in some cases. One randomly selected paper in the general subject area: https://arxiv.org/html/physics/9807003

By the way he also claims his 9 year old son could tell the difference between a 120db DAC and 150db DAC while watching Shaun the Sheep on TV. If you have any explanations for that, I'd love to hear that too.

 

[Wanman]

He didn’t even really prove that it’s audible at all.

True. But that doesn't mean the claim is automatically false either.

 

[Wanman]

By the way he also claims his 9 year old son could tell the difference between a 120db DAC and 150db DAC while watching Shaun the Sheep on TV. If you have any explanations for that, I'd love to hear that too.

Maybe he was measuring the wrong things? Sigma delta DACs have dynamic distortion/noise effects that don't occur in other types of audio gear. What looks like noise on an FFT can be a non-PSS deterministic signal that sounds ugly. It may take some review of how DFTs work to see how that can happen, although to put it in simple terms a non-PSS signal changing over the time an FFT is acquired may have its energy spread across bins. Thus it may not appear as a distinct spur. Not sure if Watts was measuring everything that could be measured when he described the dacs his son was said to be able to discriminate.

 

[voodooless]

True. But that doesn't mean the claim is automatically false either.

Since nobody has demonstrated otherwise since the release of the products, I think we can be quite confident regarding the audibility.

 

[fpitas]

Just to be sure, Rob the atomic clock claims there is no limit to how accurate smalls signals needs to be, and that this explains why pure copper cables produce better depth perception and presumably better soundstage. Here is a screen shot of his slide. You are claiming there might be some truth to this? I would love to understand better.

View attachment 282800

That really is a collection of nonsense. I wonder if he believes it.

 

[ahofer]

True. But that doesn't mean the claim is automatically false either.

What to do? Believe something that is ridiculously unlikely and that he hasn’t actually demonstrated? Or defer to the microscopic possibility that it might be true and shift the burden of proof to others?

Hmmm.

 

[fpitas]

Once again if falls to us to (dis)prove something ridiculous. I think I've heard this one before.

 

[Soria Moria]

The people who are extremely confident about these things never go out of their way to prove them.

 

[fpitas]

The people who are extremely confident about these things never go out of their way to prove them.

That's our job, apparently. I mean, Galileo made somebody else prove his conjectures. All great science works that way

 

[IAtaman]

Maybe he was measuring the wrong things? Sigma delta DACs have dynamic distortion/noise effects that don't occur in other types of audio gear. What looks like noise on an FFT can be a non-PSS deterministic signal that sounds ugly. It may take some review of how DFTs work to see how that can happen, although to put it in simple terms a non-PSS signal changing over the time an FFT is acquired may have its energy spread across bins. Thus it may not appear as a distinct spur. Not sure if Watts was measuring everything that could be measured when he described the dacs his son was said to be able to discriminate.

Did you watch the video in which he makes these claims?

 

[Jimbob54]

That really is a collection of nonsense. I wonder if he believes it.

I think he believes it's the only card he can play to distinguish his (and let's not fanny around saying he has no stake other than consultants fee in Chord DACs, they are his ) products from any other competent DAC.

 

[fpitas]

I think he believes it's the only card he can play to distinguish his (and let's not fanny around saying he has no stake other than consultants fee in Chord DACs, they are his ) products from any other competent DAC.

Well, he got people on the internet to harass us about it. So that's success. I guess.

 

[antcollinet]

What to do? Believe something that is ridiculously unlikely and that he hasn’t actually demonstrated? Or defer to the microscopic possibility that it might be true and shift the burden of proof to others?

Hmmm.

Personally I choose to believe in the teapot.