Serious Question: How can DAC's have a SOUND SIGNATURE if they measure as transparent? Are that many confused?

[pablolie]

I'm going to guess your background is in research science. Some of the applied science ("engineering") folks get upset when people ask questions about things they believe to be (or want to be) black and white.

HarmonicTHD has been quick to accuse others here of lacking knowledge. It's not just you.

It's never black and white. There are no *truths*, leave that to philosophy,

It's about which *repeatable* model works best in physics and math, which engineers then leverage. The moon landing was not an exercise in belief, it was a methodical approach based on experimentation to create a repeatable working model. In engineering, we measure success: did we meet design goals? The "truth" of the universe is never considered in that exercise - not even in physics, which are also domain specific. Newtonian and Quantum physics *do* happily co-exist,.

*Belief* on the other hand has only built ruins and failure when mis-applied as an implementation guideline. Belief doesn't try to be a repeatable, deterministic model for measurable success.

 

[preload]

The question, though, is whether DACs sound the same. Generally they do, and there’s very little scientific evidence to the contrary.

What that has to do with the frontiers of science, I don’t know. Maybe if we identified an unmeasurable but audible difference, we could investigate it and find out. But pants before shoes please.

I think it's fair to ask questions. And it's more than fair to ask how/why we believe something to be true.

Everyone has a different threshold for the degree of evidence necessary to be compelled.

For some, DAC measurements (e.g. FR, THD, SNR, etc) within certain parameters is sufficient to assume that modern DACs are generally transparent.

For me, it was the descriptions of experiments where they daisy-chained the inputs and outputs if multiple ADCs and DACs and asked listeners to differentiate vs a single DAC. And then they posted the actual comparison audio files for anyone to compare and try out in a PC ABX (like the foobar plugin). That, to me, is an elegant experiment and it is very compelling, far more compelling than measurements.

 

[FlyingFreak]

I am trying to understand what could be validated in the belief DACs sounds different.

Are people hearing differences imagining that
engineers building the 'superior' dac know something about science of sound reproduction or electronics no one else on Earth does?
Or do they thing it is a matter of chance?
Or is magic involve?

 

[preload]

I am trying to understand what could be validated in the belief DACs sounds different.

Are people hearing differences imagining that
engineers building the 'superior' dac know something about science of sound reproduction or electronics no one else on Earth does?
Or do they thing it is a matter of chance?
Or is magic involve?

Welp for starters, consider the case of DACs that use an analog pot for volume control.
They can have a L>R mismatch of >1dB, even with a nice Alps component.

Is that transparent? Arguably not. Could that result in perceived differences in soundstage? Maybe. Could that result in FR differences? Maybe, if your L and R transducers have different FR curves (for instance, some headphones).

But here's the kicker: do we measure L and R channel balance at ALL pot levels on DACS that use an analog pot? Because if we don't, we would miss this channel imbalance. Meanwhile, all of our other fancy measurements taken on our fancy Audio Precision would tell us there should not be any audible differences. Think about it.

Note: Based on all of the evidence I've seen, I do NOT believe that, in general, modern DACs sound different. However I cannot rule out the possibility that a design decision or oversight (whether engineering-driven or not) could result in an audible difference in the occasional DAC design.

 

[DonR]

Welp for starters, consider the case of DACs that use an analog pot for volume control.
They can have a L>R mismatch of >1dB, even with a nice Alps component.

Is that transparent? Arguably not. Could that result in perceived differences in soundstage? Maybe. Could that result in FR differences? Maybe, if your L and R transducers have different FR curves (for instance, some headphones).

But here's the kicker: do we measure L and R channel balance at ALL pot levels on DACS that use an analog pot? Because if we don't, we would miss this channel imbalance. Meanwhile, all of our other fancy measurements taken on our fancy Audio Precision would tell us there should not be any audible differences. Think about it.

The phenomenon is called non-linear tracking and can be overcome by using a stepped attenuator using precision resistors (my preference), a higher-quality pot or measured and compensated in-circuit.

 

[preload]

The phenomenon is called non-linear tracking and can be overcome by using a stepped attenuator using precision resistors (my preference), a higher-quality pot or measured and compensated in-circuit.

Yep, or better yet, attenuation can be done digitally. But that wasn't my point, was it.

Small design features can result in audible differences. And they may not show up on "routine" DAC measurements. I used this as an example. So perhaps we can lay off the member here who decided to raise a question.

 

[DonR]

Yep, or better yet, attenuation can be done digitally. But that wasn't my point, was it.

Oh no. Digital attenuation is a no-no... just in case. As to your point, I think volume controls on DACs are superfluous. Set the output to be 2Vrms at 0db and be done with it. Let the pre-amp control the volume.

 

[Mnyb]

Welp for starters, consider the case of DACs that use an analog pot for volume control.
They can have a L>R mismatch of >1dB, even with a nice Alps component.

Is that transparent? Arguably not. Could that result in perceived differences in soundstage? Maybe. Could that result in FR differences? Maybe, if your L and R transducers have different FR curves (for instance, some headphones).

But here's the kicker: do we measure L and R channel balance at ALL pot levels on DACS that use an analog pot? Because if we don't, we would miss this channel imbalance. Meanwhile, all of our other fancy measurements taken on our fancy Audio Precision would tell us there should not be any audible differences. Think about it.

Note: Based on all of the evidence I've seen, I do NOT believe that, in general, modern DACs sound different. However I cannot rule out the possibility that a design decision or oversight (whether engineering-driven or not) could result in an audible difference in the occasional DAC design.

The phenomenon is called non-linear tracking and can be overcome by using a stepped attenuator using precision resistors (my preference), a higher-quality pot or measured and compensated in-circuit.

That why it’s good to have ASR , things like these and other stuff would show up if a DAC sounds “different” it’s a design error somewhere . Or sadly enough a setting, the fad with selectable filters I suspect some of the settings are not transparent in some DAC’s and should not be exposed to end users , could be hidden somewhere for testing purposes .

Many off the very emotionally invested audiophiles does not see what folks here actually says.
Absolutely no one really says “all DAC’s sound the same” it’s small but import bit missing “all properly designed DAC’s sound the same” ( no filterless NOS DAC for you ) and actually “sound the same” is not really a proper description of the matter there another piece of the puzzle “all properly designed DAC’s are transparent to the source material for a human listener” same could mean equally bad ?

You can get all that for 100$ .
The next misconception.
Audio reproduction quality and price has a very weak correlation. That’s very true , it can even be negative if you buy high end hifi DAC’s that’s not properly designed like an very expensive Audio Note DAC, it’s worse than any wall mart 50$ 20 year old DVD player .

But product quality and maintenance can have some correlation with price up to a point. It’s usually the whole company not only a specific product that’s takes care to design . So sometimes you get what you pay for.
Audio quality alone does not make a good product. That’s why measurement of old products give us insights to this , is the thing up to its original spec 10 years later ? Who knows ?

Good examples here are RME and Benchmark. Cost more but not silly prices and well run operations that makes quality products.

 

[HarmonicTHD]

Wow, so much hate for my post. I think people are deliberately misunderstanding and frankly resorting to insults - HarmonicTHD, SIY - what makes you think I lack knowledge and what is IFLS?

(I have two science degrees, one undergrad, one postgrad, FWIW. By 'belief' I meant the belief that we understand - fully - what's going on, because we don't. You tell me what dark matter is, for example).

Anyway that's enough about that.

Well could then explain please what the relativity theory and the boundaries of physics have to do with Audio? And the implication that we need to discover new physics to understand audio electronics or DACs for that matter (because „old physics“ are dead according to you)?

 

[pma]

Just recently I have had a feeling that two different, quite well measuring DACs do sound different. So I made an A/B test, level matched, and it was hopefully possible to make it in completely same hardware setup that brought apparent sound difference. This level matched test showed no sound difference. I have then made a set of measurements and also Deltawave files. I will prepare a new thread with results and descriptions.

 

[solderdude]

I am trying to understand what could be validated in the belief DACs sounds different.

Are people hearing differences imagining that
engineers building the 'superior' dac know something about science of sound reproduction or electronics no one else on Earth does?
Or do they thing it is a matter of chance?
Or is magic involve?

A small look in the kitchen through some keyholes. (my emphasis)

Even "good sound" can be quantified. By listening tests. I have learned to take typical AP Test results (or similar) as a proxy, but I have not found the result reliable enough to make a reliable rule.

Important here is how the listening tests were done.

There are enough outliers to my own view of "this should sound good" to suggest that rules I use are insufficient. It is easier to predict "bad sound" from measurements but even this is not 100% reliable.

To me this suggests a hidden variable that is not quantified. That is not saying it cannot be quantified or is mystical, it is just outside current scopes. I might even have a rough idea and a hypothesis. In part Hawkesford's works points the way ("Fuzzy Distortion") I think. Do I want to work on it? Why not, if someone foots the bill.

This suggests 'choosing components/circuits' that sound 'more pleasant' to be found by accident. It also isn't known by other designers.
At least that's what can be assumed, on the other hand it would be easy for some designers/brands (note:not Thorsten) to claim they do and keep it as trade secret.

On the other hand, what seems to me to be much more unexplored is psychoacoustics, i.e. perception and awareness.

The thing with audio is that from the moment the recording(s) are made down to the listener there is mechanics, transducers, (psycho)acoustics, electronics, theories about operation, science, engineering, production limitations being involved in all stages.

My conclusion was the specific measures taken in the circuit design and execution of similar but subtly different circuits manifested as a reliable listening preference in one metric and differences in (T)HD in another metric.

It seems that the 'positive difference' that was found was pure by chance and not researched further as to the why.
So 'pure chance' and 'going with it' seems to be a way to go about this.

Could be the difference in circuits, could be used parts, could have been the testing for preference in Thorsten's case.
In any case it remains 'mysterious' yet producible with current components.

 

[Blumlein 88]

A small look in the kitchen through some keyholes. (my emphasis)



Important here is how the listening tests were done.



This suggests 'choosing components/circuits' that sound 'more pleasant' to be found by accident. It also isn't known by other designers.
At least that's what can be assumed, on the other hand it would be easy for some designers/brands (note:not Thorsten) to claim they do and keep it as trade secret.



The thing with audio is that from the moment the recording(s) are made down to the listener there is mechanics, transducers, (psycho)acoustics, electronics, theories about operation, science, engineering, production limitations being involved in all stages.



It seems that the 'positive difference' that was found was pure by chance and not researched further as to the why.
So 'pure chance' and 'going with it' seems to be a way to go about this.

Could be the difference in circuits, could be used parts, could have been the testing for preference in Thorsten's case.
In any case it remains 'mysterious' yet producible with current components.

You have more faith in Mr. Loesch than I think is warranted. The most likely reason is his listening tests don't pass muster. He would never quite come clean about the details of those tests.

 

[Robbo99999]

I think it's fair to ask questions. And it's more than fair to ask how/why we believe something to be true.

Everyone has a different threshold for the degree of evidence necessary to be compelled.

For some, DAC measurements (e.g. FR, THD, SNR, etc) within certain parameters is sufficient to assume that modern DACs are generally transparent.

For me, it was the descriptions of experiments where they daisy-chained the inputs and outputs if multiple ADCs and DACs and asked listeners to differentiate vs a single DAC. And then they posted the actual comparison audio files for anyone to compare and try out in a PC ABX (like the foobar plugin). That, to me, is an elegant experiment and it is very compelling, far more compelling than measurements.

I think I took that test, it was a really old thread, I think started by Blumlein, but not sure. I couldn't reliably tell the difference between a music track that had undergone one DAC process vs a music track that had undergone (was it) 8 loops of DAC->ADC.

 

[solderdude]

You have more faith in Mr. Loesch than I think is warranted. The most likely reason is his listening tests don't pass muster. He would never quite come clean about the details of those tests.

I was merely showing what the line of thinking is with some designers that venture 'beyond' what is known to exist.
Thorsten does know how design transparent enough (for us mortals) equipment but does like to 'allow' headroom, as it were, to please the subjective crowd as well.
Basically enlarging the possible market for audio products.

Not agreeing with some of the claims.

 

[Blumlein 88]

I think I took that test, it was a really old thread, I think started by Blumlein, but not sure. I couldn't reliably tell the difference between a music track that had undergone one DAC process vs a music track that had undergone (was it) 8 loops of DAC->ADC.

I think I have included links to those tests 5 times in this thread. BTW, there were two of them with different gear used in each test.

Serious Question: How can DAC's have a SOUND SIGNATURE if they measure as transparent? Are that many confused?

...that is the best take on the subject I have seen so far... ...it is clear to any thinking person that there are asspects involved we have no way to measure... ...it was just stated there are only three asspects measured at the current state of the art...

www.audiosciencereview.com

 

[pma]

Just recently I have had a feeling that two different, quite well measuring DACs do sound different. So I made an A/B test, level matched, and it was hopefully possible to make it in completely same hardware setup that brought apparent sound difference. This level matched test showed no sound difference. I have then made a set of measurements and also Deltawave files. I will prepare a new thread with results and descriptions.

So, here is the new thread, with A/B test of two DACs:

A/B test of two DACs - do they sound different?

A/B test of two DACs – do they sound different? Doing some measurements and changes in my audio chain, once again I had a feeling that the DAC that I have just connected into the system sounds different. Wow, better clarity and tighter, well defined bass! Not the first time such situation...

www.audiosciencereview.com

 

[Aleksandar RS]

Yes.
Look at the review of the $46500 CH 1.2 in the latest Stereophile. It’s no better than mediocre. I’m sure it sounds OK, though, because our ears are very poor discriminators.
They’re selling a fancy box and lots of hand-waving, and that’s what some people want to buy.

I completely agree that there are such examples, quite a few, but that is not the rule. I was thinking more about audio companies with decades of experience and great reputation. Although their prices are realistically higher than what you get. Quite simply, part of the price is their name behind the product. I think it's the same outside of the audio sphere.

I don't think we should discuss these relatively new fancy companies that release their first (second, third) device and sell for crazy money.

 

[charleski]

I completely agree that there are such examples, quite a few, but that is not the rule. I was thinking more about audio companies with decades of experience and great reputation. Although their prices are realistically higher than what you get. Quite simply, part of the price is their name behind the product. I think it's the same outside of the audio sphere.

I don't think we should discuss these relatively new fancy companies that release their first (second, third) device and sell for crazy money.

Well, CH Precision was founded in 2009, so it's not that new. The founders had over a decade of experience working for Goldmund and realised the way to make money in hifi was to satisfy those looking to buy magic for a lot of money.
I think the spell is "Expecto Aurum", but you need the right wand.

 

[Killingbeans]

I completely agree that there are such examples, quite a few, but that is not the rule. I was thinking more about audio companies with decades of experience and great reputation. Although their prices are realistically higher than what you get. Quite simply, part of the price is their name behind the product. I think it's the same outside of the audio sphere.

I don't think we should discuss these relatively new fancy companies that release their first (second, third) device and sell for crazy money.

I'm not really sure what point you're trying to make?

That those 'decades of experience and great reputation' automatically gives real and verifiable superiority in audible performance?

 

[Aleksandar RS]

Well, CH Precision was founded in 2009, so it's not that new. The founders had over a decade of experience working for Goldmund and realised the way to make money in hifi was to satisfy those looking to buy magic for a lot of money.
I think the spell is "Expecto Aurum", but you need the right wand.

I know, that's why I wrote relatively young, and as for the experience at Goldmund is extremely useful thing. But besides that, CH Precision is not a name that has much significance in the audio world. I'm not saying that this device is bad, but as a company (of the rank you are) you have to position it where it really belongs. And that's much, much lower on the price scale. I don't know what card they are playing, if they think someone will buy something like this.

This is something that I don't think can happen with companies with big reputations (who, surely, pretty much build their name into the price), but they won't make a fool of the customer. There are examples of crazy prices from such companies as well (Nagra Reference Anniversary Turntable) with a price of around $170,000. The version is limited, I think seventy units. It's about them selling this turntable as a piece of art (sculpture) that plays records. I saw and heard it at an audio show a few months ago accompanied by Nagra amplification and Marten speakers. It sounds nice, but the sound quality has, of course, nothing to do with the price. It's true that the whole system looks like a modern piece of art, so from that point of view it's ok, but now it's a different story and a different scale of evaluation. In any case, the entire system was sold for about $450,000 on the first day of the show.