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

[oleg87]

Guys, I was about to start a thread and strangely saw this one that it closely linked to what I wanted to ask.

What is stopping a DAC from being transparent with measurements but also emphasising (just for example) the upper mids and sub bass, for a certain signature.

Can this happen? If not, can someone explain so I understand it.

I guess I’m asking if there are ways (just like EQ) to colour sound by elevating certain frequencies, but almost hard baking it into the firmware.

Hope that makes sense.

Well, you could certainly add some kind of digital or analog filters to do that. It would be EQ.

 

[antcollinet]

I feel very flattered that you didn't correct my inexpert attempt to summarise some basic principles at work here!

Didn't feel the need to - any inaccuracies were below the level of audibility.

 

[antcollinet]

Guys, I was about to start a thread and strangely saw this one that it closely linked to what I wanted to ask.

What is stopping a DAC from being transparent with measurements but also emphasising (just for example) the upper mids and sub bass, for a certain signature.

Can this happen? If not, can someone explain so I understand it.

I guess I’m asking if there are ways (just like EQ) to colour sound by elevating certain frequencies, but almost hard baking it into the firmware.

Hope that makes sense.

A dac has one, and only one job : to perfectly reproduce the waveform represented by the digital data, in analogue form.

A transparent DAC does that so accurately that any deviations are so small they are inaudible. (Like perfectly transparent glass (if it existed) would be invisible - it would transmit the light with no visible defects)


If a dac is "emphasising the upper midsection and sub bass", then it is changing (emphasising) the waveform in order to do that. If that is audible, then by definition, the DAC is not transparent. - it has changed the waveform in an audible way. And if it is audible is must also show up in the measurements, so it wouldn't measure as transparent.

EDIT - one of the measurements is frequency response. Audibly perfect (transparent) Dacs have flat frequency response. EQ (even if baked in) would have to put a peak in the mids and a peak in the sub bass to emphasise them. Then the frequency response is no longer flat.

 

[Sal1950]

I have spent the last day or so reading large parts of this thread. It has been an educational and rather painful experience - realising how much money I have wasted. Schitt Yggdrasil and fancy digital transport, I'm looking at you.....

I am posting to thank those who have repeatedly taken the time to share their expertise in this thread. I have to commend many of the regulars for their patience as, time and again, variants of the same claims have been made by 'newcomers' to this thread. Attempts are typically made to explain basic principles and guide, often quite gently, toward an understanding of the basic principles at work. The 'newcomers' are often resistant to the point of abusiveness, but the patience displayed is more often than not impressive.

Well done and thank you.

@barza , Thanks for the kind words toward our members and website. We do try to keep an open mind and be inviting to new members
when discussing the incorrect emotional subjective findings that are so rampant in our industry. If you're not already aware, I'd like to expose you to a few writings on audio by the late audio luminary Peter Aczel. Since his passing in 2017 at the young age of 91, little has changed to alter the things he wrote in these two essays.
Cheers, Sal

What I have learned after six decades in audio (call it my journalistic legacy):​

The Ten Biggest Lies In Audio

 

[barza]

Sal1950 - many thanks for highlighting those two articles by Peter Aczel. I very much enjoyed reading and learning from them. Rest in peace, Peter, and thank you for leaving some directly expressed wisdom behind.

 

[ShootingStills]

Any deviation from linearity in a DAC will show up in measurements. In case you're thinking of a non-linear output stage installed as a "trick" by the manufacturer ... that will show up in measurements, too.

Scientific measurements of a DAC are similar to an autopsy. There's no way to hide anything.

Interesting and a great reply, this is what I was looking to understand.
I hear and read a lot of comments about DAC's having a wider soundstage, or being more musical, or richer, or 'insert any other adjective' to describe their sound.
The only time I ever noticed a difference in the sound quality, was going from the internal DAC of the Node 2 to the DAC of the RME ADI 2, I have since sold that and now have an Eversolo A6, which (after some initial sound issues that I think I resolved) I cannot tell the difference there either.

So if a transparent DAC had a bass boos function or even tone controls that would then suddenly become non transparent and show up in measurements?

 

[ShootingStills]

A dac has one, and only one job : to perfectly reproduce the waveform represented by the digital data, in analogue form.

A transparent DAC does that so accurately that any deviations are so small they are inaudible. (Like perfectly transparent glass (if it existed) would be invisible - it would transmit the light with no visible defects)


If a dac is "emphasising the upper midsection and sub bass", then it is changing (emphasising) the waveform in order to do that. If that is audible, then by definition, the DAC is not transparent. - it has changed the waveform in an audible way. And if it is audible is must also show up in the measurements, so it wouldn't measure as transparent.

EDIT - one of the measurements is frequency response. Audibly perfect (transparent) Dacs have flat frequency response. EQ (even if baked in) would have to put a peak in the mids and a peak in the sub bass to emphasise them. Then the frequency response is no longer flat.

Thanks very much for this, just what I wanted to know. So they are given a flat frequency response and if they show elevation in one area of the Frequency range or another this will show as colouration and 'fail' essentially?

Thanks so much

 

[Jim Taylor]

So if a transparent DAC had a bass boos function or even tone controls that would ... show up in measurements?

Correct.

 

[denzity]

If you spend some time working through the interchanges in this thread, you will likely come away as sure in that conclusion as I think you can be. I am an economist too, and am pretty sure by now I have been led into some very poor and expensive audio equipment decisions which I understand after reading this thread and some of the material it links too.

An attempt to summarise the basic principles (from one economist to another....)

-The job of a DAC is to convert a digital representation of audio into a continuous voltage signal. That voltage signal is then passed to other equipment (preamps, speakers, headphones) to generate what we hear.
-Testing equipment can assess the accuracy of the digital to voltage conversion with a high degree of precision. Devices where any inaccuracies lie well outside of the bounds of human hearing can be referred to as being 'audibly transparent'
- Technology has developed to the point that many relatively inexpensive DACS are audibly transparent
- Once a device is audibly transparent its output is basically indistinguishable from other audibly transparent devices to the human ear under controlled listening/testing conditions. There is no "better soundstaging" or similar fairy dust that lies within the bounds of both (a) remaining accurate to the source material and (b) being reliably identified by listeners under controlled conditions.

As for your list as to why the structure of the audio equipment inmdustry is the way it is, you have simply used terms recognizable from the economics literature to express ideas which have been expressed in the prior 500+ pages multiple times. What an economist's perspective probably underplays, relative to the expertise of the contributors here, is the extent to which perceptual biases and "testing" outside of controlled conditions leads consumers to believe that objective differences in audio reproduction are present when in fact they are not.

I really appreciate the summary, thank you! Looks like I have some more reading to do.

If I'm also using the DAC as a preamp into my Hypex ncx500s, are there additional things I need to be looking at besides just the DAC measurements? Would say, a Wiim Ultra be just as good as a MiniDSP SHD or even my current NAD C658 (which measures poorly compared to top DACs, but is still transparent)?

 

[antcollinet]

Would say, a Wiim Ultra be just as good as a MiniDSP SHD

From the DAC point of view, absolutely. The mini DSP might have the edge when it comes to DSP - especially with the Dirac option. And of course - it has balanced interconnect which may be a benefit if your system suffers from ground loops.

Even the Nad, with it's poor measurements will probably sound the same to almost all people in real world listening. Though those with the best hearing, listening with headphones under ideal test conditions may well be able to pick it out.

 

[Sal1950]

So if a transparent DAC had a bass boos function or even tone controls that would then suddenly become non transparent and show up in measurements?

Sure, my Emotiva DC1 DAC has a movable circuit jumper to introduce a +3db @ 20hz boost for headphone listening if desired.
It's very appropriate to use with my Sennheiser HD 650's. I'd like a bit more.

Though those with the best hearing, listening with headphones under ideal test conditions may well be able to pick it out.

And even that remains debatable until proven by Controled Bias DBT listening.
So many claims, so little evidence. LOL

 

[kemmler3D]

What is stopping a DAC from being transparent with measurements but also emphasising (just for example) the upper mids and sub bass, for a certain signature.

Can this happen? If not, can someone explain so I understand it.

The standard measurements will always catch that kind of thing, with one (nonexistent as far as I know) type of exception:

If there is distortion that's very dynamic in nature, i.e. it only shows up on peaks or only at specific volume levels, AND it happens to emphasize bass or treble (possibly in theory) it could be audible but not show up in standard measurements.

Amir measures DACs at many volume levels, so for something like that to get missed it would have to be very exotic, and in fact I have never heard of anything other than a tube amp behaving that way.

So if a transparent DAC had a bass boos function or even tone controls that would then suddenly become non transparent and show up in measurements?

Yep.

If it "measures transparent" then that means to the best of our knowledge, you can't hear the difference between the input and output, i.e. it does nothing to the sound except what it's supposed to.

It's possible in theory for something to be hidden from normal measurements, but in practice this basically never happens.

 

[RW61485]

As I have gotten more into the hobby like everyone else I have read and seen endless commentary around DAC signatures. From the start it has always been impossible for me to notice any difference between DACs. Indeed, every DAC I have have trialed at home or at audio meetups must have been audibly transparent because I have never heard a difference. I initially thought I might have some problem with my hearing or that there might be variability across humans in their ability to interpret sound. There are variabilities in ear and ear canal shape so why can't there be differences between how how one person's brain interprets sound versus another?

I initially felt like an outlier, and this site was helpful in parsing through the BS. I find it interesting that there are so many people genuinely hearing a difference when even if I try I cannot. And I count it as a blessing as it's likely going to save money which I can devote to things that actually do make a difference like better speakers!

One other point around DAC signatures has also struck me as odd. Even if you submit DACs have different sound signatures, surely the setup of your listening area and your speaker selection will be many multiples more impactful than the DAC right? If a well-measured DAC has a THD of 0.0005% and a well-constructed speaker has 0.5% across most of the frequency spectrum, the speaker's distortion on the sound signature is 10,000x times the DAC's no? And even the speaker's distortion pales in comparison to whether the washing machine is running in the kitchen above my listening room! How could a DAC's signature EVER matter when the scale of impact we are talking about is so small relative to the other links in the audio chain?

 

[antcollinet]

I find it interesting that there are so many people genuinely hearing a difference

In nearly all cases, the difference they hear is not in the sound reaching their ears, but is created in the wetware between them. Not imagination, but perceptive/cognitive bias.

 

[Jim Taylor]

. I find it interesting that there are so many people genuinely hearing a difference ...

Remove the word "genuinely" and in its place insert the word "supposedly", and the statement will be correct.

In the past, there were abilities that were considered genuine ... scrying tea leaves or goat entrails, for instance. The people who did this considered their abilities as genuine, the leaders to whom they reported considered the abilities as genuine, and even the population at large believed the abilities to be genuine. These people were held in the highest regard, so much so that their words determined the fates of kings and emperors. To gainsay them was considered heresy or insulting, and could get you executed.

That pretty well describes the Subjectivist faction of audio even today. Their unfounded and uncontrolled opinions regarding hearing are no more "genuine" than the ability to foresee the future in a goat's guts.

Science long ago proved superstition, divination and voodoo were just based on ignorance. Through the efforts of untold numbers of scientists, we now have a better idea what is real, what is true, what is usable and what is not. It's time the audio community sees the same light.

“The root of all superstition is that men observe when a thing hits, but not when it misses.”
— Francis Bacon

 

[MaxBuck]

Without contradicting anyone here, it's worth pointing out that "inaudible" frequencies can have psychological effects. To the extent that double-blinded DACs might be perceived as sounding different (and I'm not claiming these necessarily exist), this phenomenon might be a contributing reason.

Inaudible high-frequency sounds affect brain activity: hypersonic effect - PubMed

Although it is generally accepted that humans cannot perceive sounds in the frequency range above 20 kHz, the question of whether the existence of such "inaudible" high-frequency components may affect the acoustic perception of audible sounds remains unanswered. In this study, we used...

pubmed.ncbi.nlm.nih.gov

 

[SIY]

Without contradicting anyone here, it's worth pointing out that "inaudible" frequencies can have psychological effects. To the extent that double-blinded DACs might be perceived as sounding different (and I'm not claiming these necessarily exist), this phenomenon might be a contributing reason.

Inaudible high-frequency sounds affect brain activity: hypersonic effect - PubMed

Although it is generally accepted that humans cannot perceive sounds in the frequency range above 20 kHz, the question of whether the existence of such "inaudible" high-frequency components may affect the acoustic perception of audible sounds remains unanswered. In this study, we used...

pubmed.ncbi.nlm.nih.gov

Oohashi. Let’s just say that it’s not exactly accepted as correct.

 

[MaxBuck]

Oohashi. Let’s just say that it’s not exactly accepted as correct.

That's not really a response based on science. Can you cite a well-researched rebuttal?

 

[JeremyFife]

That's not really a response based on science. Can you cite a well-researched rebuttal?

I also found that research a few months ago and was also excited at the time. The study has never been replicated and has no good basis to support the findings. Look a little deeper, it's not good science.

 

[antcollinet]

Remove the word "genuinely" and in its place insert the word "supposedly", and the statement will be correct.

I disagree. They are genuinly hearing a difference. It is just that difference is generated in the brain.

If you object to "hear", taking the view that involves ears only, rather than the full auditory system, you could replace "hear" with "perceive"