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

[Blumlein 88]

After now 176 pages still no progress in getting a glimpse of an answer in this serious matter

Audio as a hobby has, as many other things in life, 2 main aspects: A rational one and an emotional one. Although they are definitely 2 completely different things, they are often mixed up or even interchanged. IMHO part of the art of life is to let both parts work in harmony. so 1+1 can become 3. Unfortunately, quite often they don't work together or counteract each other and 1+1 stays 1 or even level each out.

IMHO both parts are playing their role in the threshold of perceiving differences. On the rational side we have the magic 120dB number, great! We can write it down. But on the emotional side there will be never a number we can write down.

Maybe recent participants in this thread could listen to the old threads where I posted 8th generation copies, and see if they hear a difference vs the original. The ADC needs to be good and the DAC needs to be good and you hear whatever they do increased by 8 passes thru each. None of the devices have -120 db SINAD either.

From earlier in this thread.

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

 

[Jimster480]

I may be wrong, but I have an idea that your comment about opamps may have been a trigger for at least one of the responses you received.

For one opamp to sound different to another in the same circuit implies something is amiss, and this is usually instability which can give the impression the opamp sounds different.

I agree that the filter employed can influence the sound, and indeed I believe that some manufacturers tailor the filter to yield their 'house sound' (i.e. not neutral).

Well when I was referring to OpAmps I was referring to the quality of the sound. However there are some boutique OpAmps that are designed to sound different. I have tested some of those from Burson in the past and they can indeed change the sound signature. I didn't ultimately keep them as I noticed other "issues". However with lower end OpAmps they can indeed make certain situations sound "dirty", especially higher frequency noises like the ones from Cymbals in rock or frequency sweeps in electronic music or high notes from violins in classical music. Some of the details are lost. However once high end OpAmps are used (like the ones which are commonly present in competent DACs) these problems go away.
I rolled OpAmps years ago just to see what the fuss was about and if I could change the sound at all. It definitely is possible, I just don't think its possible to change it for the better UNLESS maybe you have some old hardware with some old crap $0.3-0.7 OpAmp and you can upgrade to a "good" TI OpAmp that costs $2 lol
However the OpAmp that comes in the Topping A30 and D30 are already good enough, and even "upgrading" them to the top end TI OpAmp doesn't do anything.
I haven't wasted time exploring this further since that time.... I got a Burson Amp with a whole set of OpAmps and I rolled those and found either no difference, or that it would be worse / dirtier (since I was doing reviewing / testing for Burson for a while). Needless to say they weren't too happy with me and my testing lol

Lastly; as you mentioned filters, I think they are the biggest "differentiators" in DACs these days if you want to compare how they "sound". Because some DACs have changable filters and others don't and maybe you don't "like" the filter that the company selected for you.... I mean its hard for me to really think about this being a serious scenario, but it would be more plausible then detecting the actual sound difference between 2 modern DAC chips which are well implemented.

 

[DonR]

Well when I was referring to OpAmps I was referring to the quality of the sound. However there are some boutique OpAmps that are designed to sound different. I have tested some of those from Burson in the past and they can indeed change the sound signature. I didn't ultimately keep them as I noticed other "issues". However with lower end OpAmps they can indeed make certain situations sound "dirty", especially higher frequency noises like the ones from Cymbals in rock or frequency sweeps in electronic music or high notes from violins in classical music. Some of the details are lost. However once high end OpAmps are used (like the ones which are commonly present in competent DACs) these problems go away.
I rolled OpAmps years ago just to see what the fuss was about and if I could change the sound at all. It definitely is possible, I just don't think its possible to change it for the better UNLESS maybe you have some old hardware with some old crap $0.3-0.7 OpAmp and you can upgrade to a "good" TI OpAmp that costs $2 lol
However the OpAmp that comes in the Topping A30 and D30 are already good enough, and even "upgrading" them to the top end TI OpAmp doesn't do anything.
I haven't wasted time exploring this further since that time.... I got a Burson Amp with a whole set of OpAmps and I rolled those and found either no difference, or that it would be worse / dirtier (since I was doing reviewing / testing for Burson for a while). Needless to say they weren't too happy with me and my testing lol

Lastly; as you mentioned filters, I think they are the biggest "differentiators" in DACs these days if you want to compare how they "sound". Because some DACs have changable filters and others don't and maybe you don't "like" the filter that the company selected for you.... I mean its hard for me to really think about this being a serious scenario, but it would be more plausible then detecting the actual sound difference between 2 modern DAC chips which are well implemented.

Look inside an analog mixing desk at a studio and look at the op-amps. Almost every time it will be a NE5532. Less than 40 cents in bulk. It is an unfortunate fallacy in this hobby to equate cost with audible sound quality.

 

[egellings]

NE5532 & NE5534 have withstood the test of time. They're competent op amps.

 

[digicidal]

Without a 100% transparent recording and production process... minuscule variations in the playback decoding shouldn't have a significant impact on the listening experience. Substantial differences could... but for the bulk of DAC components, filters, etc. we're talking about differences of a few dB (at levels under -100dB) in a few harmonics and at worst 1-2% variation in a few frequencies (mostly those already outside human hearing). Even then, without the ludicrous proposition of selecting a specifc DAC per recording, any actual "signature" will be present all of the time - so (at best) as "damaging" as having the air conditioning on or living somewhere with traffic, birds, etc. outside.

Until I hear the same number of people arguing over the preferred posture of a musician in the studio, incorrect mic placement for a specific drum kit, or the fact that the vocalist clearly had too much of a specific beverage before the session... I'll continue to consider "DAC signature" appraisals as mostly masturbation. Especially in a world where speakers and headphones have differences significantly greater than those between OpAmps, etc.

 

[Jomungur]

Until I hear the same number of people arguing over the preferred posture of a musician in the studio, incorrect mic placement for a specific drum kit, or the fact that the vocalist clearly had too much of a specific beverage before the session... I'll continue to consider "DAC signature" appraisals as mostly masturbation. Especially in a world where speakers and headphones have differences significantly greater than those between OpAmps, etc.

True. You know what finally convinced me? Playing with Roon's equalization settings. Small, measurable adjustments to frequencies made differences in sound that far outweighed any perceived change resulting from switching DACs. Better to just make adjustments in the software if I wanted to change the sound.

 

[digicidal]

True. You know what finally convinced me? Playing with Roon's equalization settings. Small, measurable adjustments to frequencies made differences in sound that far outweighed any perceived change resulting from switching DACs. Better to just make adjustments in the software if I wanted to change the sound.

Exactly. While a truly broken DAC design might cause audible deviation from the source - room modes, transducer/enclosure deficiencies, etc. always do. The problem is that it's much more instantly gratifying to unpack and install a shiny new component, cable, etc. - than it is to spend an afternoon with a laptop and microphone running sweeps and tuning DSP/EQ settings.

 

[Glint]

A long time ago I had a "Valab NOS DAC". Sure as hell wasn't transparent, though I think that was the point. Was said to sound 'organic and analogue', but in truth it was simply 'lacking' any reason to exist, music-reproduction wise. I suspect it would have measured as badly as FM radio or something

 

[Palladium]

I'm wondering, does the varying output impedances of various line-out affect the sound even if the measured output voltages were the same?

 

[DonR]

I'm wondering, does the varying output impedances of various line-out affect the sound even if the measured output voltages were the same?

Ideally, the load impedance should be so high as to negate any changes in the output impedance of the DAC.

 

[solderdude]

I'm wondering, does the varying output impedances of various line-out affect the sound even if the measured output voltages were the same?

When they truly varied (a lot as output resistance of line-out is usually low) and the load resistance is low or also varying a lot then Ohm's law dictates the frequency response will not be flat. (voltage division).

Fortunately most line output stages consist of a buffer (usually an opamp) followed by a resistor. So the output resistance does not vary much within the audible band.
Another fortunate thing is that most amplifiers have a relatively much higher input resistance (within the audible range) and not a complex impedance such as speakers/headphones. This means that acc. to Ohm's law. the small voltage division that takes place (input resistance of amp and output resistance of source) is very, very small and linear (again within the audible range) and thus no change of tonal balance can/will occur.

So unless one has some really poorly designed output stage (transformers, high impedance tube circuits) and or a weirdly low input resistance (say a few hundred Ω max.) or the amp has a very unusual impedance (transformers unusual high capacitance) then it is not going to be an audible issue.
Measurable (as in very tiny variances well below 0.01dB or so) could well exist. Inconsequential for sound.

 

[audio_tony]

I'm wondering, does the varying output impedances of various line-out affect the sound even if the measured output voltages were the same?

An excessively high output impedance (greater than 5..10k) when used with a high capacitance cable, can result in some minor top end loss.

Such devices are thankfully rare, although some valve gear can have a high output impedance (depends on design, and if transformers are used).

So called passive preamps can suffer from the same issue, as the output impedance will vary according to the position of the pot.

 

[Wseaton]

I apologize if I've responded already on this, but to rehash my experience on this:

About 20 years I used to sell high end audio. I only did it for a bit to make some extra cash, but I always had VIP access and made some great friends with the sales staff. If they needed a clinical set of ears we sat around after we were closed many night with a case of beer and did A/B testing until our wives starting calling.

One of the perplexing issues was differences between DACs sonically. Sometimes the differences were dramatic. Sometimes the differences were minor.

Case in point. I bought Toshiba DVD player based on strong video quality reviews, and indeed it was strong in that suite. Bought it to replace by aging Sony MDM laser disc player. I long since switched to DVD, but the MDM had Sony's much revered 1-bit DAC, and it was a superb reference player.

After hooking up the Toshiba I noticed that it just didn't sound right. Was running some B&W Matrix speakers at the time on a B&K stack, and though my gear wasn't the problem. The Sony sounded fine. The Toshiba sounded flat and compressed and the soundstage lacked resolution.

I brought the Toshiba back to the shop because my old buddies were dying to hear what I was griping about. Plus, they sold me the thing. Through some finagling of gear we were able to set about an A/B/X using the toshiba as a transport, but using the toslink and coax outputs to different units. We set up another Sony CD player running the 1-bit DAC like my MDM and then through a Adcom GDA 600 external DAC. We could then flip through the inputs quickly and hear all the DACs seamlessly. We did the same test on a variety of speakers including a pair of Wilson's somebody had traded in along with a pair of Legacy Whispers. The later being one of the most dynamically resolving speakers I've ever heard.

The DAC in the Toshiba was clearly having issues. The soundstage was smaller, lacked detail, and instruments we're somewhat compressed together. The Sony 1-bit had excellent detail, but sounded a bit 'dark' if I can pick a term. The Adcom was the best of the bunch. When we flipped to the Adcom the soundstage literally bloomed outward dramatically and expanded in depth quite a bit. We were all a bit dumfounded. We found the same problem with some other midfi units, but I don't want to give names.

Somebody then got smart and pulled out a Pioneer Elite disc player given the Pioneer uses the same Burr Brown DAC as the Adcom GDA. The sound was identical. Other players using the same Sony 1-bit were identical.

I realize this was awhile ago, but it was a big holy sh_t moment for all of us, and I wish I still had access to all that gear to ABX stuff. I have never, ever got a technical explanation on why those DACs sounded distinctly difference. Equalization was the same. Line leveling was the same, but the perceptual resolution and soundstaging was staggeringly different. I have always wanted to discuss this with an electrical engineer who's familiar with DAC design and flesh out what is happening electronically to cause the difference. What is causing the toshiba DAC to sound mushy and flat, and can this actually be tested. I'm sure most of the bits apply to modern DACs.

Note that I've heard DAC / Amp combos from various friends, and what they claim is different is to my ears is differences in amplifier / equalization. Boosting the treble might work for B&W, but it doesn't full me. That's not the DAC. It's something after the fact, IMO.

 

[Jimster480]

Look inside an analog mixing desk at a studio and look at the op-amps. Almost every time it will be a NE5532. Less than 40 cents in bulk. It is an unfortunate fallacy in this hobby to equate cost with audible sound quality.

You are right, but there are still better products that do produce better cleaner sound. NE5532 is quite good for its price though, but there are worse Op Amps used in some products.

Without a 100% transparent recording and production process... minuscule variations in the playback decoding shouldn't have a significant impact on the listening experience. Substantial differences could... but for the bulk of DAC components, filters, etc. we're talking about differences of a few dB (at levels under -100dB) in a few harmonics and at worst 1-2% variation in a few frequencies (mostly those already outside human hearing). Even then, without the ludicrous proposition of selecting a specifc DAC per recording, any actual "signature" will be present all of the time - so (at best) as "damaging" as having the air conditioning on or living somewhere with traffic, birds, etc. outside.

Until I hear the same number of people arguing over the preferred posture of a musician in the studio, incorrect mic placement for a specific drum kit, or the fact that the vocalist clearly had too much of a specific beverage before the session... I'll continue to consider "DAC signature" appraisals as mostly masturbation. Especially in a world where speakers and headphones have differences significantly greater than those between OpAmps, etc.

Just isn't that simple. Music contains many different sounds at many different frequencies at the same time. The reproduction of that sound can lead to things sounding different. These small differences can cause listener fatigue (especially with excess distortion often being exacerbated by higher frequency tones, which can come off as "sharp" or "harsh").
When listening to classical music (especially recently recorded tracks with proper mastering) such as some of the songs from Lang Lang, Itzhack Perlman or Yo Yo Ma (or even all 3 in the same song) it is much "easier" to pick out differences if there are any audible ones. The violin as well as a few higher notes of a proper Piano have a real knack for exposing weakness.
There are also some songs in the world of Trance & Progressive where the many beats at once can expose other flaws that are not commonly present. There are many songs with 400+BPM (or even 250BPM) which means more than beats per second (not including other background noises).
Over the years I have put together a playlist which has specific points in songs which can draw out certain flaws with setups, whether DAC or Amp or DAP or some other device (or even a headphone itself).
As I mentioned before; it doesn't mean that the flaw is always present or that it becomes annoying quickly (it depends) but there are sometimes differences that can be heard and sometimes these differences can cause listener fatigue or other annonyance which can disrupt concentration (I always typically listen to music while programming).

 

[CapMan]

Exactly. While a truly broken DAC design might cause audible deviation from the source - room modes, transducer/enclosure deficiencies, etc. always do. The problem is that it's much more instantly gratifying to unpack and install a shiny new component, cable, etc. - than it is to spend an afternoon with a laptop and microphone running sweeps and tuning DSP/EQ settings.

I could not agree more with this post. Spot on sir

 

[CapMan]

Over the years I have put together a playlist which has specific points in songs which can draw out certain flaws with setups,

I use Day Is Done on the Brad Meldhau album of the same name as a bass test - there is a double bass solo that should be centred between speakers, before I fixed my room acoustics and used DSP some of the double bass notes would jump from centre to hard right like some crazy 70s prog rock LP as a result of room modes and other nasties

 

[DonR]

You are right, but there are still better products that do produce better cleaner sound. NE5532 is quite good for its price though, but there are worse Op Amps used in some products.

No, it's pretty much past the threshold of audibility which is why it's used everywhere. There are better measuring op amps but all the improvements are beyond audible. There may be worse op amps used as well but why bother when an audibly sufficient one is available for so cheap?

 

[Jimster480]

No, it's pretty much past the threshold of audibility which is why it's used everywhere. There are better measuring op amps but all the improvements are beyond audible. There may be worse op amps used as well but why bother when an audibly sufficient one is available for so cheap?

Well I actually tested this specific op-amp next to a TI version and I found that the ti version did indeed sound better in complex high frequency audio scenarios. Or in a scenario where there is fast base on the bottom and then a frequency sweep to the top or a high note played simultaneously. Like if you have drummers in the back in a classical song where there is a violin being played hitting a high note.
However these details are only typically present on Modern high-res recorded tracks. So the number of songs with which you could run into anything that could be Audible with this off amp is very small.

I remember also some years ago when I was testing cheaper Chinese products that there are products that use a more expensive DAC chip and TI op amp but the overall topology is not that good or the off-amp is actually too powerful for the circuit and putting in a less powerful one results in better performance. I have some documentation about what I did years ago somewhere and I will try to find it to post here. Because I forget the model that I improved by actually reducing the op-amp that was in there.

 

[mansr]

Uh-huh.

 

[DonR]

Well I actually tested this specific op-amp next to a TI version and I found that the ti version did indeed sound better in complex high frequency audio scenarios. Or in a scenario where there is fast base on the bottom and then a frequency sweep to the top or a high note played simultaneously. Like if you have drummers in the back in a classical song where there is a violin being played hitting a high note.
However these details are only typically present on Modern high-res recorded tracks. So the number of songs with which you could run into anything that could be Audible with this off amp is very small.

I remember also some years ago when I was testing cheaper Chinese products that there are products that use a more expensive DAC chip and TI op amp but the overall topology is not that good or the off-amp is actually too powerful for the circuit and putting in a less powerful one results in better performance. I have some documentation about what I did years ago somewhere and I will try to find it to post here. Because I forget the model that I improved by actually reducing the op-amp that was in there.

Double-blind? Did you take measurements? NE5532 has a gain-bandwidth product of 10MHz so it is doubtful there is any benefit in hi-res. What do you mean by "too powerful"? Did it draw too much current?