Does DSD sound better than PCM?

[noel_fs]

So you think images above the audible band, which is 20 kHz and not 100 kHz, at - 85dB are of audible consequence?

Do you have any evidence of that, or is this just a theoretical desktop exercise?

Sorry, Im not following you with the relative bit. If an image in the transition band is at - 85dB then it's at - 85dB. Pretty much inaudible even if it were in the audible range, but it is above the audible range. What you are forgetting is that numbet was relative to full scale signal, which you simply will not have in the transition band.

I'm evaluating what's of concern based on engineering and acoustic knowledge, it's not arbitrary.

Very few people can hear above 20kHz.

If you look at good amplifiers a full scale 19+20khz imd test may create in band tones below - 110dB. Transition band images at significantly lower levels will produce IMD products a lower levels again. You shouldn't conflate amplifier performance with this filter stuff.

Hey, i have a question for you.

If you need to pay 40$ for a bill, do you prefer to have in your wallet 40$ or 60$

 

[Blumlein 88]

Hey, i have a question for you.

If you need to pay 40$ for a bill, do you prefer to have in your wallet 40$ or 60$

Well using audiophile logic. I'd need at least $600,000 to be able to spare $40 for a bill.

This is not what Miska is saying. And I don't think Alan would be at all averse to having all spurious tones in or out of band below - 120 db. I'd have to wonder about our ability to hear -85 db in the midst of any other sound. Usually at normal listening volumes you'll not hardly hear it at all even if it this low level sound were presented along by itself.

 

[Blumlein 88]

Okay here is a few seconds of a FLAC file. It has a Zildjian Fast Crash cymbal recorded at 176 khz rates using a wide bandwidth microphone. I forget if it were an Earthworks or another. This is recorded at a HIGH LEVEL you have been warned. Don't play it too loud at first. Prior to the loud cymbal is the same thing reduced by 80 db. And a 3 khz tone reduced to -80 db. You can listen and see which comes first the tone or reduced cymbal crash. But don't turn it up too high because the loud cymbal crash a couple seconds later is LOUD.

 

[March Audio]

Given for example said 22.8 kHz, and it's aliasing product at 21.3 kHz, it is not down by -85 dB in this case. And yes I consider 21.3 kHz likely audible for notable portion of population on earth (at least more than 1 out of entire population).

But again, I don't evaluate audibility of N within population X people. That is much harder evaluation to do than just fixing the problem, instead of speculating about audibility. More, evaluating audibility of some objectively found problem is trying to find excuse for not fixing an obvious problem.



How am I ignoring the real world? If I can fix something I can find, objectively, as fault in real world, I will fix it.

Can you describe why you are against fixing some obvious problems?

For me, this is non-issue. It is one of the many things I've fixed already and I move on to many of the other issues to be fixed. No need to make a lot of posts about one particular fix, which is among the easier ones.



In real world (pun intended) not everybody has a perfect amplifier... How about playing nice with those with less than perfect ones too?

Otherwise we very quickly move to talk about amplifiers and then I start hearing statements "if a DAC has good quality it doesn't output ultrasonic images". Just IMO, you cannot push problem of component A to shoulders of component B. Every component has enough of design challenges in it's own domain.



I don't see it that way. Usually 2+3 kHz twin tone is quite different from 19+20 kHz win tone, both have the same 1 kHz difference.


But really, I don't know why I'm having this conversation. It is off-topic, except for the nice detail that with properly done DSD content you don't have digital decimation filters (so whatever we discuss doesn't exist)... You can make you own decisions about what is important to you and I make mine. When I can objectively measure differences, I'm usually fine talking about those, from mathematical perspective. If you want to tell people what they can and cannot hear, it's your decision. Personally, based on my experience, I don't do that. People have surprised me enough many times.

You can "consider" all you like, doesn't make it in any way a realistic proposition. There is no speculation here, and nothing difficult about it. We know that few people can hear above 20kHz at high signal level, let alone any signals at low level (the example being - 85dB).

You are ignoring the real world if you try to fix a problem that isn't actually a problem.
Let's make a dac with a SINAD of 200dB. Great idea. Shame no one will hear the difference between that and the dac with 120dB SINAD.

You are correct, not everyone has a perfect amplifier. However an amp that has an IM problem with inaudible signals at 20 to 22 kHz also has problems with IM with signals in the audible band. Normal audible signals will cause you a problem. The dacs low level of transition band output is the least of your worries. The amp distorts with whatever input you feed it. Blaming the dac it's totally inappropriate and erroneous.

Out of band signals are a fact of life. We are surrounded by RF. As such an amplifier should be designed to deal with these signals. It's a normal condition of operation. That's good design. Whilst it is good to have a dac with minimal RF output (no argument there) the dac is not the only source of problem that needs to be dealt with

If we take your view to its logical extreme (which is where you appear to be) a dac would have no US output and an amp would have infinite bandwidth with no phase shift and no input filtering. Not a realistic proposition. Oh sorry forgot to add, followed by the infinite bandwidth phase linear speaker........preceeded by infinite bandwidth recordings that have no phase shift in any part of the recording chain.

I don't understand your later point about IM.

You are misunderstanding me. I am not against you fixing issues. However you have made some statements that don't stack up, such as about the transition band signals. Or ascribing importance where there is little/no issue. I'm questioning rather than denigrating what you do. That's what this forum is about and why we are having this conversation.

Also, my view on what people can hear isn't an arbitrary position, it's based on decades of research by many different people. It's your position that signals above the accepted audible range at extremely low levels are of significance - ie audible. So really it's up to you to back up that assertion with some evidence.

 

[March Audio]

I agree with you in principal, but there is the constraint of what is possible. Reducing those constraints is good. But for instance if we took your approach to amplifier distortion you'd be unhappy until there is none. There are at least as many ways amps can be tripped up as can ADC and DAC. But if you apply their use well distortion is low enough not to be heard with good amps, and any pathological conditions have been made so rare they aren't much of a concern.

We need people like you trying to reach perfection, but the phrase don't throw the baby out with the bathwater comes to mind.

This. It's the point I made originally about bigger fish to fry. Perspective.

 

[March Audio]

Well using audiophile logic. I'd need at least $600,000 to be able to spare $40 for a bill.

This is not what Miska is saying. And I don't think Alan would be at all averse to having all spurious tones in or out of band below - 120 db. I'd have to wonder about our ability to hear -85 db in the midst of any other sound. Usually at normal listening volumes you'll not hardly hear it at all even if it this low level sound were presented along by itself.

That - 85dB we are all quoting is based on a 0dB signal. You would never see a 0dB signal in the transition band. What aliasing levels do you get if you try a - 60dB signal which would be a more realistic level?

 

[March Audio]

Funnily enough I have just been made aware of this article which very neatly confirms my point. With measurements showing real world music signals (not never encountered pathological high level high frequency test signals) there is not an issue.

http://archimago.blogspot.com/2019/02/musings-measurements-on-why-2496.html

Lets be clear I am what you might call a "measure-ist", I want to explore technical performance in many ways, stressing equipment to see what happens. However this has to be tempered by relating it to real world usage and operation.

You might test an amplifier into a highly capacitive 0.5 ohm load, it might be academically interesting to find out how it behaves. If it performs well you could claim what a marvelous amplifier it is. Problem is there are no real world speakers that present that sort of pathological load, so the test has produced data of no actual value.

 

[Blumlein 88]

Funnily enough I have just been made aware of this article which very neatly confirms my point. With measurements showing real world music signals (not never encountered pathological high level high frequency test signals) there is not an issue.

http://archimago.blogspot.com/2019/02/musings-measurements-on-why-2496.html

Lets be clear I am what you might call a "measure-ist", I want to explore technical performance in many ways, stressing equipment to see what happens. However this has to be tempered by relating it to real world usage and operation.

You might test an amplifier into a highly capacitive 1 ohm load, it might be academically interesting to find out how it behaves. If it performs well you could claim what a marvelous amplifier it is. Problem is there are no real world speakers that present that sort of pathological load, so the test has produced data of no actual value.

I own some Soundlab electrostatics. They present more like a .75 ohm capacitive load.

 

[March Audio]

I own some Soundlab electrostatics. They present more like a .75 ohm capacitive load.

Arse!

I knew some clever dick would spoil things Im aam revising my post to 0.5 ohms.

Which model is it? Are you sure its 0.75 ohms?

Thats not actually a speaker, thats a link of wire. That will cause a world of pain for amps.

 

[Miska]

You are misunderstanding me. I am not against you fixing issues. However you have made some statements that don't stack up, such as about the transition band signals. Or ascribing importance where there is little/no issue. I'm questioning rather than denigrating what you do. That's what this forum is about and why we are having this conversation.

Also, my view on what people can hear isn't an arbitrary position, it's based on decades of research by many different people. It's your position that signals above the accepted audible range at extremely low levels are of significance - ie audible. So really it's up to you to back up that assertion with some evidence.

I'm happier with fixable things fixed and will continue fixing such things to improve performance. I don't care if you or someone else consider something little/no-issue, that's your opinion. Any issue that can be detected is an issue to be fixed for me and of significance to me.

Mine is also based on decades of R&D... So far I've been showing evidence, not seeing much from your side.

I'm failing to understand why not fix something.

If we take your view to its logical extreme (which is where you appear to be) a dac would have no US output and an amp would have infinite bandwidth with no phase shift and no input filtering. Not a realistic proposition.

There are already D/A conversion solutions that have no ultrasonic noise or images in the output. I've shown one example of such earlier in this thread. And for amps, you can take a look at for example Spectral amps.

Just as a starting point.

Let's make a dac with a SINAD of 200dB. Great idea. Shame no one will hear the difference between that and the dac with 120dB SINAD.

I would still buy the one with 200 dB SINAD, if it performs equally better in the other aspects too. Like SINAD would be only thing that matters... There are so many other things too that people don't seem to be even looking at at all.

But I will now leave you with your opinions, I don't want to waste my time more on this pointless discussion. I have more things to improve and I know how to get there!

I'd have to wonder about our ability to hear -85 db in the midst of any other sound. Usually at normal listening volumes you'll not hardly hear it at all even if it this low level sound were presented along by itself.

Aliasing in transition band is much higher with those hb-filters as you checked yourself.

You can actually compare this by using some of the NOS R2R DACs running at 44.1k sampling rate and see if you can hear difference. They typically have a lot of out of band distortion and also lot of IMD products at about those levels. At some point (when I find time) I could play some of your reference tracks through such and record it with ADI-2 for comparison.

Based on statements from both of you it should sound perfectly fine.

 

[March Audio]

I'm happier with fixable things fixed and will continue fixing such things to improve performance. I don't care if you or someone else consider something little/no-issue, that's your opinion. Any issue that can be detected is an issue to be fixed for me and of significance to me.

Mine is also based on decades of R&D... So far I've been showing evidence, not seeing much from your side.

I'm failing to understand why not fix something.



There are already D/A conversion solutions that have no ultrasonic noise or images in the output. I've shown one example of such earlier in this thread. And for amps, you can take a look at for example Spectral amps.

Just as a starting point.



I would still buy the one with 200 dB SINAD, if it performs equally better in the other aspects too. Like SINAD would be only thing that matters... There are so many other things too that people don't seem to be even looking at at all.

But I will now leave you with your opinions, I don't want to waste my time more on this pointless discussion. I have more things to improve and I know how to get there!

Aliasing in transition band is much higher with those hb-filters as you checked yourself.

You can actually compare this by using some of the NOS R2R DACs running at 44.1k sampling rate and see if you can hear difference. They typically have a lot of out of band distortion and also lot of IMD products at about those levels. At some point (when I find time) I could play some of your reference tracks through such and record it with ADI-2 for comparison.

Based on statements from both of you it should sound perfectly fine.
View attachment 22478
View attachment 22479

Oh I could supply endless evidence that tallies with Archimagos data in the link above. It's a simple fact that music does not contain lots of high level high frequency content.

I am interested in the behavior with representative signals.

You have created a problem where one doesn't exist by using test signals that never appear in the real world.

Your presentation of data this way is misleading, but hey keep on fixing stuff that doesn't need fixing.

I'm not sure how you are deriving your data but when you play real music

Problem? What problem?

 

[Veri]

Based on statements from both of you it should sound perfectly fine.

Sounds fine to me

I think I am IMD-tone deaf guys.

 

[Miska]

Oh I could supply endless evidence that tallies with Archimagos data in the link above.

Your own please, instead of copying other people's work.

It's a simple fact that music does not contain lots of high level high frequency content.

I already gave you figures from music that contains high level high frequency content.

I'm not sure how you are deriving your data but when you play real music

Problem? What problem?

I already explained why this kind of plot easily leads you to fallacy about frequency content. Music signals and their frequency content is not stationary. You think four fairly randomly selected songs are representative of all that is out there? Although for example Wild World and Le Mal de Vivre here already shows relatively high normalized level at 20 kHz.

Example, I have a complex test tone I use for testing certain things. It contains repetitive Dirac pulses at about -1 dBFS level, which is full-flat spectrum throughout. The spectrum plot from Audacity looks like this:

Not much high frequency content! Allthough the peak you can see in the waveform contains all frequencies flat from 0 to 22.05 kHz and has amplitude of 0.8, yet it's level in this spectrum plot looks like -91 dB. This is because it's energy is concentrated for such short period compared to other, lower level signals.

If we zoom in and look at shorter period around it and make FFT shorter, it comes up a lot:

 

[svart-hvitt]

I'm happier with fixable things fixed and will continue fixing such things to improve performance. I don't care if you or someone else consider something little/no-issue, that's your opinion. Any issue that can be detected is an issue to be fixed for me and of significance to me.

Mine is also based on decades of R&D... So far I've been showing evidence, not seeing much from your side.

I'm failing to understand why not fix something.



There are already D/A conversion solutions that have no ultrasonic noise or images in the output. I've shown one example of such earlier in this thread. And for amps, you can take a look at for example Spectral amps.

Just as a starting point.



I would still buy the one with 200 dB SINAD, if it performs equally better in the other aspects too. Like SINAD would be only thing that matters... There are so many other things too that people don't seem to be even looking at at all.

But I will now leave you with your opinions, I don't want to waste my time more on this pointless discussion. I have more things to improve and I know how to get there!



Aliasing in transition band is much higher with those hb-filters as you checked yourself.

You can actually compare this by using some of the NOS R2R DACs running at 44.1k sampling rate and see if you can hear difference. They typically have a lot of out of band distortion and also lot of IMD products at about those levels. At some point (when I find time) I could play some of your reference tracks through such and record it with ADI-2 for comparison.

Based on statements from both of you it should sound perfectly fine.
View attachment 22478
View attachment 22479

@Miska , you must understand, @March Audio is a hardware producer with a profit motive. So he will make a DAC Mk1, which he says is transparent, only to make a DAC Mk2 at a higher price at a later point.

So even if he says his current dac is transparent, he will make transparenter dacs in the future.

I don’t see why he makes so much fuzz about a software package at €150, while his own hardware is much more expensive - and he even has plans to obsolete his own hardware in the future.

I still haven’t seen @March Audio ’s evidence that @Miska ’s software doesn’t do what it says or that the software does harm.

 

[Miska]

@Miska , you must understand, @March Audio is a hardware producer with a profit motive. So he will make a DAC Mk1, which he says is transparent, only to make a DAC Mk2 at a higher price at a later point.

So even if he says his current dac is transparent, he will make transparenter dacs in the future.

Sure, I'm a bit confused about the response though.

I first explained how I think particular choice of ESS digital filter in his DAC fixes certain common problem from source material used in the test tracks created by @Blumlein 88 . I'm kind of surprised of the response. Based on the response he could of course also go back to his chip vendor (ESS) and complain that they in first place even offer any filter options at all because none of that matter. Instead of complaining about that to me... I have my opinion about the matter and he is not going to change that anyway.

But this thread has gone massively off-topic at least. The only relevance to OT is that with DSD ADC you don't need digital decimation filters at all so this issue being discussed now doesn't exist.

 

[svart-hvitt]

Sure, I'm a bit confused about the response though.

I first explained how I think particular choice of ESS digital filter in his DAC fixes certain common problem from source material used in the test tracks created by @Blumlein 88 . I'm kind of surprised of the response. Based on the response he could of course also go back to his chip vendor (ESS) and complain that they in first place even offer any filter options at all because none of that matter. Instead of complaining about that to me... I have my opinion about the matter and he is not going to change that anyway.

But this thread has gone massively off-topic at least. The only relevance to OT is that with DSD ADC you don't need digital decimation filters at all so this issue being discussed now doesn't exist.

The beauty of software is that an update will give you a «new» product.

Hardware - even when marketed as «transparent» - will be obsoleted when you go from Mk1 to Mk2, Mk3...etc.

So from a sustainability point of view, software has more beauty in it. Hardware will overcrowd our landfills.

Maybe there’s a dose of software envy here?

 

[solderdude]

@Miska

First let me say I admire the desire to improve on things, whether they may become audible in some specific situations or not.
I appreciate the efforts and above all that you base on measurements.
The fact that you leave the perceptual part out of it is commendable and questionable at the same time.
The below is not 'bashing' on my side but simply asking questions that popped up in my small penguin brain as it were.
You were obviously compelled to start your search in this area .. what triggered it ?
I mean you must have 'heard' something and related it to something and investigated it.

Why not measure into the GHz range ?
In the many EMC tests I have done often there is a lot of life well above 5MHz.
For these frequencies to become audible, however, they need to be large enough to be rectified (as in AM detector) in general. -80dB signals do not qualify. These frequencies travel like a knife through butter across PCB's and cables.

Most amplifiers do start to roll-off above 100kHz though some do not. granted this is mostly 6dB or 12dB/octave as input filters are generally only 6dB/octave and the amp itself as well.
Do you have any practical shots/analyses of certain (preferably often used/well known) amplifiers loaded with actual speakers/headphones that clearly show problematic behavior of amplifiers when fed signals in the audible range and say around 1MHz combined which seem to show that when the 1MHz signal (preferably not a multiple of the test tone) is included or not that the output of the amp differs within the audible range ?
I assume your very extensive research has a basis and is clearly documented.
Preferably with the main frequency say -6dB and the higher frequency at -60dB or so.

Do you have any indication how big this issue is and how relevant it is to music reproduction ?
I mean one can measure well outside of any audible threshold. What one can measure does not need to be audible. Some claim vice versa as well but I am not one of those people.

Examples of amps that are known to exhibit this behavior ?

You mention

So any intermodulation products they may create in later stages (like power amp) are also fully correlated. So if you hear it sounds like artificial hardness.

Is this your experience ?
Can you clearly correlate this ?
What amplifier / transducers are you using to listen for this (so I can avoid that one) as you clearly state they MAY create it I assume you own one to test.

You say:

We cannot know what kind of amps there are after the DAC, so we need to account for all kinds of possibilities, such as analog class-D amps, etc. One dominating factor is how the amp's THD/IMD vs frequency profile looks like.

Agreed. What would be so poor THD/IMD that people could possible hear effects and what would constitute a good profile.
You seem to suggest that the often reported 'synergy' between DAC's and amps (almost infinite possibilities here) could (maybe) be related to this.

This should be very possible to capture with an RME after an amp when loaded with a real load.
Do you happen to know of any such recordings that show this is the case ?

You mention:

Looking at my own tests,..... Last detectable output from microphone for this instrument is at 65.7 kHz before harmonics disappear in increasing noise slope of the ADC noise shaper. But my measurement microphone is not claimed to have flat frequency response that high anyway... Too bad I don't have money for the Sanken CO-100K...

Do you correlate listening tests with measurements ?
Are those listening tests 'blind'.
It's one thing to measure HF content but another thing to perceive it.
What's your current age (may I ask) and what is the max. frequency you can hear and what is the max. bandwidth you need to correlate recorded and actual sound quality ?

You state:

World is full of all kinds of amplifiers. These transition band (20 - 22.05 kHz) images are pretty much around -3 dB at 22.05 kHz. But you'll have much more problems with the images around 352.8 kHz.

Do you have plots/graphs/examples indicating this within the audible band (preferably using nulling) ?

As mentioned before... I appreciate the effort and feel that in principle any effort to objectively improve output quality is a good thing.[/quote]

 

[March Audio]

Your own please, instead of copying other people's work.



I already gave you figures from music that contains high level high frequency content.



I already explained why this kind of plot easily leads you to fallacy about frequency content. Music signals and their frequency content is not stationary. You think four fairly randomly selected songs are representative of all that is out there? Although for example Wild World and Le Mal de Vivre here already shows relatively high normalized level at 20 kHz.

Example, I have a complex test tone I use for testing certain things. It contains repetitive Dirac pulses at about -1 dBFS level, which is full-flat spectrum throughout. The spectrum plot from Audacity looks like this:
View attachment 22481

Not much high frequency content! Allthough the peak you can see in the waveform contains all frequencies flat from 0 to 22.05 kHz and has amplitude of 0.8, yet it's level in this spectrum plot looks like -91 dB. This is because it's energy is concentrated for such short period compared to other, lower level signals.

If we zoom in and look at shorter period around it and make FFT shorter, it comes up a lot:
View attachment 22482

I completely independant came to the same conclusion as Archanigo. I had no idea he had already gone through the same conversation with you until someone pointed out to me this afternoon.

Your whole premise is based on high levels of high frequency signal which simply doesn't exist in real music.

Would you like me post post dozens of spectra of different tracks to demonstrate this, or will you just accept what you already know to be true?

I am fully aware of Fft processing limitations in the time domain. That still doesn't mean that there is significant high level high frequency content in music. Try high pass filtering and look in the time domain if you like.

 

[March Audio]

@Miska , you must understand, @March Audio is a hardware producer with a profit motive. So he will make a DAC Mk1, which he says is transparent, only to make a DAC Mk2 at a higher price at a later point.

So even if he says his current dac is transparent, he will make transparenter dacs in the future.

I don’t see why he makes so much fuzz about a software package at €150, while his own hardware is much more expensive - and he even has plans to obsolete his own hardware in the future.

I still haven’t seen @March Audio ’s evidence that @Miska ’s software doesn’t do what it says or that the software does harm.

That's pretty low. The same argument can apply to miskas product - taking advantage of naive audiophiles - but I wouldn't dream of suggesting it is the case.

It is a solution to a theoretical problem. If you believe it is a problem you can purchase the software. It is of no threat to my product sales. How could it be? You still need a dac to use the software.

I am challenging the propositions that miska is presenting. That's what this forum is about.

 

[pjug]

@Miska

Do you have any practical shots/analyses of certain (preferably often used/well known) amplifiers loaded with actual speakers/headphones that clearly show problematic behavior of amplifiers when fed signals in the audible range and say around 1MHz combined which seem to show that when the 1MHz signal (preferably not a multiple of the test tone) is included or not that the output of the amp differs within the audible range ?

This is what I would like to see too. Not that it should be on Miska to do it but it would be great if someone would make these measurements. I think both @Miska and @March Audio have good points, but without measurements it seems the conversation can just go around and around.