• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Does DSD sound better than PCM?

Status
Not open for further replies.

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,643
Likes
12,874
You could look here for the aliasing behavior of the ADC I've been using.
https://www.audiosciencereview.com/...aring-aliasing-in-three-adcs.3272/#post-80806

The Zen Tour has aliasing around -88 to -90 dbFS with a Max ultrasonic signal. Of course this aliasing drops if the ultrasonic signal drops. Otherwise at least up to the 96 khz bandwidth I can check it has no surprises. The ESS9038 seems well behaved from what I've seen past the nyquist rate.

Miska checks on up for signals in the megahertz region. I've seen some of what he checks. I'm not convinced it is something that has consequences to us listening to music. But some DACs are not well behaved up there. Miska has pointed out recording the output of the DAC with an ADC filters out ultrasonics that would otherwise be present. So in effect it cleans up any such effects that might cause issue with the rest of the electronics in the chain. A wide bandwidth power amp or feedback loops on class D some such could conceivably be effected. So listening to my recorded files wouldn't be identical to listening to the DAC itself due to this filtering. I believe he suggests it would be best to record 44.1 material at least out to 384 khz sample rates. I don't have any ADCs that run above 192 khz. I could record at 192 to get some of it.

As has been said, people aren't having an easy time hearing differences in 8th and original. The original playback will have whatever ultrasonic artifacts are present in the listener's gear. My 8th gen copies will have that too even if they don't have the artifacts the March DAC itself has. You would be recording the compounding effects of distortion, noise, uneven FR effects, and jitter. Perfect it is not. I can even hear the 8th gen copies, but not easily. I can't hear 1st gen copies as different vs the original.
 

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,643
Likes
12,874
:) Im making no claims about my dac, it is an ESS9038 implementation so very representative as to whats currently out there from other manufacturers.

...and yes people are very much struggling to identify which is which, I have seen the results. I think it highly unlikely that my dac is fixing any issues of the ADC or the original recordings. Thats really not plausible proposition that it provides an inverse to those issues.
In any case surely it will still have the DAC filter issues you allude to? Plus you have already stated that the differences are readily visible in an audio editor.

So if we further band limit and filter (repeatedly) do we not have worse problems with phase?

Perhaps we could ask @Blumlein 88 to repeat the test with a square wave?

I think the point I would make is that I am most certainly not denigrating your work, however I think there are bigger fish to fry - it doesnt address the issue anywhere else in the chain beyond the DAC, which means it can only have very limited impact, but also that we need to establish correlation to audibility to say if there is actually any impact.
We do have worse phase problems with more copies. As for the square wave I could do that 8 times. My expectation would be once you've bandlimited it the first time thru, further trips won't make lots of difference except in the transition zone. If I get a chance to do it what frequency would you want. 7 khz, or 1 khz?
 

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,643
Likes
12,874
PCM and DSD are so 20th century.

Why isn't this thread asking about MQA?
Boooooo! Hissssssss!
MQA be gone.

Haven't noticed you posting recently @watchnerd or maybe I'm hanging out in the wrong threads. Hope all is well.
 

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
5,764
Likes
7,383
Location
Albany Western Australia
We do have worse phase problems with more copies. As for the square wave I could do that 8 times. My expectation would be once you've bandlimited it the first time thru, further trips won't make lots of difference except in the transition zone. If I get a chance to do it what frequency would you want. 7 khz, or 1 khz?
I suppose 7kHz, but dont put yourself to trouble on this, it was more of a musing which I already know the answer to ;) .
 
Last edited:

Miska

Active Member
Joined
Feb 20, 2019
Messages
174
Likes
187
:) Im making no claims about my dac, it is an ESS9038 implementation so very representative as to whats currently out there from other manufacturers.

...and yes people are very much struggling to identify which is which, I have seen the results. I think it highly unlikely that my dac is fixing any issues of the ADC or the original recordings. Thats really not plausible proposition that it provides an inverse to those issues.
In any case surely it will still have the DAC filter issues you allude to? Plus you have already stated that the differences are readily visible in an audio editor.
This was about the particular digital filter selected in ES9038 (looks like the "Hybrid" one) which cleans to some extent aliasing region (about 20 - 22.05 kHz) of the half-band decimation filter used in the original material. I don't know if the material originated from an ADC running at 44.1 kHz or if it was down-converted during mastering for example from 96 kHz which is quite typical thing these days. Such decimation filters are quite typical in "modern" ADC chips.

So if we further band limit and filter (repeatedly) do we not have worse problems with phase?
Phase shift increases every time due to analog filters in both. Depends on design of the particular analog filters how much.

Perhaps we could ask @Blumlein 88 to repeat the test with a square wave?
I'm not sure how useful that would be. Of course from step response one can calculate impulse response and from impulse response you can calculate both frequency and phase response. Nth generation loop results are mostly relevant to studio things. Nowadays it is quite common to do mixing with analog desks in old way, tape machines are just replaced with ADCs and DACs. In such case you get multi-generation effects, but also the desk in question is one of the dominating factors.

I think the point I would make is that I am most certainly not denigrating your work, however I think there are bigger fish to fry - it doesnt address the issue anywhere else in the chain beyond the DAC, which means it can only have very limited impact, but also that we need to establish correlation to audibility to say if there is actually any impact.
What issue are you talking about now? For me, everything matters, in all parts of audio chain. I just don't want to bring all aspects into this discussion topic because it becomes too messy. We are already going offtopic... :)
 

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
5,764
Likes
7,383
Location
Albany Western Australia
What issue are you talking about now? For me, everything matters, in all parts of audio chain. I just don't want to bring all aspects into this discussion topic because it becomes too messy. We are already going offtopic... :)

Im talking about the points you have not addressed but I have mentioned several times.

You are extremely concerned about the DAC filter phase response, yet oblivious to the recording chain which has no respect for phase integrity. Equally, the subsequent amp and speaker.

Everything might matter matter to you, but you have to put things into context and importance. Can you hear it? If you cant then its not important and doesnt matter. You can clean up the DAC filter and its phase response as much as you like, but if the signal up to that point and beyond that point is completely messed up in that respect, then its not important and it doesnt matter.

But yes the topic is DSD better than PCM. In my view no.
 

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,643
Likes
12,874
How much aliasing does it have for example with full level 22.1 kHz or 23 kHz input?
I ran it at 48 khz sampling. With 25 khz full level input an alias at 23 khz is - 17 dbFS. By 28 khz input the alias at 20 khz is - 84 dbFS, and at 19,500 and lower on the alias signals they all lie -88 to -90 dbFS all the way to 80 khz inputs at max level.
 

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
5,764
Likes
7,383
Location
Albany Western Australia
This was about the particular digital filter selected in ES9038 (looks like the "Hybrid" one) which cleans to some extent aliasing region (about 20 - 22.05 kHz) of the half-band decimation filter used in the original material. I don't know if the material originated from an ADC running at 44.1 kHz or if it was down-converted during mastering for example from 96 kHz which is quite typical thing these days. Such decimation filters are quite typical in "modern" ADC chips.
Cleans? is it dirty? At what level is this dirt? Can you hear 20-22 kHz?
 

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
5,764
Likes
7,383
Location
Albany Western Australia
I ran it at 48 khz sampling. With 25 khz full level input an alias at 23 khz is - 17 dbFS. By 28 khz input the alias at 20 khz is - 84 dbFS, and at 19,500 and lower on the alias signals they all lie -88 to -90 dbFS all the way to 80 khz inputs at max level.
Thats interesting. So what is your guess of a typical music signal level in the transition band? Maybe -60dB? What would a -60dB instead of full scale signal result in?
 
Last edited:

Miska

Active Member
Joined
Feb 20, 2019
Messages
174
Likes
187
I ran it at 48 khz sampling. With 25 khz full level input an alias at 23 khz is - 17 dbFS. By 28 khz input the alias at 20 khz is - 84 dbFS, and at 19,500 and lower on the alias signals they all lie -88 to -90 dbFS all the way to 80 khz inputs at max level.
So it is one of those "modern" ones with quite strong aliasing band at the top. Luckily that can be fixed at playback time with suitable choice of oversampling filter.

Cleans? is it dirty? At what level is this dirt? Can you hear 20-22 kHz?
Yes, as usual lot of content is dirty, especially inside transients. Level of the dirt depends on the particular content, but can be just -20 dB or so down.

Thats interesting. So what is your guess of a typical music signal level close to the transition band?
Close-miked drums and percussions can reach it pretty much flat. Claves are good example:
https://www.cco.caltech.edu/~boyk/spectra/11.htm#a

Note to readers: the challenge looking at spectra is that since these frequencies are not continuous but transient, you need to take into account that the edge has much higher concentration of HF content than is apparent from the spectra, because it's length is shorter than length of the FFT which means it's contribution to total power spectrum is less even if it's short term contribution would be higher.

Looking at my own tests, for example for castanets the level difference at 20 kHz is -20 dB compared to highest level. For soprano glockenspiel, highest level tone above 22.05k is at 22.8 kHz and it is -20 dB too compared other high peaks which are at 3.4, 7.4, 9.3, 13.8 and 16.1 kHz. 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...
 

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
5,764
Likes
7,383
Location
Albany Western Australia
Yes, as usual lot of content is dirty, especially inside transients. Level of the dirt depends on the particular content, but can be just -20 dB or so down.



Close-miked drums and percussions can reach it pretty much flat. Claves are good example:
https://www.cco.caltech.edu/~boyk/spectra/11.htm#a

Note to readers: the challenge looking at spectra is that since these frequencies are not continuous but transient, you need to take into account that the edge has much higher concentration of HF content than is apparent from the spectra, because it's length is shorter than length of the FFT which means it's contribution to total power spectrum is less even if it's short term contribution would be higher.

Looking at my own tests, for example for castanets the level difference at 20 kHz is -20 dB compared to highest level. For soprano glockenspiel, highest level tone above 22.05k is at 22.8 kHz and it is -20 dB too compared other high peaks which are at 3.4, 7.4, 9.3, 13.8 and 16.1 kHz. 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...
I would dispute that.

Where would you find claves or castenets mixed to 0dB in a musical recording? You wouldnt.
 

Zek

Active Member
Joined
Feb 17, 2018
Messages
296
Likes
134
Hi Miska,
I read the HQPlayer Manual and it's very complicated for us oldman novices.
How to set up HQPlayer so that any audio file (PCM16/44 - 24/192 and DSD64-128) is converted and played back in DSD256 (or DSD512) audio format?
Please for a detailed step-by-step guide.
Thank you very much in advance.

P.S. In what format should the album image be uploaded by HQPlayer in its window?
 

Miska

Active Member
Joined
Feb 20, 2019
Messages
174
Likes
187
I would dispute that.

Where would you find claves or castenets mixed to 0dB in a musical recording? You wouldnt.
Nothing should be ever mixed to 0 dB, but there are periods in recordings (some Spanish music (Flamenco) and such) where castanets play solo. What matters are the relative levels. And what stops me from publishing my recordings of castanets playing? I would be upset if someone considers my musical creation unimportant to reproduce properly.

You can also find similar effects in periods where someone drives ADC into clipping. That shouldn't be done either, but it still happens. And recordings shouldn't contain digital clipping, but most RedBook content does, thanks to loudness wars.

I never ever make assumptions about what "musical recording" contains. There are so many different types of recordings in the world. From trash metal to Taiko drummers and sounds of nature. Roger Waters' Amused to Death album contains a very nice scene of fighter plane flying towards you and over you and dropping a bomb that explodes. Regarding earlier comment about MQA, that is one of the mistakes they make. In their presentations they present typical spectra of classical music, like that would be all there is.

So playback gear better behave properly in all conditions. That is important to me, other people may have different priorities.
 
Last edited:

Miska

Active Member
Joined
Feb 20, 2019
Messages
174
Likes
187
Hi Miska,
I read the HQPlayer Manual and it's very complicated for us oldman novices.
How to set up HQPlayer so that any audio file (PCM16/44 - 24/192 and DSD64-128) is converted and played back in DSD256 (or DSD512) audio format?
Please for a detailed step-by-step guide.
Thank you very much in advance.

P.S. In what format should the album image be uploaded by HQPlayer in its window?
In the "DSDIFF/DSF Settings" uncheck the "Direct SDM" box to allow DSD-to-DSD processing. Then in main window, select "SDM (DSD)" as output format. In settings, set for example "44.1k x256" or "44.1k x512" as rate limit and make sure "Auto rate family" in unchecked.

Cover art embedded in files is shown automatically. External cover art is shown when playing from the library, in which case JPEG or PNG format files with "cover" or "folder" in the file name are picked up.

For further questions about my software, please send me email, start a new thread here, or post at the HQPlayer thread at AS: https://audiophilestyle.com/forums/topic/19715-hq-player/
Just to avoid derailing this thread too much from the original topic. Thanks! :)
 

March Audio

Major Contributor
Manufacturer
Joined
Mar 1, 2016
Messages
5,764
Likes
7,383
Location
Albany Western Australia
Nothing should be ever mixed to 0 dB, but there are periods in recordings (some Spanish music (Flamenco) and such) where castanets play solo. What matters are the relative levels. And what stops me from publishing my recordings of castanets playing? I would be upset if someone considers my musical creation unimportant to reproduce properly.

You can also find similar effects in periods where someone drives ADC into clipping. That shouldn't be done either, but it still happens. And recordings shouldn't contain digital clipping, but most RedBook content does, thanks to loudness wars.

I never ever make assumptions about what "musical recording" contains. There are so many different types of recordings in the world. From trash metal to Taiko drummers and sounds of nature. Roger Waters' Amused to Death album contains a very nice scene of fighter plane flying towards you and over you and dropping a bomb that explodes. Regarding earlier comment about MQA, that is one of the mistakes they make. In their presentations they present typical spectra of classical music, like that would be all there is.

So playback gear better behave properly in all conditions. That is important to me, other people may have different priorities.
But as Blumleins earlier numbers show an alias could be 85dB down by the time you hit the audio band. So where is the problem?
 
Status
Not open for further replies.

Similar threads

Top Bottom