• 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.

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
1,308
Likes
885
Location
SoCal
I've finally spent some time comparing PCM vs. PCM converted to 4xDSD in JRiver through the Khadas Board, and there are differences in sound.
 

Wombat

Major Contributor
Joined
Nov 5, 2017
Messages
4,097
Likes
2,419
Location
Australia
I've finally spent some time comparing PCM vs. PCM converted to 4xDSD in JRiver through the Khadas Board, and there are differences in sound.
Really? Please elaborate your credible method and its repeatability outside of your situation(vs sample of one). You did say there ARE differences, not I PERCEIVED differences.
 

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
1,308
Likes
885
Location
SoCal
Really? Please elaborate your credible method.
No need for sarcasm. It is sighted, I listen for details in a short section of a recording multiple times switching back and forth. Given the conversion process is different between on the DAC filters on externally I don't see why they can't sound different.
 

Wombat

Major Contributor
Joined
Nov 5, 2017
Messages
4,097
Likes
2,419
Location
Australia
No need for sarcasm. It is sighted, I listen for details in a short section of a recording multiple times switching back and forth. Given the conversion process is different between on the DAC filters on externally I don't see why they can't sound different.
Your gear is hardly definitive for such tests. Your post is not enlightening. Just another "I can hear it opinion". Nothing personal but you have been around ASR for a considerable time but still don't get its essence. I can't see why you posted.
shrug.gif
 
Last edited:

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
1,308
Likes
885
Location
SoCal
Your gear is hardly definitive for such tests. Yes it is probably the gear but you claimed it was the format.
Make an effort and read my post again. I claimed only that I hear differences between PCM and PCM converted to DSD.
 

Wombat

Major Contributor
Joined
Nov 5, 2017
Messages
4,097
Likes
2,419
Location
Australia
Make an effort and read my post again. I claimed only that I hear differences between PCM and PCM converted to DSD.
Yes, another credibly unsupported 'I can hear differences" opinion.

I am not being personal here but 'take my word for it' does not measure up to scientific scrutiny. A fact of life on ASR is that one can be challenged to be more credible re their posts.
 
Last edited:

Wombat

Major Contributor
Joined
Nov 5, 2017
Messages
4,097
Likes
2,419
Location
Australia
No need for sarcasm. It is sighted, I listen for details in a short section of a recording multiple times switching back and forth. Given the conversion process is different between on the DAC filters on externally I don't see why they can't sound different.
No sarcasm. Scientific scrutiny.

But you have been here on this forum many times before, to no avail, re taking questioning personally .
e417.png
 
Last edited:

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,177
Likes
1,163
Location
Le Mans, France
Good morning from France ! :D

I try to report my experience ...

I play WAV and DSD files by means of the AudioGate 4 software from KORG, via the DS-DAC 100 from KORG.

KORG says : Audio Gate 4 is a convenient feature that utilizes a high-speed filter and direct 1-bit DSD output. In order to experience the exceptional fidelity provided by 1-bit audio, please use the Korg DSDAC Series, a certified USB audio device, or a device that supports 1-bit audio, such as the Korg MR Series.

For the sample rate of the audio device, I have the choice between :

- leaving AudioGate 4 automatically switch the sample rate of the DAC to match the sample rate of the current song
- changing the rate on the fly, from 44.1 thru 192 kHz for the WAV files and 2.8 / 5.6 MHz for the DSD files

For the buffer size and latency, I set those at the maximum values.

Listening a 44.1 kHz WAV song with automatic sample rate : I hear as I expect it to hear. (I have made A/B comparisons with the CD versions for many of the WAV songs of mine.)
Listening the same 44.1 kHz WAV song increasing the sample rate to 192 kHz on the fly, step by step : I hear gradually more details, to the point that at 192 kHz, the listening becomes a little bit fatiguing, with the desire to go back to the native frequency of 44.1 kHz.
Listening the same 44.1 kHz WAV song switching the sample rate to 2.8 or 5.6 MHz : what I hear becomes substantially different, as if the song would come out from a different recording, I am not saying it is better than the WAV (CD) version I know, just different. I little bit of difference between the 2.8 and 5.6 MHz also. The same result applies to 192 kHz upscaled to the DSD frequencies.
Listening a DSD (2.8 or 5.6 MHz) song with automatic sample rate : I hear as I expect it to hear since I have also the SACD version in some cases.
Listening the same DSD song downscaling the sample rate to 44.1 kHz on the fly, step by step : the song becomes gradually "unlistenable" compared to the original DSD version ...

My conclusions are very simple :

WAV ---> upscaled to the PCM higher frequencies ---> to the DSD frequencies : OK, why not, if one has the desire to get a "different version" out from the same WAV file. (This is a totally different result from what one can get playing with the EQs.)
WAV <--- downscaled from DSD : no interest at all
 

Wombat

Major Contributor
Joined
Nov 5, 2017
Messages
4,097
Likes
2,419
Location
Australia
Good morning from France ! :D

I try to report my experience ...

I play WAV and DSD files by means of the AudioGate 4 software from KORG, via the DS-DAC 100 from KORG.

KORG says : Audio Gate 4 is a convenient feature that utilizes a high-speed filter and direct 1-bit DSD output. In order to experience the exceptional fidelity provided by 1-bit audio, please use the Korg DSDAC Series, a certified USB audio device, or a device that supports 1-bit audio, such as the Korg MR Series.

For the sample rate of the audio device, I have the choice between :

- leaving AudioGate 4 automatically switch the sample rate of the DAC to match the sample rate of the current song
- changing the rate on the fly, from 44.1 thru 192 kHz for the WAV files and 2.8 / 5.6 MHz for the DSD files

For the buffer size and latency, I set those at the maximum values.

Listening a 44.1 kHz WAV song with automatic sample rate : I hear as I expect it to hear. (I have made A/B comparisons with the CD versions for many of the WAV songs of mine.)
Listening the same 44.1 kHz WAV song increasing the sample rate to 192 kHz on the fly, step by step : I hear gradually more details, to the point that at 192 kHz, the listening becomes a little bit fatiguing, with the desire to go back to the native frequency of 44.1 kHz.
Listening the same 44.1 kHz WAV song switching the sample rate to 2.8 or 5.6 MHz : what I hear becomes substantially different, as if the song would come out from a different recording, I am not saying it is better than the WAV (CD) version I know, just different. I little bit of difference between the 2.8 and 5.6 MHz also. The same result applies to 192 kHz upscaled to the DSD frequencies.
Listening a DSD (2.8 or 5.6 MHz) song with automatic sample rate : I hear as I expect it to hear since I have also the SACD version in some cases.
Listening the same DSD song downscaling the sample rate to 44.1 kHz on the fly, step by step : the song becomes gradually "unlistenable" compared to the original DSD version ...

My conclusions are very simple :

WAV ---> upscaled to the PCM higher frequencies ---> to the DSD frequencies : OK, why not, if one has the desire to get a "different version" out from the same WAV file. (This is a totally different result from what one can get playing with the EQs.)
WAV <--- downscaled from DSD : no interest at all
Another sincere post asking us to believe "what I can hear".

Personal anecdotes. We need more authoritative affirmation.

I have to ask why ASR keeps getting this sort of input given that its masthead clearly explains ASR's purpose?
hide.gif
 
Last edited:

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,177
Likes
1,163
Location
Le Mans, France
Another sincere post asking us to believe "what I can hear".

Personal anecdotes. We need more authoritative affirmation.
Well, my intention was not asking you to believe to what I can hear, but merely to report an experience ... If you do not like it, tant pis ...
What our ears hear is music, not graphs, those are very helpful to compare materials, but what we hear ultimately is music.
No graph can convince whoever wants to enjoy the pleasure of getting his body fed with music whether one format is better than another, look at the 320kbps as clear explanation ... o_O
 

Sal1950

Major Contributor
The Chicago Crusher
Patreon Donor
Joined
Mar 1, 2016
Messages
5,965
Likes
3,647
Location
Central Fl
WAV ---> upscaled to the PCM higher frequencies ---> to the DSD frequencies : OK, why not, if one has the desire to get a "different version" out from the same WAV file. (This is a totally different result from what one can get playing with the EQs.)
Well, my intention was not asking you to believe to what I can hear, but merely to report an experience ... If you do not like it, tant pis ...
What our ears hear is music, not graphs, those are very helpful to compare materials, but what we hear ultimately is music.
No graph can convince whoever wants to enjoy the pleasure of getting his body fed with music whether one format is better than another, look at the 320kbps as clear explanation ... o_O
Don't know what's different about it, though graz seem to come to a proper conclusion.
Two possibilities bring you to the same answer.
Any perceived difference was induced by sighted bias.
Any real difference was a departure from what the file was intended to sound like.
Why bother, listen to a file in the manner it was recorded/released, anything else is a modification of such.
Or use REAL eq to get to a place you like.
 

Graph Feppar

Active Member
Joined
Nov 19, 2018
Messages
195
Likes
156
DSD MQA Hi Res... its all snakeoil.

Dithered and noiseshaped 16bit 44.1KHz with FLAC loseless compression is best audiophile sound format.It takes little space - is most efficient in sound quality per file size,audible noise is 120db down,is true loseless on all DACs,doesnt require any license,is most compatible sound format on earth,effects and DSP can be used without conversion to another format,no DRM,it have no quantization distortion becose it have enough bits for superior triangular dither,easy oversampling - no need for steep reconstruction filter.

Even 24/96 is BS for consumer.Dont know about you guys but I listen to 95db peak volume in room with 30db noisefloor.I probably could not be able to tell difference between high quality reel to reel tape player and modern 100db + SINAD DAC.When the music is playing,I have to concentrate like super sayan to hear anything bellow -40db.
 

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,177
Likes
1,163
Location
Le Mans, France
DSD MQA Hi Res... its all snakeoil.

Dithered and noiseshaped 16bit 44.1KHz with FLAC loseless compression is best audiophile sound format.It takes little space - is most efficient in sound quality per file size,audible noise is 120db down,is true loseless on all DACs,doesnt require any license,is most compatible sound format on earth,effects and DSP can be used without conversion to another format,no DRM,it have no quantization distortion becose it have enough bits for superior triangular dither,easy oversampling - no need for steep reconstruction filter.

Even 24/96 is BS for consumer. ...
I have the feeling I agree with your comments, the best "easy listening experience" is via 16/44.1 FLAC ... but as I have already posted in an earlier post, for "serious listening" I want the highest resolution possible and for me that is both DSD and high resolution PCM recorded SACDs. :cool:
Ohh ... "serious listening" happens only when I am home alone, no dogs, no wife, no daughter messing around with their chatting ... or thru the headphones for a more intimate experience ... ;)
 

Kal Rubinson

Major Contributor
Industry Insider
Joined
Mar 23, 2016
Messages
1,118
Likes
1,033
Location
NYC/CT
The layout and graphics of that guys site gives me a eye-ache after a few minutes. It tells me to just go somewhere else before I freak out. :(
His poor grammar makes reading it an unnecessary strain.
 

Pluto

Active Member
Patreon Donor
Joined
Sep 2, 2018
Messages
182
Likes
306
Location
Harrow, UK
DSD is usually at -6dB or so relative to the same PCM recording
It's a tricky business designing a converter that is accurately level matched between PCM and DSD modes. PCM is absolute in terms of it's headroom; when you reach the maximum, there is nowhere else to go. You hit the wall. Ouch.

DSD is rather different insofar as the coded state is rather more representative of the rate of change in the analogue waveform than its absolute
state at any instant.
Here's a nice illustration courtesy of Wikipedia.

Most DSD material is prepared to be SACD-legal and the Scarlet Book (the SACD rule book) states that the number of ones within any 28 sample 'window' must be more than 4 and fewer than 24. This is a physical limitation of the optical reading process and it relates to the use of eight-to-fourteen modulation. In short, this rule exists so that the optical reading sensor does not spend so much time in either (1 or 0) state as to cause it to loose synchronization. In this respect, the critical distinction between SACD and Red Book is that, in the latter, this rule does not affect the programme content in the slightest whereas, given the bitstream nature of SACD, it might.

So, with PCM we have an absolute, clearly defined, upper limit. With DSD the maximum modulation is, in some degree, programme dependent and, therefore, ill-defined compared to the exactness of PCM. Were this not the case, manufacturers of dual-mode converters could match the maximum modulation of PCM to generate the same analogue level as the DSD maximum. But as it is, if you design your analogue output stage to cope with the worst-case DSD scenario, you are likely to end up with a different level compared to PCM in which you know exactly what maximum level will be encountered.

Were I a cynic (which, of course, I am not o_O) I might suggest that this inherent discrepancy is highly convenient in terms of preventing the creation of a converter system capable of switching instantly between level-matched PCM and DSD.
 
Status
Not open for further replies.

Similar threads

Top Bottom