• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Octave Music Don Grusin High Resolution Music Analysis (Video)

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
Here is a PCM1792 (Mono mode) outputting -150dB signal in DSD64 mode. For DSD256 shift the noise lower and to the right. No in band problems to see here.
You cannot compare graphs with different configurations (FFT size and averaging) and different test signals (frequency and amplitude). These are common sense.
 

TCD333

Member
Joined
Nov 12, 2020
Messages
34
Likes
13
You cannot compare graphs with different configurations (FFT size and averaging) and different test signals (frequency and amplitude). These are common sense.
With -60dB the noise is basically the same. With DSD64 in Mono 9V OP mode the DR will be -130dB. With DSD256 it will reach the chips max of > -132dB. This is far in excess of 16 bit NF. So the format itself is capable of huge DR in audio band. Conversion software and hardware is another issue.
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
With -60dB the noise is basically the same. With DSD64 in Mono 9V OP mode the DR will be -130dB. With DSD256 it will reach the chips max of > -132dB. This is far in excess of 16 bit NF.
From the same datasheet:
dsd.png




pcm.png
 

TCD333

Member
Joined
Nov 12, 2020
Messages
34
Likes
13
That's correct. You are looking at DSD64 not DSD256.
You were asserting that "DSD256 is still noisier than 44.1k PCM" in band. This is just not the case.
 
Last edited:

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
That's correct. You are looking at DSD64 not DSD256
An amusing excuse. Did the datasheet mention the graph you posted is from DSD64 or DSD256?
You were asserting that "DSD256 is still noisier than 44.1k PCM" in band. This is just not the case.
Since you edited your post...
I was clearly commenting Miska's RME AKM DAC measurement results, and it is totally irrelevant to how you interpret the TI datasheet.
 
Last edited:

Dogcoop

Active Member
Joined
Feb 12, 2021
Messages
136
Likes
269
It has been explained to me that the PCM format may have problems with: glitches, idle tones and some imaging at low levels at a few hundred kilohertz or in the megahertz range. What other problems are there with PCM that I am missing? Ringing? There is incredible technological firepower on display in this thread, but no one has explained why we need to ‘improve’ the PCM format. If someone could explain in a couple of sentences what the problem with PCM is that must be addressed, I will leave this thread a happy man. If the problem with PCM cannot be identified, then there is no need to look for a solution, whether it be DSD or some as yet undiscovered new format.

Happy Listening!

PEACE
 

TCD333

Member
Joined
Nov 12, 2020
Messages
34
Likes
13
An amusing excuse. Did the datasheet mention the graph you posted is from DSD64 or DSD256?
All DSD measurements in the data sheet are at 64x. In the TI forum, application Engineers couldn't guarantee that it will run higher than DSD64.
However this DAC has successfully been used by people at DSD256 rates and the max DSD clocking spec supports this.

Maybe you could cut the condescending approach and take note that I have spoken to product application engineers and other designers
WRT what this chip can and can't do.

This forum is clearly very anti DSD amongst other things. I'm neither a proponent or a detractor but I do like to understand what it is
capable of and how to best work with it.

TCD
 

TCD333

Member
Joined
Nov 12, 2020
Messages
34
Likes
13
Then don't post that graph in the first place.
???? Surely you understand that whatever the NF is at DSD64, it will be lower at DSD256.
You clearly stated that "DSD256 is still noisier than 44.1k PCM". I'm at a bit of a loss why but maybe it
was a response to Miska's input which seems to have polarized a lot of people here.

TCD
 

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
index.php

index.php
View attachment 196694

It is amusing that everyone put the focus on ultrasonic noise without paying attention to stuff below 20kHz. Even Miska's trusted AKM DAC with his trusted modulator and his trusted Prism analyzer showed that DSD256 is about 3dB noisier than 44.1k PCM, so no more excuse about ADC aliasing and such. If the rise of <20kHz noise is really caused by ADC aliasing, then it simply means DSD256 is still noisier than 44.1k PCM, even in a "Miska approved" environment. Archimago got similar results as well:
Summary.png

Regardless of audibility, the whole DSD upsampling thing is simply a waste of time.

Good point! Here's ADI-2 Pro FS, DSD256 (blue) vs PCM192 (red). Captured by Cosmos ADC @ 192k.

By the way, both, DSD and PCM playing through HQPlayer :)
1648729693299.png


Same, but zoomed out
1648729922393.png
 
Last edited:

TCD333

Member
Joined
Nov 12, 2020
Messages
34
Likes
13
Good point! Here's ADI-2 Pro FS, DSD256 (blue) vs PCM192 (red). Captured by Cosmos ADC @ 192k.
Yes, that looks about right.

The DR spec for AKM DAC in ADI-2 is about 3dB worse in DSD mode.
The Cosmos ADC (used for measurement) is really good. One of the very few ADC's to have very flat noise floor up to 100kHz -> great for
WB measurements.

TCD
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
???? Surely you understand that whatever the NF is at DSD64, it will be lower at DSD256.
You clearly stated that "DSD256 is still noisier than 44.1k PCM". I'm at a bit of a loss why but maybe it
was a response to Miska's input which seems to have polarized a lot of people here.

TCD
Higher sample rates (and bit depth in the case of PCM) must have a potential to achieve lower noise floor in digital domain, that's for sure. So "DSD256 is still noisier than 44.1k PCM" is a comment to Miska's measurements in analog domain. If you have to interpret my previous post in the way you think, then it is really up to you.
 

TCD333

Member
Joined
Nov 12, 2020
Messages
34
Likes
13
Higher sample rates (and bit depth in the case of PCM) must have a potential to achieve lower noise floor in digital domain, that's for sure. So "DSD256 is still noisier than 44.1k PCM" is a comment to Miska's measurements in analog domain. If you have to interpret my previous post in the way you think, then it is really up to you.
I'd say the 3dB difference in analog domain is probably the RME DAC. If he was using Prism H/W I'd be looking at that. Very unlikely to be HQplayer.


TCD
 

Music1969

Major Contributor
Joined
Feb 19, 2018
Messages
4,636
Likes
2,809
Yes, the same. In fact, that's why I posted the zoomed out version to show that PCM output was at the same level at 0dBc as it was with DSD.
Is that measuring Vrms of the analogue outputs?

Input signal might be 0dBFS for both, but that's seperate to output level?

And is that ADI-2 in "DSD direct" mode?
 

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
Is that measuring Vrms of the analogue outputs?

Input signal might be 0dBFS for both, but that's seperate to output level?

And is that ADI-2 in "DSD direct" mode?

It's the output level, as measured by the ADC. The plots I posted show only the signal recorded by the ADC. How could it be a separate output level?

And yes, DSD Direct = ON
 

Rottmannash

Major Contributor
Forum Donor
Joined
Nov 11, 2020
Messages
2,969
Likes
2,606
Location
Nashville
It's the output level, as measured by the ADC. The plots I posted show only the signal recorded by the ADC. How could it be a separate output level?

And yes, DSD Direct = ON
Well, that settles that now doesn't it...
 
Top Bottom