• 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.

AES Paper Digest: Subjective Evaluation of High Resolution Recordings in PCM and DSD

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
35,856
Likes
142,200
Location
Seattle Area
A much more recent paper/study on differences between PCM at 24/192 versus DSD at 2.8 and 5.6 Mhz (single and double speed DSD respectively). http://www.aes.org/e-lib/browse.cfm?elib=17166&rndx=224452

Subjective Evaluation of High Resolution Recordings in PCM and DSD Audio Formats

Atsushi MARUI1 , Toru KAMEKAWA1 , Kazuhiko ENDO2 , and Erisa SATO Faculty of Music, Tokyo University of the Arts, 1-25-1 Senju, Adachi, Tokyo, 120-0034, Japan TEAC Corporation, 1-47 Ochiai, Tama, Tokyo, 206-8530, Japan Correspondence should be addressed to Atsushi MARUI ([email protected])

Presented at the 136th Convention 2014 April 26–29 Berlin, Germany

The study is a paired AB comparison. It puts PCM 192 against DSD 1X and 2X and DSD 1X against 2X. What makes it odd is that the raw results are not given, only the "p" factor or probability of chance. The test is one of preference, not whether the listeners could tell them apart to start.

The testing used speakers and two clock synced TASCAM players:

"Two TASCAM DA-3000 (from the same production lot with the same firmware version installed) were used for playback of all the stimuli. They were set to master- and slave-mode for playback synchronization. Hence, the same digital-to-analog converter was used for all stimuli played back. Outputs from DA-3000 were sent to a remote controllable monitor switcher (operates in analog domain) which enabled a listener to switch between one of the two playback sources.

Two loudspeakers were positioned in the standard stereo playback according to ITU-R BS.775 [6], with 2.70 m (≈ 8.86 feet) from the listening position (Figure 2). Two Genelec 1032A were used at Site A and Genelec 8050A were used at Site B. A stereo volume controller was installed as a precaution for loud noise exposure to human subjects. Because no loud noise was emitted by accident, the level was kept at constant level throughout the experiment. Esoteric C-02 preamplifier was used at Site A and Tomoca TCC-100ST was used at Site B for the volume controller."


ITU rooms are supposed to represent the typical living room we have. In reality it doesn't do that but is a standard configuration making it easier to compare one study to another performed in a similar room.

Here are the test results:

upload_2016-3-16_18-20-21.png


Can't understand this? It is OK. I don't think most people can :). Essentially it says the smaller the value the more the listeners prefered the format on the left to the one on the right in that category of effect. First time I see such a comparison table so not sure how valid it is.

Here is their conclusion:

"The three formats were compared by 46 participants on six sound programs and eight attributes. From the result of binomial test applied on the data from pairwise comparison experiment, statistically significant differences between PCM and DSD but not between the two sampling frequencies (2.8 MHz and 5.6 MHz) of DSD.

[...] formats, stimuli having broad spectra and clear temporal transients (such as Vocal, Jazz Trio, and Piano) and attributes such as spatial width, spatial depth, timbral richness were able to be used to discriminate between DSD and PCM. Overall quality and preference showed similar tendency of in favor of DSD (5.6 MHz) over PCM (192 kHz/24 bit)."
 
Last edited:

Opus111

Addicted to Fun and Learning
Joined
Mar 2, 2016
Messages
666
Likes
27
Location
Zhejiang
tascamda3000.png


So not really a comparison with PCM at all - the PCM1795 being an S-D DAC.
 

Mivera

Major Contributor
Joined
Mar 4, 2016
Messages
2,322
Likes
83
Location
West Kelowna

Mivera

Major Contributor
Joined
Mar 4, 2016
Messages
2,322
Likes
83
Location
West Kelowna
Don't say I have never done anything for you Mike. :D

Ha ha, and that's just double DSD with a mediocre ADC. Imagine quad with a ADC based on dual AKM 5578's in dual mono mode with Sonny's Jfet folded cascade discrete gain stage input buffer! Man I need to talk to Sonny! :)
 

dallasjustice

Major Contributor
Joined
Feb 28, 2016
Messages
1,270
Likes
848
Location
Dallas, Texas
The Tascam DA-3000 will always sound better with DSD. I'd bet PCM192 converted to DSD128 would sound equivalent to the native DSD128.

Of course TEAC sponsored the study and their ADCs and DACs are all SDM designs. So it's unlikely they would want to do as I suggested.
 

Mivera

Major Contributor
Joined
Mar 4, 2016
Messages
2,322
Likes
83
Location
West Kelowna
The Tascam DA-3000 will always sound better with DSD. I'd bet PCM192 converted to DSD128 would sound equivalent to the native DSD128.

Of course TEAC sponsored the study and their ADCs and DACs are all SDM designs. So it's unlikely they would want to do as I suggested.

Your Mola Mola handles both PCM and DSD the best way anyways by upsampling it all to 1 bit. So you get the best of both formats either way.
 

dallasjustice

Major Contributor
Joined
Feb 28, 2016
Messages
1,270
Likes
848
Location
Dallas, Texas
Your Mola Mola handles both PCM and DSD the best way anyways by upsampling it all to 1 bit. So you get the best of both formats either way.
I agree that DACs should be format agnostic.

The only thing which could be concluded from this TEAC study is that the DA-3000 sounds better with DSD sources; nothing else.

It would be more interesting to compare different DACs with various formats.

For me 1 bit DSD has always been a dead letter due to its incompatibility with DSP.
 

Mivera

Major Contributor
Joined
Mar 4, 2016
Messages
2,322
Likes
83
Location
West Kelowna
I agree that DACs should be format agnostic.

The only thing which could be concluded from this TEAC study is that the DA-3000 sounds better with DSD sources; nothing else.

It would be more interesting to compare different DACs with various formats.

For me 1 bit DSD has always been a dead letter due to its incompatibility with DSP.

My testing has shown the main benefit with DSD is how it's handled in the DAC anyways. So your Mola Mola is perfect for what you're doing because it allows you to take advantage of both DSP, and upsampling to 1 bit. Even if you tried something like HQplayer with that DAC, you would likely see little or no benefit.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
13,413
Likes
20,894
This test would indicate a difference. It does not indicate the DSD was more accurate or of higher fidelity however. I am reminded of comparisons I did with triode and wideband SS amps. In such a test, the triode would have cleaned up at 100% every category listed above. Then I could run the output of the triode amp into the SS amp. It was the same sound. So the SS wasn't incapable of that sound. It simply wasn't adding the euphony to the signal.

Nevertheless, a genuine perceptible difference is worthy news. The one we need added is converting 192 PCM to double DSD then comparing to the original 192.

I also would like to know more about the details of the presentation. Did they play PCM first in every instance for example? One would think they knew enough not to do this.
 
Last edited:

AK1

New Member
Joined
Aug 6, 2021
Messages
1
Likes
0
Hello!

My first post here.

I wonder about the following issues, which might have skewed the results:
  1. I have not seen any reference to whether exact volume matching has been performed between the PCM and the DSD samples being compared, plus/minus what fraction of a dB. IF no such matching has been performed, or not performed accurately enough, the louder version has an advantage over the softer one

  2. The DSD recorder might have done a better job than the PCM recorder -- this is possible even if the same chip has ben used for both
  3. Likewise, the DSD replay might have been better than the PCM replay, even if the same chip has been used for both
For avoiding the skews resulting from 2 and/or 3, prior sets of comparisons would have to be performed for selecting the best PCM recorder and the best PCM player, and likewise for DSD, among a few top performing candidates, and do the comparison between the winning models. I does not guarantee that the handlers are equally good, but the comparison is between the best obtainable PCM result vs. the best obtainable DSD result (in any case, the comparison is not of the formats directly, but between the results of converters of the formats to/from analog).

Also, I don't know how the conversion chip works internally, but it is possible that when DSD is being handled, the chip performs the required operation (ADC or DAC) in PCM and converts the result to DSD, or vice versa. IF that is the case, this is not a PCM vs. DSD comparison, but rather DSD-->PCM vs. PCM or PCM-->DSD vs. DSD comparison. Maybe the chip's specs specify this, or the manufacturer needs to be asked.

What do you think?
 
Top Bottom