Oh! My God!
That is a great resource !
Thank you very much for that !
Nevertheless, it must be said that DSD is a form of sigma-delta modulation, not delta modulation, which is a different thing.
-
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!
The sound quality of DSD?
- Thread starter Saidera
- Start date
That is a great resource !
Thank you very much for that !
Nevertheless, it must be said that DSD is a form of sigma-delta modulation, not delta modulation, which is a different thing.
OP
Saidera
Senior Member
- Thread Starter
- #782
Thank you all for the insightful discussion on "The Sound Quality of DSD" and the other DSD audio-related threads I started. Your expertise and willingness to share make this community an invaluable resource, especially for executives and decision-makers still archiving recordings with DSD or considering whether to discontinue its use or sell off high-end DSD equipment.
Every now and then, a 'DSD Defender' emerges, and I find it refreshing to revisit some basic premises and concepts in plain language. Much like those who still enjoy physical mediums like vinyl, cassette, or MD, I continue to gravitate towards converting to DSD, exploring DSD recording, and seeking DSD audio support in dongles—despite the facts. Knowledge of DSD is almost like a prerequisite for understanding the nuances of high-resolution audio formats and making informed decisions about archival and recording technologies.
Looking forward to learning more from everyone here!
Every now and then, a 'DSD Defender' emerges, and I find it refreshing to revisit some basic premises and concepts in plain language. Much like those who still enjoy physical mediums like vinyl, cassette, or MD, I continue to gravitate towards converting to DSD, exploring DSD recording, and seeking DSD audio support in dongles—despite the facts. Knowledge of DSD is almost like a prerequisite for understanding the nuances of high-resolution audio formats and making informed decisions about archival and recording technologies.
Looking forward to learning more from everyone here!
@fatoldgit @pkane
It is my understanding that DSD and SACD are two completely independent projects within the conglomerate of Sony.
I have no inside info other than being in my formative audiophile days when SACD was launched so this is me repeating what I have heard over the years.
The original DSD project was to archive analog tapes digitally. Somehow Sony determined that recording the 1-bit data straight of a sigma delta ADC and then playing it back was better than the PCM ADC/DAC of the time.
Based upon claims that it was a Crystal DAC, it might be the CS5390. If that sets the timeline as 1993, the CS DAC of that time could be the CS4303.
It is possible that recording the 1-bit playback and then using a pure analog HF filter gave better results than the standard 20/44.1 or 20/48 PCM encoding at the time. We would have to find some 90s era products to measure to see if that was true, but maybe the avoiding of the digital filter at 44.1 kHz made sense for superiority of DSD *and* the lack of knowledge of 44.1 or 48 kHz would be the standard.
SEPARATELY, the music and CD division saw the opportunity to do patents, multichannel, leverage DVD manufacturing, etc. That separate group pulled from the Sony IP war chest to make SACD. That was entirely royalty and consumer marketing driven.
It is my understanding that DSD and SACD are two completely independent projects within the conglomerate of Sony.
I have no inside info other than being in my formative audiophile days when SACD was launched so this is me repeating what I have heard over the years.
The original DSD project was to archive analog tapes digitally. Somehow Sony determined that recording the 1-bit data straight of a sigma delta ADC and then playing it back was better than the PCM ADC/DAC of the time.
Based upon claims that it was a Crystal DAC, it might be the CS5390. If that sets the timeline as 1993, the CS DAC of that time could be the CS4303.
It is possible that recording the 1-bit playback and then using a pure analog HF filter gave better results than the standard 20/44.1 or 20/48 PCM encoding at the time. We would have to find some 90s era products to measure to see if that was true, but maybe the avoiding of the digital filter at 44.1 kHz made sense for superiority of DSD *and* the lack of knowledge of 44.1 or 48 kHz would be the standard.
SEPARATELY, the music and CD division saw the opportunity to do patents, multichannel, leverage DVD manufacturing, etc. That separate group pulled from the Sony IP war chest to make SACD. That was entirely royalty and consumer marketing driven.
Yet again, you completely put aside Philips, who was the co-developer of SA-CD/DSD.
Even if it is true that Sony was the first to publicly put forward the use of DSD at the AES as soon as 1996 [1], Philips has already embraced the project at that time. You can read the testimony of Mr Andrew Demery [2], who used to work for Philips, then Sony, about the very first multichannel DSD recording he was associated with on the Internet Archive [3]. This recording was made with a Philips's prototype 8 channels DSD recorder using hard disk drives :
The very same guy has written many testimonies on stevehoffman.tv about the internal struggle inside the Dutch and Japanese corporations that dispel many myths, for instance:
- https://forums.stevehoffman.tv/thre...and-format-history.199295/page-2#post-4939355
- https://forums.stevehoffman.tv/thre...-happening-on-sacd.198103/page-4#post-4906272
- https://forums.stevehoffman.tv/threads/why-does-vinyl-analog-sound-better.84273/page-30#post-1887773
- https://forums.stevehoffman.tv/thre...hi-end-cd-players.991673/page-2#post-26331963
From this testimonies (and numerous others), we know that, at that time, Sony and Philips have chosen DSD for product differentiation and that Sony Music never really embraced SA-CD and was reluctant to adopt either the disc format or DSD to archive its analogue tapes.
All reasons that have nothing to do with some mystical sound qualities. That's not to say that there are no technical advantages to the system. There are. But there are challenges also. As with anything technical, there are pros and cons and it is up to the engineers to balance a design within its contraints to get the required results.
Today, we know perfectly well that both systems, high-res PCM or DSD, can work equally well when done right. They are just different systems having different characterics that must be well known and understood to get the best of either of them.
[1] A. Nishio, G. Ichimura, Y. Inazawa , N. Horikawa, and T. Suzuki "Direct Stream Digital Audio System" Proceedings of the AES 100th Convention, Copenhagen, Denmark, May 1996.
[2] A name that should be well known for those who have worked in the disc industry at the beginning of the deployment of SA-CD, notably for he is the co-author (with Mr Kobie Crawford), of the Sonoma DSD digital audio workstation user guide.
[3] https://web.archive.org/web/20191225060532/http://www.superaudiocenter.com/DemRep.htm
Last edited:
Waxx
Major Contributor
DSD was relevant to make easier ad/da convertors when it was invented. At that time they did not have the convertor chips of today. But today PCM convertors are on a level that it's not relevant anymore and then dsd looses on other parameters like said. PCM as format itself is fully transparent, it was the conversion to analog and back that was the limiting factor in the past, and convertors that were able to do it in high enough resolution were very expnsive mostly. Today you can buy PCM format convertors for 100€ (or even less) that beat all those systems from 30 years ago. Even 15 years ago Philips and Sony knew that already and sold that tech and patents to another company because they knew it would go nowhere as classic PCM convertors were getting better. But the music industry saw an opportunity to sell the same music an other time to their customers, but this time in DSD format...
My experience with digital listening (in ABX of course which is the basis of this site I think) but also in measurements, very far from the threshold of audible improvement is that the first players/converters (still sometimes sought by some amateurs here and there, but this is not proof, okay) were already excellent. I have quoted here many times the famous theorem and I do not see any possible improvement by increasing the sampling rate or frequencies of more than 20,000 Hz.We could still increase the dynamics, which would make the discs inaudible in a domestic environment but that possible with the Red Book CD is only rarely used, so...
Sony has use 1-bit converters (Pulse Conversion) in their CD players for a long-time, including many ES series ones.
The hybrid discs are basically two discs in one. They use different layers.
When DSD was developed (let's say the last decade of the old millennium) it was a definite improvement on the conversion then available outside the laboratory. At that time 16/48, despite its notional 16 bit operation, suffered insofar as the two least significant bits were pretty much random.
Even though a few 20+bit converters existed at that time, their bottom bits were essentially random. We didn't start seeing reproducible 20bit+ PCM until the best part of ten years later in standard units that you and I could buy.
But as intimated elsewhere, the real motivation for introducing SACD, DSD and their concomitant copy protection was the clearly visible writing on the wall that the handling of 600MB of data, (near-impossible to manage when Red Book made its appearance), was becoming routine largely thanks to CD-R.
Certainly not. After all, It would be quite unlike a company like Sony to attempt to alter the world's digital audio direction purely for commercial gain…
Uncomfortable Fact 1 —
If you require your PCM layer to reproduce at the same volume as your DSD layer, you are likely to leave one of the two a few dB shy of peak.
Uncomfortable Fact 2 —
If you are in a world where your paymasters insist that all released audio tracks peak as high as possible (i.e. hit the ‘0’ line), you have a problem. Go to Uncomfortable Fact 1.
All this is not helped by slightly uncomfortable fact 3, that as the maximum modulation is defined differently for PCM and DSD, the two may well hit peak modulation at different points in the music. This is another facet of the loudness war that does not help anybody. If A&R execs. were not so obsessed with maximum loudness, it would be possible to work around all these issues and still have more dynamic range than human hearing finds comfortable.
Last edited:
Similar threads
