A Final…But Not Last Word On DSD.
In researching the posts of the last few days, I done a lot of reading and exchanged a bunch of emails with people who’s opinions I respect in the area of audio engineering, equipment design, and product manufacturing. Today, I’d like to share a few things that came my way in the form of emails or comments. I’m not going to reveal the authors of these statements because they were sent to me privately…I might even paraphrase here and there. But I believe the audiophile segment of the population needs to carefully consider what the get when they opt for DSD recordings.
I got the following from a customer of PS Audio, the company that recently offered upgrades to the owners of the Perfect Wave DAC. Their “upgrade” to the Perfect Stream DA forces every source recording to be converted to DSD prior to playback. I’ve written about this rather harsh design consideration. Click here to read about the new Direct Wave.
I got this note from a PS Audio customer that purchased the upgrade.
“I just recently updated my PS Audio Perfect Wave DAC to the Direct Stream. I ran it straight into the amp using the onboard volume control.
Now since I have ‘updated’ the DAC to the Direct Stream, I feel I have lost my way. Despite having been run in over 300 hours I feel the presentation has been dulled or truncated. It feels as if it has lost speed and clarity, like it needs to blow its nose. I have tried putting a preamp in the chain as well as a passive attenuator and still I am want for clarity.
I could reinstall the Perfect Wave components and sell the Direct Stream kit but wanted to get a recommendation from you on what DACs I should take a look at should I decide to break camp with PS Audio.”
I haven’t responded to this individual yet but found his “needs to blow its nose” comment compelling (I admit it brought a wide smile to my face). His comments are interesting. I don’t know if they are typical because I haven’t reached out to any other PS Audio customers that purchased the “upgrade”. I think PS Audio made a critical error by forcing everyone to adopt DSD in the Direct Stream unit. Why not augment the unit with that “feature” instead of mandating it? I’m always in favor of providing choices to my customers.
Another comment came from a long time reader. He was compelled to download a track from Channel Classics in native DSD to hear just what pure DSD can sound like. He wrote:
“Just to check things out, I downloaded an album from Channel Classics in DSD 64. The recording was captured straight from the analog mixing board in DSD. No PCM processing or I wouldn’t have bothered. MY DAC does not do DSD but Audirvanna converts it to 24/176.4 for my DAC to process. Yes, it is a warm sound but lacking in HF detail and air. I compared it to a 24/192 captured recording of similar material, which has more spatial detail and a less rolled off sound, which to these ears sounds more lifelike. I have been an audiophile kind of guy since the 80’s and I can see why some audiophiles love DSD as being older guys, their setups are too bright and the DSD sound fits right in to their systems. To me, the good PCM stuff sounds more like the live sound I experience at venues, especially small classical and jazz venues. On this one trial there is no doubt that the DSD recording (of very fine music I might add) sounds very pleasing, just not as accurate as PCM. The good news out of all of this is at least, with computer-based music, we have choices to pick the particular flavor that we enjoy. Some like warm sound. I want as accurate to the live event that I can get. But at least we have choices. Keep up the good work Mark.”
And finally, I received this from an EE guy that designs and builds world-class converters:
“Thanks for posting this often overlooked detail in the history of DSD”. He’s referring to my post from the 18th…
“By the time Sony and Philips completed the design and manufacturing of their first SA-CD machines (which were limited to stereo playback initially), the chip makers had already made advances in the design of their chips. It turned out that analog to digital converters had abandoned clocks of 64fs (64 times 44.1 kHz) and moved to 128fs. And they determined that using only a single bit wasn’t the best technology either. The state-of-the-art moved beyond 1-bit (from 1.5 to 5 bits). The imagined simplicity of the original Sony plan failed…to stay current with the new chips they needed additional processing, which was not part of the new player designs. Sony and Philips were forced to live with the older design based on the capacity of the discs (which held the same amount of data as a standard DVD…4.7 GB). The original SA-CD players and discs used DSD 64 and analog filters…and they were produced from analog and PCM masters.”
Tomorrow I’ll be writing about a study that was done to test whether subjects can perceive a sonic difference between DSD 64 and PCM at 176.4/24-bits. Check back then.
13 thoughts on “A Final…But Not Last Word On DSD.”
I see that I am not the only one who characterizes DSD/SACD as having a warm, smooth sound that is easy to listen to but lacks life in many instances. I will say as a super-fan that the Rolling Stones SACDs’ sound fabulous, master tape city, no question, having heard their stuff in every format and under a rather wide variety of conditions for decades to say the least.
My comment about high rate PCM being more realistic but not necessarily easier to listen to appear to also find agreement in other comments here; more true sonically, but also a bit more clinical. Some folks will trade off clinical accuracy for ease of listening and access to the emotional content; that’s what the ‘tubes and vinyl’ thing is all about.
Anytime anyone says “tests prove that people can’t hear the difference,” before I get sick, I always know that it can easily be the stress of the test circumstance that causes mis-perception, or lack of perception.
Acknowledging the barriers to accurate perception that do cause far too many bogus reports, the only way to let your ears work the way they want to is to listen under familiar circumstances in a normal fashion. Since hearing is part of biorythym, anything that upsets normal biorythym is likely to affect aural perceptions. This is the exact platform that allows publications such as Stereophile to exist, and while I have my hang-ups with them related to grossly over-priced and sometimes questionably engineered products, if my postulate of “normal listening for correct perception” has enabled the mag to thrive all these years and help many, many listeners, there just might be something to it.
Craig, you’ll be interested in the post I’m planning for tomorrow.
Hi Dr. W,
First things first: Massive thanks for fighting “the good fight” to differentiate science fiction from engineering facts. For some reason, the Audio world seems to be a place where the traveling medicine show hucksters have found safe haven in the 21st century. Bravo Dr. AIX! And thanks for asking us to use our ears and listen for the realities of tape hiss, vinyl surface noise, limited channel separation, encoding artifacts and sad, poor recordings. Trust our ears indeed!
Now the great news: I’ve reconnected with audio after three decades, initially intrigued by the Pono marketing splash, and then drawn back into a relationship with recorded music again. As an EE and audio enthusiast, I did a “deep dive” into what had happended since 1985, with keen interest in the improvements in digital and the essential source material. Even on a modest budget, with more of my dollars (and sense) devoted to the source material, I’m able to have a playback experience that would have been impossible in 1985.
Knowing from my training that the biggest improvements would come from starting with pristine source material, and then making sure the delivery chain did its best, I put emphasis on native playback/DAC quality (Fiio X5) driving a preamp free (one less link in the chain) direct/passive input amplifier (Acurus DIA) to the best speakers that suited my ears and budget (KEF Q and Sub). Then using tools only recently available (Onyx Apps/Faber Acoustical, etc.) to time-align, properly place and acoustically treat my listening space. Bye-bye turntable, ruby cantilever stylus, rack of components, moving coil head amp, 2nd Gen CD player, 66% interconnects, strapped Mono amps, power conditioner, and speaker overkill. Less is more.
And most certainly, budget for HD source (the real stuff not Pono ponzi) and revisiting my CD collection to rip only the best AE masterpieces. This process has provided the vital ear-fuel that has re-ignited my passion for recordings. I am thrilling to re-discovery of old treasures, re-mastered classics and a whole new world of HD digital masters.
I come full circle to another thanks — Thank you for helping me clarify, de-mystify and have this terrific adventure!
With most sincere appreciation,
Chris, it sounds like you’ve dialed in to the essence of great music reproduction. From source to delivery the simplest path is usually the best.
Sorry… I meant for my last post to be a “general” comment, not related to the topical thread. Moderate as you see fit. CK
My own experience with DSD, from files I’ve subsequently converted to PCM for use with my PCM only systems is that they come across as somewhat monochromatic, ie lack tonal colours, and yes somewhat smooth sounding as well. I’m surprised Craig finds PCM clinical; I find PCM to be nearest thing to live music I’ve ever heard. I’ve been into Hi-Fi for many years and believe that both my headphone and speaker systems are the nearest I’ve ever been to experiencing an accurate musical event.
Yesterday evening I was listening to the Steely Dan SACD “Gacho.” I was watching the VU meters on my Yamaha CA-810 amp. Between cuts, after a song ended, I happened to notice that the amp’s VU meters did not go to zero. Instead, they hovered at about the 200 mw level (with the amp set at normal listening levels.)
Concerned that the amp could be oscillating, I pressed the stop button on the SACD player and the meters immediately dropped to zero. I restarted the SACD and watched the meters carefully between cuts of a different song set on the disc. Once again the meters dropped on the song fade, but stopped falling at about the 200 mw level. Then a fraction of a second before the next song started they fell to zero.
The Yamaha CA-810’s frequency response is specified to 100 KHz (-1.5 db.) We know that the flaw of SACD DSD encoding process is that the noise and other encoding junk is pushed up above 20 KHz. My guess is that I am seeing this being reflected on the VU meters.
I have tried five different SACD discs. Each had the same noise on it between cuts (about 200 MW at normal listening levels.)
I have tried a number of CDs in the same player (Denon DVD-2910) and there was no measurable noise between cuts.
The Denon SACD player has its ultrasonic SACD filter set to 50 KHz. (The other choice is to set the filter to 100 KHz.) The filter doesn’t eliminate the noise.
I fed a DSF file from my Mac Mini through the Shiit Loki DSD DAC to the Yamaha and, no surprise, the HF noise was present.
In theory, the ultrasonic filters in the SACD player are supposed to take care of the HF junk. In addition, the CA-810 has a high frequency filter that drops to -12 db at 20 KHz.
Interestingly, while the SACD player’s ultrasonic filter did nothing, the Yamaha amp’s High Filter was very effective in removing the noise from both the SACD and DSD files. And, to my ears, there was little to no significant difference with the High Filter in or out. (Yes, I’m getting old!)
This suggests that I can keep the High Filter on when playing SACDs or DSF files with minimal loss of information to my old ears. In fact, I can compensate somewhat for the lost highs with a boost of the amp’s treble control which, happily, is after the 10 KHz filter in the amp’s circuit.
But now to the key question. Is this HF junk hurting anything?
Of course it can be argued that any HF crap hurts the audio quality. It robs power and, at best, adds nothing. But, over time, will it harm either the amp or the speakers? My guess is that the amp won’t be harmed. It’s designed to handle frequencies up to 100 KHz and the junk is simply perceived and handled as just another audio frequency. However, this may be a bigger deal to the tweeters as they may try to reproduce the noise. My concern is that this will cause them to overheat.
I started out as a huge SACD/DSD fan. The perceived difference in audio quality on SACD discs (more likely due to careful mastering than the format itself) was noticeable and remarkable. Now, I’m not so sure. I have reached the point where a well mastered Redbook CD or a well mastered uncompressed file has enough audio quality going for it.
This is worthy of an entire post. What you’re experiencing is the ultrasonic noise that is inherent in the DSD format. As the frequencies of the SA-CD begin at 22-23 kHz, a filter at 50 kHz doesn’t help much at all. Whether we can actually hear the stuff or not isn’t all we need to consider…the electronics and speakers in your playback system CAN hear it and will try to work with frequencies that can only harm the sound and possibly the equipment. DSD at any rate is a sales and marketing phenomenon and nothing more. Read today’s post.
Thank you very much for your series of posts about DSD. I personally much enjoy SACDs, as I find that they bring much warmer sound compared with CDs of the same material, and I own many of them. However, in the instances when I have been capable of comparing DSD and high-res PCM versions of modern recordings, I have always found PCM to sound more natural, closer to my experience when I am in a concert hall, while DSD is smoother. I can understand that some can prefer DSD as there is something “euphonic” to it, but to my ears it is slightly less accurate than high-res PCM.
If someone tried to engineer a warm-sounding recording via CD, so you think they would fail? I doubt it.
Of course, any sound can be put on a compact disc.
Mark, your converter expert said, “The original SA-CD players and discs used DSD 64 and analog filters…and they were produced from analog and PCM masters.”
I worked for the converter maker dCS during the time DSD was being developed and launched. I asked one of the senior Sony engineers I know who was involved in the DSD project about the early SACD release and he confirmed:
“All TELARC recordings (except for reissue titles), all DMP recordings, San Francisco Symphony’s Mahler cycles, Yo-Yo Ma recordings, and following titles were native DSD made using Sony’s Blackline recorder, or Philips’ multi-channel recorder.
Regarding A/D converter makers moving from 64Fs to 128Fs, the dCS 904 A/D converter was the first commercially available DSD A/D and is a 64Fs design. It was regarded as state of the art then and there are still many in use in studios around the world. Indeed your recent post about s recent study of perceptual differences between DSD and PCM was done with a dCS 904. The D/A converter in the current dCS audiophile products is a development of the 5 bit D/A used in the 904’s noise shaping loop.
Happy Holidays Robert! It’s true that the Telarc and SFS Media recordings were made using native DSD techniques, but they represent a very small fraction of the SA-CDs or DSD recrdings that were made. The vast majority (85%) were transfers and not native tracks. Plus the outputs of the players themselves were filtered using digital means rather than analog to keep the cost of the players down…and they were done using PCM.