So thanks for hanging in there for the last 5 days. I felt it was important to clear away some of the smoke and misinformation and present accurate information from a very knowledgeable analog and digital audio product designer. I couldn’t refute the misrepresentations of Richard from BitPerfect without John Siau’s assistance. Some comments have positioned this conversation as a debate on a philosophical subject. My view is that Richard called me out and accused me of contradictions, and playing fast and loose with the “facts” with regards to the simple statement that 96 kHz/24-bit PCM provides more audio accurate recording and reproduction than a DSD 64 recording of the same source. The last five days have unambiguously established that high-res PCM trumps DSD.
This does not close the issue of DSD and 1-bit encoding in general. Plenty of audiophiles, download sites, equipment manufacturers, and audio publications will remain faithful believers and won’t even bother to consider the information presented on this site and others. They’ll stick to the “DSD is warmer and sounds more like analog” theme and keep dreaming that DSD is going to take over the world of audio production. It’s not going to happen. And it shouldn’t happen.
So here are the last few exchanges…
BitPerfect: “In designing DSD Master, we make those design compromises on the basis that the purpose of these conversions is to be used for final listening purposes. But if a similar functionality is being designed for the internal conversion stage of a PCM SDM-ADC then we know that a residual ultrasonic noise peak in the output data is not going to be acceptable. In our view, this means that design choices will be made which do not necessary coincide with the best possible sound quality.
As a final point, all the above observations are specific to ‘regular’ DSD (aka ‘DSD64’). The problem with ultrasonic noise pretty much goes away with DSD128 and above, something I have also written about in detail in a previous post.”
John Siau I agree that the ultrasonic noise problem is less of an issue with DSD128, but it still needs to be filtered out. The low pass filter can be moved up in frequency by a factor of 2. The complication is that DSD128 and DSD64 need different analog lowpass filters. This means that a disk player or DAC needs to have switchable analog lowpass filters in order to fully benefit from DSD128. Given the complexity of these DSD post filters, I doubt that there will be many playback devices that offer switchable filters.
Dr. AIX I thoroughly enjoy learning new things. And I had no idea the higher rate DSD encoding requires independent analog lowpass filters to get rid of the ultrasonic noise. And there would be yet another filter for the DSD 256 files. Thanks for that one John.
As we read the last paragraphs from the BitPerfect article, Richard acknowledges the BitPerfect’s DSD Master makes “design compromises” in order to craft a viable product for listening to PCM and DSD music files. They allow the ultrasonic noise…the uncorrelated, non-harmonic noise…to remain along with any ultrasonic music components (and yes, there are money up to and beyond 40 kHz) until a 30 kHz corner frequency lowpass filter kicks in.
While those using PCM to record and deliver high-end audio can provide the exact master file from the studio to the consumer without any conversions (and the loss of information associated with them), at a lower price, in a smaller file, and with extended fidelity. What’s to debate?
BitPerfect: “So, from the foregoing, purely from a logical point of view, it seems somewhat contradictory for Dr. AIX to suggest that 24/96 PCM is inherently better than DSD, since DSD comes directly out of a SDM in its native form, whereas PCM is derived through digital manipulation of an SDM output with, among other things, a ‘brick-wall’ filter with a less-than-optimal configuration. I’ll also point out that his argument suggests that DSD (i.e. the output of an SDM) will not deliver the full bit depth that he offers up as a key distinguishing feature of 24/96. Of course, those arguments apply only to ‘purist’ recordings, which seek to capture the microphone output as naturally as possible. In that way the discussion is not coloured by any post-processing of the signal, which in any case is not possible in the native DSD domain.”
John Siau Wrong conclusion for the reasons outlined above.
These days, PCM originates in a multi-bit SDM. In contrast, DSD originates in a single-bit SDM. Every added bit improves the raw SNR by 6 dB. Single-bit SDM was abandoned by most converter manufacturers at least 10 to 20 years ago.
Benchmark has been using multi-bit SDM for 20 years (since 1995). In 1995 we combined two 1-bit sigma-delta modulators to create a multi-bit sigma-delta modulator. We did this to reduce the ultrasonic noise produced by the sigma-delta D/A converter. Our current D/A converter (the DAC2) uses four 5-bit sigma-delta modulators running in parallel. This massively parallel architecture offers significant advantages over single-bit SDM.
Dr. AIX Richard’s last paragraph states that “logically” my argument and that of John Siau (and Robert Stuart of MQA and Meridian and others) about DSD vs. PCM is “somewhat contradictory”. He went through a long explanation of why DSD 64 is inherently better than PCM, “since DSD comes directly out of a SDM in its native form, whereas PCM is derived through digital manipulation of an SDM output with, among other things, a ‘brick-wall’ filter with a less-than-optimal configuration.”
Somehow DSD is more native and natural than digitally manipulating the SDM output from a convertor. However, consider that the “digital manipulation” described by John above (5-bit SDM parallel modulators) produces better phase response, better frequency response, and better dynamic range than DSD 64. Perhaps being more natural is not an advantage in the DSD vs. PCM issue?
And by the way, 96 kHz/24-bits doesn’t require a “brick-wall” filter…just the facts here, please.
Finally, Richard challenges my argument about the superiority of high-resolution PCM over DSD because, “those arguments apply only to ‘purist’ recordings which seek to capture the microphone output as naturally as possible. In that way the discussion is not coloured by any post-processing of the signal, which in any case is not possible in the native DSD domain.”
Read this carefully. My take is that Richard is agreeing that High-Res PCM delivers “better” dynamics when used in “purist” recordings like the ones that I produce. Don’t all recording engineers want to “capture the microphone output as naturally as possible?” Apparently, he’s counting on the myriad of post production digital signal processors to diminish the “pure” source fidelity as EQ, reverb, audio compression, and other plug ins are used.
His last line admits that native DSD can’t do any post-processing, which makes it very difficult to use for most real music productions AND guarantees that most DSD releases are conversions from something else. I’m done here.
See you tomorrow. Thanks for your patience in reading through these past five days of posts. I believe this is important. What’s at issue is not whether you want to listen to PCM or DSD files, but rather the lengths that people will go to skew the facts in their favor. We can do better.