Why Does DSD sound better than PCM?

I’ve always been curious about this. I’ve ready a variety of reasons for this but from my reading it seems the main factor is to do with filtering artefacts being moved out of the audible range, or at the very least out of the frequency range where our hearing is most sensitive.

This would also explain some high res audio files don’t sound better than their 16 bit versions - especially when they don’t use higher sampling rates.

Anyway I would be love to hear from those with expertise in the topic like @tedsmith - as well as others on what their theories are.

It doesn’t help that a lot of DSD releases were recorded as PCM, because that’s what virtually all studios use, and you cannot get directly comparable PCM and DSD files of the same recording. Blind tests have been done with mixed results. The format has not been widely accepted in the marketplace and studios are not going to change hardware and software. I bought a few DSD issues and there was nothing noticeably better about them. The streaming services have almost entirely ignored DSD. I now have a system that only operates in 24/192 PCM. It can convert DSD to PCM. So some may consider DSD better, but it is purely academic as it has not caught on in any big way.

1 Like

I’m gonna put it to you that DSD and PCM are both capable of capturing and conveying essentially perfect sound. And instead of answering “why does DSD sound better” I want to answer as if you asked “why do DSD and PCM seem to sound different?”

DSD and PCM differ in two ways: sampling frequency and encode/decode complexity.

The sampling frequency matters from a low-pass filtering perspective. Every filter degrades the audio to some degree through phase shifting, pre-ringing etc. The steeper the filter the worse the impact. Constraining CD audio to 44.1kHz sampling required the use of “brick wall” filters somewhere during recording/production. DSD, even at its lowest rate, has another six octaves of headroom so it can use a very gentle filter during A-to-D conversion and minimise those impacts. But of course PCM can be operated at higher rates too, so with care the advantage of DSD there can be caught up.

The encode/decode complexity is a physical challenge. Creating a perfect 24-bit ladder DAC requires 99.99999% precision in the resistance values of your encode/decode circuit. That’s not just me mashing the keyboard, that’s the actual order of magnitude you get when you calculate 1 - 1/(2^23). 16 bit DACs are 256 times less demanding but still, look at the effort and expense involved even in doing that in the current crop of discrete R2R DACs. Imperfections in these resistor values lead to non-linearity in the D/A conversion and that degrades the “sound” of PCM in an audible way.

Contrast that to the DS DAC which just needs two precision voltage rails (one positive, one negative) and one ultra-low-jitter clock to convert binary data into an analogue electrical signal.

For that reason, the vast majority of ADC and DAC devices which produce or consume PCM don’t actually use pure PCM techniques internally. They use a much smaller number of bits (from 1 to about 6 at most) and much higher sample rates and sigma-delta modulation with noise shaping… much like DSD.

DSD’s major downside, of course, is ultrasonic noise.

So the lines in practice are really blurry. Are you hearing the effect of brick-wall filtering in your CD audio? Are you hearing non-linearities in your R2R DAC even when decoding 24/192 PCM? Does the ultrasonic energy of pure DSD cause intermodulation distortion somewhere in your playback chain? Does your off-the-shelf DAC chip’s internal processing treat DSD and PCM differently such that converting one to the other in your PC would make you think that “DSD sounds better than PCM”?

Ultimately, I think that DXD (aka 24/352.8kHz PCM) is the format with the least compromise for recording, processing and distribution (file size and transport compatibility aside). And I think that the DS DAC – which converts DXD to DSD256 on the fly – showed us the very best architecture for D-to-A conversion at lounge room scales (with the caveat that our amps and speakers need to be OK with the ultrasonic noise that makes it through the low-pass filter).

3 Likes

There are two main benefits for DSD that I am aware of:

  1. The sharp DSP filters required in PCM ADCs and DACs are avoided completely in DSD. Even though most PCM uses oversampling (like DSD) it gets subjected to fairly steep filters that cause ringing and smearing in the microsecond timing of really fast impulses and transients that your brain uses to discern three dimensional space. That is why PCM seems to get better in some ways when you raise the sample rate, higher sample rate equals greater bandwidth, equals more accurate timing resolution. DSD has massive bandwidth and essentially no digital filtering at all. The filtering that needs to happen to reduce noise can be done with gentle analog components like resistors or transformers. In PCM most of the time the filters are on tiny, limited-DSP chips that add their own processing artifacts.

Some people say that going to 192khz or 352khz (DXD) PCM fixes this problem, but it really doesn’t. I have recorded audio with high end ADCs at both sample rates, and they lack the three dimensional, fluid, fast, and resolving sound of DSD. You can get some of that back with the Directsream DAC, Playback Designs or Meitner DSD upsampling DACs, but there is always going to be a little depth missing. (this isn’t to say that PCM doesn’t have its own advantages)

  1. The other one, which Ted talks about a lot in the DS manual, is that 1 bit is more linear and resolving at low levels than multi-bit PCM.
4 Likes

As said question should really be why do they sound different?
Here is my simplistic understanding. PCM is digital it captures the signal as discrete digital values. DSD is 1 bit PDM (Pulse Density Modulation) and analog. It captures the signal as 0’s and 1’s bits the density of which ‘mirror’ the rising and falling signal. Both capture systems are very different and therefore it stands to reason on playback you will hear a difference. DSD benefits from being closer to the original signal as PCM effectively requires a further conversion step so perhaps preserves the music more accurately?

1 Like

DSD is not analog as I understand it. It is just a slightly different way of doing digital.

Until recently I would have agreed. My current understanding is that though it uses 0’s and 1’s they have no digital word value - they simply mimic/mirror the varying signal. PCM uses definite word/bits values 16/24 whatever and so codes the music into a digital blueprint. One of the big criticisms of DSD is it cannot be digitally mixed like PCM and that is because it has no value word information in its bitstream to process.

I think you are misunderstanding what digital means. Having discrete fixed values that are either there or not is digital. It has nothing to do with word lengths. Also the reason it can’t be digitally mixed is because the hardware/software doesn’t exist to do it easily. If DSD had caught on it would definitely be available.

Really difficult to do any sort of direct comparison that is particularly meaningful, for many of the reasons stated. I think nowadays, a very good PCM-based DAC playing back high res PCM-recorded material should be equivalent to a DSD DAC playing back DSD material.

But again - tough to compare, as then you’re going to have differences in the analog stages of the devices used (throughout the chains) that are more significant sonically than the choice of PCM or DSD as recording and playback format.

1 Like

Sorry I’ll try again, PCM is digital it has samples expressed as fixed value words. DSD does not have fixed value words. You cannot digitally mix DSD as DSD because it has no value words to process. You have to convert it to either PCM - frequently DXD ‘Digital Extreme Definition’ which is a fancy name for PCM at 352.8 kHz 24 bit or as in the Sonoma system to what is called Bit Wide i.e more than 1 bit. Many people mix it through an analog mixing desk to avoid the PCM conversion. You say it isn’t available? try looking at NativeDSD.com or bluecoast records for starters many amazing recordings there. It’s interesting that many small labels recording fantastic acoustic music in particular choose this system over far easier PCM based output, it suggests to me they feel it has a sonic advantage?

The reason people need to resort to workarounds is not because it is not possible to do it but because the tools to do it properly don’t exist. That does not mean it could not be done digitally. If it exists as 1 and 0s it can be edited and manipulated which is all mixing, editing and everything else is.

Also not sure why you are telling me that there is a sonic advantage to DSD. I already think there is.

Correct, the tools don’t exist and won’t do unless someone fronts up a very large sum of money. Again please be aware of how I’m referring to digital. The 0’s and 1’s in DSD have no ‘value’. All you can do with DSD in its native form is edit it by ‘butt splicing’ literally cutting it and sticking back together again. As I understand it a workstation to mix DSD without conversion to PCM is preferable as this in many peoples opinion is detrimental to the sound however small it may be. The Sonoma system creates several individual bitstreams that allow mixing without conversion to PCM.

I’m not telling you anything, simply making a valid observation for general discussion. I agree with you, I believe DSD offers a sonic advantage for the reasons explained earlier. An additional observation is the number of people who are converting PCM to DSD in music software before it reaches their DAC’s. It makes sense given the large number of DAC’s using Sigma Delta convertors that prefer being fed a DSD stream.

2 Likes

OK - let me try to construct a Thought Experiment.

Pick whatever front end you like, a pair of the best mics in the world into the most neutral preamps in the world, with your favorite artist playing acoustically in a nice-sounding room to these mics.

This analog signal is passed via The Best ADCs Available for each format, to DxD and DSD recording chains simultaneously, with no difference between the signals delivered to the analog ins of the ADCs.

This (now digital) stereo recording is sent through both chains without any further processing. And for the purpose of the thought experiment, the “resolutions” of the two recordings add up to a similar number of total bits or a similar-sized file - or whatever you think is fair.

These recordings are then played back through whatever the best/most transparent DACs are available for Each Format, to the same amp and speakers/system of your choice.

(Nota bene: Here’s where it falls down, as the DACs and their analog output stages would tend to be different.)

Do you Imagine (because, practically, one can pretty much only imagine this) that you would be able to tell which format is which? Aside from the DAC-makers’ “sound”?

I’m suggesting that this Format Question - nowadays - is of less importance/impact than most other aspects of the recording chain.

I say this as a lover of the “Sound of DSD” compared to PCM from DSD Day 1. But that was like…forever ago in digital terms. And I think it perhaps had more to do with things other than the storage medium even then.

I just don’t know now, as PCM is SO MUCH LESS crappy than it started out ; )

Sorry for injecting my bias into the Gedankenexperiment : )

2 Likes

Agreed, and I unable to think of anything less important than format at this point of in the history of recorded music. Both PCM and DSD are superior to our ability to play back the recording in any event.

1 Like

From my admittedly somewhat limited exposure to DSD64 vs. Hi-res PCM (96/24) I’ve come to the observation that, with superbly engineered recordings, both formats are stunning in retrieving detail, ambiance and imaging/soundstaging but DSD has the advantage of being more natural sounding with regard to timbre to my ears than PCM.

Mastering is still more important, in my opinion.

2 Likes

It’s always interesting to understand when and why Americans use German words within their own language…and they often do.

Partly I guess these are words that somehow sound „cool“ to them in their wording/accent, like „angst“ or „über“…but sometimes there are just very bulky sounding words like „Gedankenexperiment“ that just seem to sound very strange and complicated to them, still worth using … is there anything that explains this? Just curious to know and understand, thanks :wink:

I studied Physics at a British university, and notional experiments were always referred to as gedanken experiments, probably because the phrase originated in Germany. English contains a significant number of ‘loan’ words (unlike French, where the Academy loathes them and always tries to force the usage of ‘French’ equivalents). Since English is basically Anglo Saxon, with origins in Old High German, with a substantial injection of Norman French, this is hardly surprising.

1 Like

Thanks, I wasn’t aware that German words are so commonly used even in scientific environments

Einstein used the term, popularizing it - especially in the States where he was doing his work.

As to language origins, I have always found fascinating that roughly a third of English is from French.

1 Like