I’m pretty new to DSD…but, what I’ve been reading from the ol internets is that it’s difficult to record and edit entirely in DSD throughout the entire process.
I’ve been reading that due to this difficulty, sessions are often recorded in high resolution PCM(or recorded in analog then converted to PCM) - then edited in PCM - and then converted to/encoded in DSD as a final product for SACDs. Is this always the case - no matter what - or can PS Audio’s Sonoma editing station record, edit and stay purely in DSD through-out the entire process?
One of the reasons why DSD original recordings typically sound good is because the engineers know that there is very little to edit, at least if they don’t want to convert to and from PCM a couple of times.
Consequently much more effort is put into capturing the sound correctly instead of trying to tweak it till it sounds reasonable with a lot of loss of details on the way.
I have not encountered one poorly recorded song at Blue Coast Records, who also record to DSD natively.
The result are distributed as PCM as well but the process is DSD, as such the DSD files you download there are real unconverted DSD files.
Hi Ryan. It is true one cannot edit DSD directly. Most of the DAW systems out there using DSD convert to PCM, edit etc., then convert back to DSD. Sonoma is different.
In the Sonoma DAW system, the music remains DSD throughout the chain except when there’s an edit. If the engineer is splicing or cutting, just that tiny bit is converted to PCM and then immediately back again after the edit. Plus, the way they
convert it to PCM and back again is unique (though I don’t remember how).
According to Gus Skinas who (along with Ted Smith) was part of the original Sony team that built Sonoma, the edits are indistinguishable from pure DSD.
On the Sonoma when DSD is edited, it’s allowed to widen to 8 bits which is PCM like, but it’s still at the DSD rate. That is clearly lossless. Then when needed, it gets remodulated to 1 bit. So over all it doesn’t stay one bit, but it’s never exactly PCM: it is always noise shaped (i.e. it needs to be lowpass filtered to become PCM) and it’s sample rate is never changed (which would require convertion to PCM for heavy filtering.)
Another way to use DSD with no PCM involved is to use DSD like tape, always edit in analog by playing the DSD and record the result back to DSD. That may seem round about, but it sounds better than using PCM with a sample rate less than the sample rate of DSD.
P.S. Just in case, let me clarify Paul’s statement above, I never worked for Sony, nor was I involved with DSD development. I did know Gus Skinas before then, and I’d met Andreas Koch and Dr. Andrew Demory (who were integral to DSD development) back then.
Andreas Koch sells equipment, he probably rightly claims is very much Sonoma based.
I looked at it. it’s probably extremely high end and will sound very good, but the price. Even if I had the budget. I’d happily settle for the direct stream and add the BHK S250 and (now there you go, I already forgot the new name for the AN3) a pair of FR (???) 30, for approximately the sam budget.
I wouldn’t think of it that way: it’s still noise shaped the same way. Think of it this way, DSD is essentially a stream of mixed +1’s and -1’s. If you multiply each sample by 2 or 0.5 have you really changed anything? How about summing a stream x with the doubled stream x, now you have a stream 3 × x, but nothing has changed format, the samples are just a little wider.
PWM is a stream of times, at each time the output flops. Linear processes on DSD don’t some how magically cause the format to change from a sampled format to a PWM format.
8-bit DSD is still in the essence of DSD, it’s not the fact that DSD is usually delivered to the customer as single bits, it’s that the stream has a high sample rate, that it is noise shaped and that only gently sloped filtering is needed to go to PCM. (To get to PCM it needs to be lowpass filtered.) Similarly when editing PCM the samples usually gets widened to not lose accuracy, at the end (and perhaps other times during editing) those samples needs to be dithered back to, say 24 or 32 bits… DSD and PCM editing aren’t that different in that sense: during editing, widen precision of the samples to keep accuracy and narrow the samples when required with either dither (for PCM) or remodulation (for DSD.)
At the time the Sonoma was being designed editing DSD at, say a full 24 or 32 bits, wasn’t practical so they used 8 bits (and FPGAs). Any new high quality DSD editor would probably work on wide samples (24 or 32 bits) at the final sample rate (or higher), perhaps store the intermediate results in wide DSD, and narrow them down for final delivery to the customer.
Some of the best recordings out there are done by playing a tape thru a console and recording the modified material. Arguably DSD can out do tape these days so (once again arguably) it’s the best way to edit.
IMO, PCM’s problem when editing, etc. is simply too low of a sample rate: Get the sample rate up and I think it would sound as good as DSD (which is almost what I was saying a few posts ago.) Then the question is if high rate PCM or DSD is more faithful to the source thru a sigma delta ADC (which is what most ADCs are these days.) Arguably DACs based on filtering DSD rather than decoding PCM are more faithful to the material.
Question: Is there a SOFTWARE that could do this? or needs the Sonoma?
If no Software, that seems to be the Achilles heel, and one major reason it didn’t take off. Which really sucks. I am in strong belief that one of the most troubling part of high end 2-channel audio is the LACK of standards! Unlike Video, or even home theater audio.
Almost all Audio/Video companies move with standards set by industry, being 1080P, 4K, HDR, Dolby Digital, DTS, Dolby Atmos, HDMI, AptX audio, BT 5.0, etc… As soon as a new standard is introduced, companies RACE to offer that on their next product, hence it revitalizes the industry, ppl interest, and keeps the industry alive! You see now everyone wants a 4K TV, WITH HDR, OR, want AptX with their latest headphones… as that’s the latest and greatest.
NO SUCH standard exists for 2-ch audio, hence DACs are all over the place, some support DSD, some don’t, PCM all over the place, some high-res, some not, 16-bit, 20-bit, 24-bit; Some MQA, some not. Digital audio is in many different formats, no standard exist since CDs introduced in 1980… that’s WHY non audiophiles all get confused, turned off, and don’t adopt any of these all over the place standards.
There are professional editing systems that do a lot in software these days. The Sonoma hardware isn’t needed anymore (in principle.) But those professional editing systems aren’t cheap. There’s no reason someone couldn’t build a great, cheap DSD editing suite. I’ve had many ask me to do it over the years.
Interesting comment and I don’t disagree, but exactly those arguments are why I don’t touch video - way too many standards (as you listed), far more file formats, confusion over what will play what, etc.
Absolutely, I do utilize an AV receiver for stereo only. The reasons I do that are:
it’s a singular control station for ALL audio and video sources, hence a single remote
it has enough power stages on board it can bi-amp my full range floor standing speakers
optically a perfect match with my multi source digital player and turntable, thus minimalistic (from components point of view) set up that does it all including streaming with my little AppleTV
But I never ever utilize any Dolby or DTX algorithms, those sound horribly artificial, even for movies but certainly for music, the best mode is “source direct” which bypasses all Dolby and DTX DSP processing.
The best audiophile thing closest to this is the MOON 390 pre amp. It features all connections I need in an what I would hope better sound quality.