p.s. that’s not to say it will sound identical, it won’t, which is why different implementations sound different, but it’s not inventing junk or even simple interpolations to fill in the “gaps” 
1 Like
Incidentally, as for tonal content and monitoring, I commented a little while ago about studio monitoring (obliquely, in passing, as part of an anecdote 
) in the Now Spinning thread.
I will dig in over next couple days, thank you
Thanks. Just to be clear what this is. It isn’t “pure” DSD in that pure DSD is 1-bit and this is not 1-bit.
What the system does is keeps it as close to 1-bit as possible and extends to 5-bits or so. So, technically still PDM (which is what matters), but not technically DSD.
Hope that makes sense.
1 Like
Does the complexity of the process speak against using it for you or other reasons?
Oh no. It’s just the klunkiness of the interface and the limitations of the process. The only system I am aware of that can mix with this is the Jussi HD player that we use. It has no real user interface, doesn’t support automation (we routinely spend hours tweaking the track levels to get a good recording) and no plug ins. It’s likely perfect for what they are doing, but not for us. We spend weeks in the recording and mix rooms creating studio masters. If were instead to be recording only live music that required nothing more than a basic mix, then it’d be worth considering.
Yes, I guessed it’s not ready (yet) for larger production use. But if this audible difference identifies part of the loss of the DXD mixing process, it seems worth observing.
What is the point of imaging and soundstage? Surely it is primarily to be able to hear each instrument clearly with a pleasing overall balance. It may be quite arbitrary in a jazz performance where people actually stand/sit, depending on who is leading, who needs to see who etc. I got to a jazz venue where the band is in the middle of a church and are surrounded by the audience. We mostly hear reflected sound anyway.
I went to a show yesterday, in the second half there were 9 or 10 instruments sitting in a semi-circle, left to right: violin (1 or 2), viola, cello, double bass, horn, bassoon, oboe, clarinet and flute. A substantial part of the music was the visual as well as sonic interplay between instruments and sections. When you get to the baroque oratorios, then life gets really complicated. Yesterday were were in a smallish 350 seat hall that doubles as a top class recording studio. The sound is very clear and a little on the dry side.
Unless the listener knows where instruments are meant to be, and in studio-based music it’s all artificial anyway, as long as I can hear everyone, I’m happy. I assume a lot of mixes produce an overall sound image that never existed in real life. When I know where things are meant to be and it gets confusing, I listen to something else.
NativeDSD says about their “pure DSD” offerings… “No DXD or PCM Processing has been used.”
You say that this is not “pure DSD.” Do I understand you correctly? I wonder if NativeDSD is being false and misleading.
Second question(s): what happens to all the data in Octave’s conversion of DSD recording to DXD (ie PCM), ie going from 11.2 million samples per second to 352 thousand samples per second? And where does the data come from to convert the DXD back to DSD?
The Positive Feedback article by @Rushton noted that data is lost in the conversions, ie the conversions are lossy.
Bugs the hell out of me that I don’t understand this stuff fully.
In any case, I purchased a “pure DSD”, not knowing if any mixing or EQ was involved. I want to hear if the files present all the sounds that I know first hand emanate from a ‘double bass’.
A white paper by Thomas Kite notes, “PDM stands for pulse density modulation. However, it is really better summarized as “oversampled 1-bit audio”, as it is nothing more than a high sampling rate, single-bit digital system.”
You said a PDM extends to 5-bits or so and is technically not DSD.
NativeDSD, by „pure DSD“ seems to understand non-mixed DSD recordings as well as what I mentioned above, which Paul explained as PDM (so no PCM/DXD), but not DSD (as 5bit inbetween during mixing). DSD or pure DSD wohld be 1bit throughout.
I’d say Paul is more correct.
1 Like
I don’t think they are intentionally misleading at all.
Let’s separate out the terms. DSD is PDM (Pulse density modulation). It is a term coined by Sony to market their “new” format.
The DSD you and I would listen to, from DSD64 to DSD256 and beyond is a 1-bit system. That cannot be mixed or volume leveled.
Multibit PDM can be mixed and volume controlled. And if they want to call it DSD then they have every right to and it’s just fine.
My only point in this is for you not to think that the DSD 1-bit audio you’re familiar with is what’s being modified. It’s not.
Sorry. I know this may be confusing.
4 Likes
To modify, eg equalization and remixing/remastering, Octave converts from recorded DSD to DXD/PCM and then back to 1-bit DSD. Two (lossy?) format conversions.
Does anyone here record DSD via RME converter?
agreed
Re Autumn Rain In A Secret Garden (Pure DSD)
Sounds great. Solo recording with a sense of avant garde too. Demonstrates a wide range of the sound that can emanate from an acoustic bass.
Recorded as DSD256 and sold as DSD256, presumably with no conversions like to DXD, PCM, PVM, FLAC, WAV etc.
From the description…You can hear every little noise from the bass, so the whole piece must be played with the utmost attention, and if that attention is distracted for as little as a moment, you need to start the recording again.
I bet no produced, purchased track does so. Looking forward to examples that prove otherwise.
EG, Paul noted recently that while a bass player was on the right during recording, he (lossy) converts the recorded DSD to PCM to change the bass player’s position to the center and then (lossy) converts back to DSD for SACD, DSD, and vinyl purchases.
anyone have an answer?
perhaps in converting to DXD/PCM, the data in the 1-bit DSD is spread across 24 bits of PCM. Would all these bits get remixed/remastered/modified/equalized and then merged in conversion back to DSD?
1 Like
I think to remember that even level changes can’t be done in DSD, so the conversion process is not just done for mixing.
I also understood, if maybe not completely lossless, the Pyramix conversion for DXD mixing is not as lossy and is technically different from normal DSD/PCM conversion processes.
1 Like
Hmmm… haven’t been following this conversation and just now catching up. A few of thoughts to add to the mix (some of which repeat what’s been said, so apologies):
-
Tom Caulfield (NativeDSD mastering engineer) once explained to me that to begin to understand DSD (or any aspect of PDM), first forget everything you think you know about PCM. PDM is just different and all of that baggage in your head about how PCM works just doesn’t transfer to understanding PDM. He said he couldn’t begin to get his head wrapped around PDM until he set aside everything about PCM. In this interview at Positive Feedback, he discusses PDM and how he tries to think about it, here.
-
What Paul is saying about level changes and other editing with DSD is correct. From what I’ve been told, you can’t do it in 1-bit PDM which is the technical definition of Sony’s marketing name “DSD.” But you can edit, volume change, fade, equalize in multi-bit PDM. For consumer conversation, DSD is the term most people know. So discussion is often short-handed to simply use the term DSD even when multi-bit PDM is actually the subject. I’ve learned just to flow with that – be a duck in water on this issue.
-
Tom Caulfield masters Eudora’s Pure DSD releases in multi-bit PDM using Jussi Laako’s Signalyst HQ Player Pro. As Paul says, HQ Player Pro does not have a visual user interface so it’s not convenient to use. It does, however, keep everything in PDM and is thus considered a Pure DSD output with no PCM processing. See the article written with Gonzalo Noque about “Mixing in Pure DSD - No PCM Allowed” here.
-
Moving PDM into DXD is a lossy process (as noted in comments above), meaning you can’t go back out of DXD to the same signal. That original is lost. Sure, DXD in Pyramix, or other tools, makes a mastering engineer’s life easier. Sliders to your heart’s content. It just doesn’t sound as good once it goes into DXD in my opinion.
-
Pyramix has lots of alternatives for mixing. It is possible to use Pyramix in a DSD project for some very limited functions, like making splices, without moving the original file into PCM/DXD. You apparently can make the edit process a “DSD Project” rather than a “DXD Project.” As the user, you see DXD for purpose of tracking the music, but the original file is never actually converted to DXD except at that splice/edit point for about 10 milliseconds (or less). In those 10 milliseconds at the edit point, the signal goes to DXD, then out again. Can you hear 10 milliseconds? I certainly can’t. The completed project is then saved as a DSD output file with the “10 millisecond” edit points. Tom Caulfield considers such a DSD Project in Pyramix to also be “Pure DSD”.
-
BUT, if the project in Pyramix is a DXD Project, then the entire file is converted to DXD for editing. Pyramix’s Album Publishing will take that DXD file and output whatever format(s) you want: 44kHz, 96kHz, 192kHz, 352.8kHz, DSD64, DSD128, DSD256. But, the output is now a PCM file that has been converted.
-
Jussi (and Tom Caulfield) will say that PDM is PDM and moving from 1-bit PDM to multi-bit PDM is not a conversion – it stays Pure PDM. The reason for this is all to do with the completely different paradigm of working in PDM versus working in PCM. I am not the person to explain this - I don’t understand it. For Jussi, it’s just math.
-
Modulating DSD to different rates of DSD (e.g., DSD64 to DSD128 to DSD256) is not upsampling as we understand this in PCM. It is a frequency modulation - like an FM radio wave. There are not samples being taken and multiplied because there are no samples as in the PCM world.
Okay, this is about it for me, and more than I really know. I’m not the expert by any stretch. I’ve just been trying to absorb what the experts try to share with me and then serve as a messenger hoping not to garble the message beyond all comprehension. My suggestion, read the articles I’ve referenced. Listen to the sample files - as I see some of you are actually doing. (Thank you!) Listen to what your ears are telling you.
15 Likes