Sorry Ted forgot to explain. Made the digital purchase. Got it in my PC now streaming to the DSjr DAC via Bridge II.
I realize now it must be because its limit that way is DSD64, right?
Yes the limit for the Bridge is DSD64.
2 Likes
Ted, I have a Stellar Direct Stream DAC, unfortunately this current generation of firmware leaves the owners in the dark about the file format/bitrate/sample rate playing. I have DSD 128 files from Blue Coast Records. Converted them using DSDMaster into hybrid DSD iTunes format in order to add it to my iTunes library. But then?
My understanding is:
When I utilise BitPerfect, bit perfect sends the DSD information to the USB port, the DSD signal gets bundled and sent as digital packages of similar size as high resolution PCM (32bit) to the USB input of the SGCD which puts the packages in the right order in a buffer (digital lens) and natively converts the DSD stream utilising the SABRE DAC chip.
Is this correct? Am I missing something?
Now, in case the display on the would SGCD display the file format, what bit-rate / sample-rate should it display.
The problem I have is that I have no clue about whether the DSD part of the hybrid file is playing or the iTunes apple lossless file.
I’m sorry but I don’t know anything about the Stellar line. You might email or call support…
1 Like
Thanks Ted, so I did more research myself.
This will help Stellar Gain Cell DAC (SGCD) users who want to ensure that BitPerfect plays the DSD part of the Hybrid DSD file instead of the PCM part.
On the “roon” site I found a link to a statement from Ted, stating: “USB will play DSD through DoP only, I2S will play up to DSD 128 native only.” The PS Audio specs of the SGCD confirm this for the SGCD too.
DoP will in no way harm the DSD signal, this is from the dCS’s website:
"FAQ: What is DoP (DSD over PCM)?
The original idea for DoP was invented by dCS in 2011. It involves taking groups of 16 adjacent 1-bit samples from a DSD stream and packing them into the lower 16 bits of a 24/176.4 data stream. Data from the other channel of the stereo pair is packed the same way. A specific marker code in the top 8 bits identifies the data stream as DoP, rather than PCM. The resulting DoP stream can be transmitted through existing 24/192-capable USB, AES, Dual AES or SPDIF interfaces to a DoP-compatible DAC, which reassembles the original stereo DSD data stream COMPLETELY UNCHANGED.
If something goes wrong and the data stream is decoded as PCM, the output will be low-level noise with faint music in the back ground, so it fails safely. This can happen if the computer erases the marker code by applying a volume adjustment."
This all means BitPerfect will play DSD on the SGCD flawless in case:
- DSD 128 files are converted into into hybrid DSD utilizing DSDMaster with 192 kHz setting (DSD Master, Output Format, preferences menu). Because dCS DoP is a 24 bit 176.4 kHz data package, 384 kHz conversion setting will not increase the audio quality in any way. Also on m SGCD, it may be the 5 m USB cable, 384 kHz does not play smooth at all. Many blips and blogs, with 192 kHz setting in DSDMaster file conversion those blips and blops are gone. So I recommend stay with the 192 kHz setting.
- “- DSD128.DSDh.” is added by DSDMaster to the music file name right before the extension of the file
- Disable “DSD 64” and “DSD128” checkboxes from BitPerfect, only enable “dCS DoP” in the BitPerfect DSD preference menu. BitPerfect WEB-site confirms that BitPerfect is not able to identify the DSD modes a DAC will support, s this must be set with caution manually by the user.
- In case “Automatically add to iTunes” is not selected in the DSDMaster, Output Location, preference menu, import the generated DSD hybrid file into the music library, drag and drop into iTunes suffices.
- Add a cover art to your taste.
- Hit the play and enjoy.
Ted, I still have a problem understanding what happens when DSD is edited/mastered. Not sure if this was discussed already in depth, but couldn’t find it.
Let’s forget all the DSD/SACD‘s containing PCM hires, but just talk of native DSD.
Then there are those like the few native DSD recordings of Todd Garfinkle / MA Recordings, who doesn’t do any editing/mastering as he says. He does everything needed during recording and that’s what’s on the DSD (or PCM) file.
But what I speak of now are all native DSD which are either edited/mastered in PCM or analog, then converted back to DSD.
How does it come, that the DSD quality is preserved although it was forth and back converted to „inferior“ formats inbetween and how does it come that the limitations of the interim formats don’t apply to the final DSD file?
Editing in DXD (32 bit floating point at 352.8kHz) is indeed audible, but barely. If the editing is a pure splice (say 10 milliseconds) then only that part of the DSD needs to be converted to DXD and back. This is pretty transparent.
Doing volume changes or other things that need DXD for extended time is less transparent.
The Sonoma doesn’t convert to PCM per se, it uses “wide DSD” which allows single bit DSD to grow to 8 bit sigma delta at the DSD rate, e.g. nothing is lost going to wide DSD. So doing volume controls of multiples of 1/256 (or 1/128 if you want to be able to invert the signal too) and mixing with other wide DSD signals is fine. At the end you need to convert back to single bit DSD, but that conversion isn’t an audible problem. So Sonoma edited discs sound pretty good.
I’d argue that analog is the gold standard in that that’s how we get the signals in the first place and how we play them. A goal could be to stay as close to that signal as possible. Then the question is “Does doing a DSD → analog, mixing in analog, and doing analog → DSD sound worse or better than using, say DXD?” FWIW there are CDs out there that were recorded digitally, mixed in analog and then mastered back to digital. I like their sound in general.
7 Likes
Hi Jazznut. I agree, a fully native path without any conversion is the path where the least information gets lost. Whether it be PCM or DSD by the way. I am a big fan of Todd Garfinkles work.
Talking about Todd Garfinkle. It’s the time of the day to put on Sera Una Noche. It is recorded with 2 custom microphones, tuned to capture the ambience sound, in a church just outside Buenos Aires in PCM 24 bit / 96 kHz.
I have the songs Malena and Nublado on 45 rpm and the entire album 13 beautiful songs on redbook CD, all converted from that on a above mentioned master file.
I can certainly hear the ambience and the details of the instruments and placing, regardless the medium or signal transport (streaming).
Like most things in live we need to make compromises based on available space, budget and taste. I can imagine that once you take your system to the extreme the path without conversion is the holy grail. But I doubt many people can differentiate the difference due to limitations of their system, room acoustics and their own hearing.
The NAD M33 was a eye opener to me. It converts the hack out of every audio source. My Blue Coast Records (native DSD recorded DSD128 files) are converted to 24 bit 96 kHz by the BluOS software.
Most rediculous case:
The above mentioned 24 bit recorded Nublado song on vinyl. It is converted to analog, processed through a RIAA converter (compressor really), then cut to master record, then pressed from master record, then fed analog into my M33 phono stage, then RIAA de-conversion (de-compression) performed analog, then converted to 24 bit 96 kHz, then PCM processed (DIRAC), then converted to analog, amplified by a class D amplifier and transported to the transducers that cause the biggest loss in sound information in your room acoustics.
We live in a very fortunate time that technology has advanced so much that the music is played enjoyable even though converted so often. I simply enjoy the flexibility and ease of use of my BluOS system
Nevertheless:
I do hear a slight difference between spinning the CD or streaming, the difference is bigger than the difference between CD and Vinyl, with sonic preference for CD in this case.
However, nothing beats putting the Nublado record on the turntable, pour in a good brandy or whiskey, look at the beautiful artwork in hand on carton cover and dream away into that church just outside of Buenos Aires.
1 Like
I have the fantastic Sera una noche as hires digital, the 45 RPM (RTI) and also the 33 RPM (Pallas) double album. Both vinyls were mastered with different equipment.
That’s what I wrote to Todd when asking for my impressions:
I got the 33 RPM today and compared it to the 45 RPM and the 24/96 file.
The 33 RPM has more weight which favors some tracks with a little brighter sounding instruments and it sometimes seems to have tiny bit more 3D rendering of single instruments or voices, but I guess that’s rather an effect of the bit more weight.
The 45 RPM sounds more transparent, mostly due to the slightly more mid and top end focused tonality. Some instruments appear more clearly than on the 33 RPM. It also sounds a little more taut in bass therefore.
Soundstage of both is similar but a slo a little different, the 45 RPM more transparent but slightly less big imaged instruments.
The 33 RPM has a quieter surface.
The 24/96 files on my DAC (PS Audio Directstream) sound more like the 33 RPM, but a bit worse than the LP’s, but that’s in this case probably due to the fact my turntable sounds better than the DAC, also in usually digital’s strengths.
Finally I’d say it’s mainly a matter of system tonality which LP to prefer. I’d say the 45 RPM provides the most information and therefore is best.
1 Like
Didn’t I read that Cookie Marenco edits DSD by going DSD->analog->DSD ?
Who needs PCM?
If I remember correctly Cookie uses the Sonoma system as well, but you could be right too. Anyway, her recordings sound amazing too.
The point is, I guess, that if recording and mastering engineers / studio’s are investing in equipment handling DSD in that way, they mean serious business and are passionate about delivering the best quality recordings. Be it DSD or PCM. Which Cookie also writes on het WEB-site. She sells DSD (native), PCM converted high res, redbook CD and sometimes vinyl. She also mentions to have no preference when it comes to signal transport and reproduction.
The music and recording quality is what matters most.
1 Like
I do. Most music I like is either never recorded in DSD but certainly not available for purchase. I have very bad experience with trying to buy Nils Lofgrens “Acoustic Live” DSD mastered by Gus Skinas, by the way. That DSD album (file) seems to leave audio enthousiasts outside of the USA forsaken. Due to bloody copyright limitations. While I was willing to pay, and not a few bucks either.
So in my opinion: Who needs DSD? -those who invest a lot into their system, do not mind the hassle of trying to keep the signal native and have access to the media. For me it has all proven to be too cumbersome. It distracts from the fun of listening to music.
2 Likes
Thanks much Ted!
What I understood is, both DXD and analog editing are quite transparent in case of tiny editings but less in case of bigger ones. In both cases of bigger editings it’s hard to still speak of a native DSD format, but the analog editing usually results in very good sound.
Sonoma editings seem to be the most native DSD pure digital versions at the end, but, so I assume, not necessarily better than analog edited ones.
But there are still a few things I didn’t understand yet:
I can imagine, that pure short PCM editings don’t have to mean that the whole DSD stream is converted. But as you said volume changes or mastering activities (EQ changes etc.) should mean, the whole stream/track must be forth and back converted, correct?
In cases where in case of bigger editing/mastering the whole track is converted forth and back either to PCM or analog, how can the end result in DSD be any better than each of them or overcome their limitations?
Is the Sonoma able to do a large extent of editing/mastering in its transparent way?
If so, why is analog editing preferred anyway e.g. by Octave Records?
Is there a reason why with DSD there’s less need for editing or mastering to expect than for other formats?
You did and that’s the same thing we do. We are working on an even more advanced system for the future, but for now that’s what we do.
A volume change (and other more complicated edits) requires leaving one bit DSD land. But going to analog or wide DSD doesn’t have to be a problem. They can be very transparent (especially in DSD128 or higher.) A volume change that goes thru lower rate PCM does seem to loose something, but still can be pretty good if used with discretion.
I never claimed that DSD is better than analog, but it doesn’t have to be worse (and it’s certainly better than sending analog thru many devices.)
Both editing with the Sonoma and editing via conversion to and back from analog can be very transparent. The Sonoma is used by many as their complete DSD editing system.
I thinks it’s more that the kind of people who use and value DSD do less intrusive editing, not that less editing is needed. Very little but careful editing can get great results in PCM land as well.
5 Likes
Is there any meaningful downside to wide DSD? This appears to be the solution if the rate is sufficiently high.
Thanks again Ted, I appreciate this very much!
I didn’t have the topic analog vs. DSD and which is better here, I just wondered how e.g. the dynamic range of DSD can be preserved with analog editing/mastering?
I understood that analog instead of Sonoma editing (as both might produce similar good results) then maybe might be chosen for sound cosmetics/tonality reasons.
I can imagine that the ideal DSD transfer might be one like other flat transfers…with a recording source as good, so no further mastering is needed.
It’s just when I e.g. listen to Patricia Barber‘s Cafe Blue (unmastered edition) and compare with the different masterings available, I hear, how much of the perceived „magic“ of recordings comes from mixing/mastering (or can do so). This reminds me of direct disc vinyl cuts vs. mastered tape transfers. The direct to disc recordings sound very dynamic and direct in an unsurpassed way, but often also sound a bit dry and missing some ambiance characteristics we like and which seem to be done during mixing/mastering.
It would be interesting to dive deeper into pure DSD production vs. formats where more mastering can be applied from a mastering engineers point of view and the options he would use if he could vs. one shot straight.
1 Like
IMO it’s a simplification of the more general “edit DSD at its own sample rate”. I.e. you can do all of your favorite PCM things on DSD fairly transparently if you don’t downsample before the edit (and upsample after,) At the time the Sonoma was designed there wasn’t enough horsepower to do full sample width editing at the DSD sample rate. Sonoma’s answer was to use multiple FPGAs to do eight bit wide editing at DSD’s sample rate. It turns out that restricting volume settings to, say 1 / 128 to 128 / 128 isn’t much of a restriction.
These days we can all of do the editing PCM enthusiasts like at DSD’s rate and have the best of both worlds. The optimizations come into play when deciding when to modulate back to one bit or if you have enough storage to keep things at 24 or 32 bits at the DSD sample rate.
4 Likes
Some editing in the past was done with analog consoles and bouncing to tape and back. Using DSD as a tape replacement and editing that same way is a great way to preserve as much quality as possible.
A lot of wonderful work was done with tape and DSD can do better than that when used carefully.
4 Likes
I think I didn’t fully get the first part, but as you said it also seems DSD „has“ to be done carefully to actually „be“ DSD and have/preserve its advantages.
I still have two questions if you can/want to answer them.
What percentage of all DSD/SACD media do you think was produced preserving DSD meaningfully (so no PCM sources pressed on an SACD, no bigger PCM editings of native DSD etc.)? 20%? 60%?
Does analog DSD editing preserve the dynamic range of digital? I assume it must, as not more than a console is used?