That's easy. Give me a short PCM test file. I could pick (or make) one myself, but then you'd probably accuse me of cheating.
Last edited:
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
The sound quality of DSD?
- Thread starter Saidera
- Start date
AnalogSteph
Major Contributor
Ever heard of digital filter periodic passband ripple? The ADC on the ADI-2 Pro natively has quite a lot of that by modern standards. This is commonly being done in modern ADCs to reduce latency, and latency is a Big Topic in pro audio circles. While the chip itself would also have a "classic" mode (low ripple, higher latency), RME eventually provided a firmware update that uses the built-in DSP to implement a filter that substantially reduces said passband ripple (though it still remains higher than the +/- 0.001 dB that some classic ADCs will deliver, e.g. AK5393, AK5394A), presumably while still delivering low latency. The unit under test may not yet have had that.
I still wish manufacturers would give you a choice, as there are plenty of applications left where the last millisecond of latency does not matter. So probably would Julian Dunn (R.I.P.), whose '90s papers on digital filters (including passband ripple) are worth a read.
That one is spec'd at -108 dB THD+N, wonder what they used in terms of ADCs? AK5390, CS5390? The CS5394 wouldn't even have been out until late '96, I think, neither would the AK5391. A fancy ADC like the Blue 245 would no doubt have had a first-rate analog stage in any case.
BTW, user @jerryfreak posted some RMAA results of the aforementioned Benchmark AD2402-96. It's very good for a 20-year-old converter, but you did see some quirks, like a modulated noise floor on silence that suggests some idle tone issues in the CS5396 ADC (a 7th-order dual-bit delta sigma job - stability of such a high-order modulator seems dubious to me).
I don't see very much of a problem with the usual "burnt steak" approach to audio A/D filters - even in single speed mode, any aliasing is confined to inaudible >20 kHz, and if you oversample and perhaps add some input filtering to boot, it pretty much becomes a non-issue. You need decently low clock jitter for the filter to do its job properly, of course, but that's been known for many years now.
As an aside, I would bet no small sum that it takes no more than some EQ with a combination of FIR and IIR filters to make just about any half-decent ADC or DAC behave pretty much exactly like an ideal one in the audible range (nonlinear distortion notwithstanding). What I've seen / heard in terms of deltas between converters usually indicated small deviations in frequency response (and/or phase) only. And yes, that's basically the equivalent of the Carver Challenge.
Last edited:
- Joined
- Oct 25, 2019
- Messages
- 11,111
- Likes
- 14,774
Mustering his troops for a well thought out counter-attack.
What now?
andymok
Addicted to Fun and Learning
To put things into context and perspectives ......
https://immersiveaudioalbum.com/2020/06/qa-with-morten-lindberg-2020-grammy-winner/
Q&A with Morten Lindberg: 2020 Grammy Winner
// There has been much focus on word length, but to me the sample rate is more important as impulse response connects directly to our primal sonic perception on a subconscious level. //
http://www.lindberg.no/norsk/artikler/004.pdf
The advantages of DXD for SACD
// To set the DXD format in perspective, it is relevant to address the initial motivation for actually defining the DSD format. There are basically two approaches to DSD. One uses DSD only as the actual audio data format on the SACD, while the other uses DSD as an acquisition format i.e. a format in which the initial sampling of the analogue audio is made.
Acquiring audio data in DSD makes sense in some recording applications, but the initial reason for introducing the format was the need for Sony Music to archive numerous analogue tapes in a digital format. For that purpose the DSD data format was used as the — at that point in time — best way of generically representing digital audio data.
...
The DXD recording is based on an A-D convertor directly outputting the DXD audio signal. The advantage of this type of recording is that the DXD signal can be edited directly thus being a higher resolution format than DSD. A benefit of this is that the out-of-band noise level is significantly lower than with the DSD signal. The editing process is relatively insensitive to changes in the dynamic range relations between the recording channels and in the final mixed music. DXD is acknowledged by Philips and Sony as a high sample rate and multi-bit processing format for recording and editing for SACD productions.
...
The data rate for a DXD mono signal is 24bit times 352.8kHz, which is 8467,2MHz, being three times the data rate of the DSD64 signal.
...
So how do you make the ideal recording for SACD? Because of the technical aspects of DXD, DXD is the most suited format for recording a signal with the highest quality. Since DXD is a PCM-like signal, editing can be made using conventional lossless digital processing, just operating at 8 times FS. The signal can be converted to DSD for authoring the SACD. DXD can also be converted to the standard PCM rates using sample-rate conversion still maintaining that the final mastering format will set the quality standard for the music production. One drawback of DXD is the high data rate of the signal as it is 8 times FS but with the evolution of workstations this is only a temporary limitation. //
i.e. production-wise it is better to use DXD; for delivery, playback and archive it's just easier on DSD to achieve at least the same technical performance.
by archive it means to store and reproduce at various incompatible fs (44.1/48) without much error/loss
https://immersiveaudioalbum.com/2020/06/qa-with-morten-lindberg-2020-grammy-winner/
Q&A with Morten Lindberg: 2020 Grammy Winner
// There has been much focus on word length, but to me the sample rate is more important as impulse response connects directly to our primal sonic perception on a subconscious level. //
http://www.lindberg.no/norsk/artikler/004.pdf
The advantages of DXD for SACD
// To set the DXD format in perspective, it is relevant to address the initial motivation for actually defining the DSD format. There are basically two approaches to DSD. One uses DSD only as the actual audio data format on the SACD, while the other uses DSD as an acquisition format i.e. a format in which the initial sampling of the analogue audio is made.
Acquiring audio data in DSD makes sense in some recording applications, but the initial reason for introducing the format was the need for Sony Music to archive numerous analogue tapes in a digital format. For that purpose the DSD data format was used as the — at that point in time — best way of generically representing digital audio data.
...
The DXD recording is based on an A-D convertor directly outputting the DXD audio signal. The advantage of this type of recording is that the DXD signal can be edited directly thus being a higher resolution format than DSD. A benefit of this is that the out-of-band noise level is significantly lower than with the DSD signal. The editing process is relatively insensitive to changes in the dynamic range relations between the recording channels and in the final mixed music. DXD is acknowledged by Philips and Sony as a high sample rate and multi-bit processing format for recording and editing for SACD productions.
...
The data rate for a DXD mono signal is 24bit times 352.8kHz, which is 8467,2MHz, being three times the data rate of the DSD64 signal.
...
So how do you make the ideal recording for SACD? Because of the technical aspects of DXD, DXD is the most suited format for recording a signal with the highest quality. Since DXD is a PCM-like signal, editing can be made using conventional lossless digital processing, just operating at 8 times FS. The signal can be converted to DSD for authoring the SACD. DXD can also be converted to the standard PCM rates using sample-rate conversion still maintaining that the final mastering format will set the quality standard for the music production. One drawback of DXD is the high data rate of the signal as it is 8 times FS but with the evolution of workstations this is only a temporary limitation. //
i.e. production-wise it is better to use DXD; for delivery, playback and archive it's just easier on DSD to achieve at least the same technical performance.
by archive it means to store and reproduce at various incompatible fs (44.1/48) without much error/loss
Last edited:
AnalogSteph
Major Contributor
-v please - comms error
To me, it sounds like this person won't win any awards for his grasp of science. Why should we care?
And what here is new information, much less information that demonstrates the audible superiority of DSD?
Comms error indeed. I still have no idea what you meant about jitter and filters.
OP
Saidera
Senior Member
- Thread Starter
- #390
http://www.vaio.sony.co.jp/vaio/solution/DSDdirect/feat1.html
mikkun.blog.so-net.ne.jp/2008-10-30
Windows Media Player vs DSD Direct Player?! How can that be fair play? No wonder they could pull off the implicit message: DSD format enables full exploitation of info on CDs and enables you to hear what you've missed up till now. But then, Windows Media Player should play CDs well enough?
That blazing GUI led me to feel as though DSD sounds good. Indeed, Nishio Ayataka-san was behind the delivery of DSD in VAIO PCs.
He also had SBMD added into SonicStage Mastering Studio. The problem is that it requires DSD Mode and so is as rare as the K-1327 SBMD machine.
One asks: why all the fuss about a sample rate converter called SBM Direct (SBMD) that's from 1998 & that is so rare (only in SonicStage Mastering Studio, Sonoma, or hardware K-1327) and can only change DSD64 to 24 or 16/44.1kHz? And now that converters are everywhere and Saracon can be paid for, why SBMD?
From various sources including Sony Music Tokyo and research papers, we know that K-1327 captures SDIF from Sonoma, and uses many boards to process each channel separately - adding up to FIR 30k taps filter for each channel, using 64 bit processing, and does 'SBM' noise shaping on top - it is uncertain whether this 'SBM' is truly different from the infamous 'SBM' from the 90s. SACD whitepapers also say that this is done as a 'single stage' process which makes it sound truer to the DSD. Unfortunately the 44.1kHz file that results is flatter and somewhat different from the DSD64. The question is whether 24/44.1kHz can retain enough information to recreate the original DSD if converted back to DSD64, using a process called DSD Direct, which is the SBMD algorithm implemented in reverse - included in VAIO PCs from 2006 or thereabouts. DSDD only accepts 24 or 16/44.1kHz and creates DSD64 or above (if hacked). If it can recreate the DSD by focusing on the audible spectrum and shaping the inaudible to sound like DSD then we know that DSD as a format creates its unique sound.
By analogy, if wav is compressed to high rate aac using FIR 30k taps filter or sth, and then we take that aac and FIR 30k tap it back to wav is it still sounding the same? This is not DSEE HX for heavy compressions, but simply a question about high rate audio.
Alternatively there are people who listen only to SBMD audio:
Mr. Otaki mentioned how he plays a CD.
What is that method···.
DA convert the CD (DAC-2000), convert the analog signal to a DSD file with SONY's commercial equipment ADA-7000R (ADA-7000R, a DSD-compatible AD / DA converter with a studio original rubidium clock - but by now surely external clocks are good enough not to cause jitter, so ADA is now obsolete?), and then down-convert it to a PCM sound source (wav file?) Using SBM (Super Bit Mapping) direct K--1327.
By doing this, you are freed from the conflicting composition of digital vs. analog in terms of sound quality, and you can now listen to CDs and vinyl records without distinction.
It goes without saying that CDs are no longer played directly, and surprisingly, everything is played in this way, including the playback of vinyl records (!), And Mr. Otaki says, "It's a body that doesn't accept anything else. I've done it "(laughs).
There was a comment that audio playback was "DA-AD-DA" in the 2011 New Year talk, but it was probably this playback method.
"The reason why I started listening to CDs in this way was that the sound of the waves recorded in the wav for radio broadcasting felt noisy, and when I consulted with an engineer, I was advised that converting the wav to DSD would eliminate that noisy feeling. That was the trigger.
It seems that wav's habit is that the sound of waves and the barking of insects can be heard noisily, but with DSD, that habit disappears.
For Mr. Otaki, it would be nice if the sound of DSD is a large-scale music such as a classical orchestra, but even if it has a large number of sounds, it basically does not match the music of the 4-rhythm organization. , When down-converted to wav, it seems that the sound became ideal for Mr. Otaki."
VAIO also included software called DSD Direct, which converts wav files to DSD, and DSD Direct Player, which converts CDs to DSD and plays them.
Mr. Otaki is using a commercial device prototyped by SONY for the studio, and there are only a few in the world, so I do not think that the sound of the file created with VAIO will be the same, but at least the work content You can do the same.
"The latest version of digital audio for sound creators" Keiji Kakizaki (Author) Byakuya Shobo
Dialogue: Eiichi Ohtaki x Keiji Kakizaki, from pages 170-206 of the book
Otaki: Generally, I don't really like the stereo method itself. I love things overwhelmingly. It feels like it's mixed up. I feel the heat there. If the separation is good, it's cold, or the heat is low.
After that, it is certain that the amount of information is large for the flowing and gorgeous sounds, but it sounds like something that is too much for me. I feel that the extra decoration is in the way of my liver. It's a comfortable sound for me that the liver is very straightforward and directly transmitted. i.e. passion rather than large information amounts?
Therefore, it's okay to use the 3-point principle, but then I don't think MP3 is the best, though it's an extreme theory. Strongly push the 3 points of the feature. However, if it is emphasized to that extent like MP3, the information is a little too missing, but it is not a mistake to say that it is such a direction. To put it broadly, it is. I'm sorry for the DSD group Kaki-chan. i.e. prefers MP3 equivalents over DSD for unexplainable reasons
Kakizaki: No, no....
Otaki Hey. Everything was a coincidence, but I finally met. Moreover, this is the last series of mastering on the 30th anniversary of Long Vacation, but at this time of year, he came out as if he had been waiting for it, a dream audio. The long journey from 35H is now perfectly over.
I AD with DSD and dropped the digital out to a PCM sound source with SBM direct, but I don't feel downgraded at all. Rather than the raw sound of DSD, it's my favorite sound.
The better the audio, the more information there is and the denser it becomes. In the movie, it's a cinerama. The screen is too wide and my eyes get tired. Even if there is a lot of information, my eyes will not get tired. If the standard screen isn't enough, expand it a little more.
In Kakizaki, it's Vista.
Otaki Vista. Vista is the thinned out DSD with SBM Direct. Then, this is DSD MP3.
Kakizaki: That's right. It's like compressing information.
Otaki Yes. that's interesting. This time, that's why the DSD MP3 was made into a CD.
Kakizaki: Not being able to go to SACD means that there is too much information.
Otaki: The basis of my sound is rock'n'roll four rhythm. Even if you can show 4 people at 70mm, 4 people are 4 people. But if you're just four people, the standard screen is fine, but my sound has a lot of instruments, so it's Vista. If you have a lot of characters and a lot of sounds, such as classical music and opera, I think 70mm cinerama, that is, a rich sound like DSD, is good, but Vista is enough for my sound.
Kakizaki: I want to reduce the amount of information.
In this book, Mr. Otaki denied the need for a higher format than the CD called SACD. This encouraged a Revolution in my thinking.
Some dude really wanted to emulate SBMD DAADDA, and is an Ohtaki fan. Seriously, DSD getting rid of waves and insects harshness doesn’t seem worth the trouble. ADA-7000R is a key item (as it captures analog rather than doing FIR filter digitally), and from Fujimoto’s articles we know K-1327 has better? SQ than SSMS SBM-D. We also know Sonoma has a SBM-D software option likely the same as SSMS’s. we know Nishio allowed MoFi to test out SBMD CD vs DSD and they felt difference – perhaps too exaggerated …amount of info relates to sound 'thickness' I guess! So if DSD also has softness of sound we haven’t explained it yet!
How is PCM harsh? Why? DSD gets rid of harshness? Is that only for analog capture or also includes FIR digital conversion? It seems digital is too accurate and keeps the harshness. One should record PCM audio of 'waves and insects harshness' and then convert to DSD to see if it is the same via analog and digital methods. One could try to broadcast PCM or DSD over radio and see which is less harsh. And one should see if SBMD really is better than any other conversion method out there, including Weiss Saracon. Given that Sony releases 24/44.1 hires files in this day and age, of Lang Lang, Glenn Gould, and others, many of which are implicitly SBMD derived from DSD Masters.
The content may be quite maniac, or, to put it bluntly, a completely unnecessary book for those who just like listening to music. However, a book with interesting content that is irresistible for those who like it:
In this volume, "Chapter 12 Jitter" and "Chapter 13 High Sound Quality Technology" were particularly interesting. For example, regarding "high-quality CDs" that have been elaborate on the latest trendy materials, "there is no one that objectively measures the played sound and proves that the sound quality is good," and the final conclusion is as follows. :
I think that "high-quality CD" should be seen as one of the products born from marketing strategy rather than high-quality technology
Not that all the SBMD is rubbish, but the best sound is not necessarily SBMD, it depends on source. Hong Kong cantopop using SBMD was quite hot, while Ohtaki is a proponent of SBMD in Japan, and Yo Yo Ma's Box set (among others) uses SBMD for all the CDs. SBMD remains a secret particularly in regard to its noise shaping method.
Perhaps someone has more experience with SBM Direct and can dismiss this post as a randomly written waste of space?
mikkun.blog.so-net.ne.jp/2008-10-30
Windows Media Player vs DSD Direct Player?! How can that be fair play? No wonder they could pull off the implicit message: DSD format enables full exploitation of info on CDs and enables you to hear what you've missed up till now. But then, Windows Media Player should play CDs well enough?
That blazing GUI led me to feel as though DSD sounds good. Indeed, Nishio Ayataka-san was behind the delivery of DSD in VAIO PCs.
He also had SBMD added into SonicStage Mastering Studio. The problem is that it requires DSD Mode and so is as rare as the K-1327 SBMD machine.
One asks: why all the fuss about a sample rate converter called SBM Direct (SBMD) that's from 1998 & that is so rare (only in SonicStage Mastering Studio, Sonoma, or hardware K-1327) and can only change DSD64 to 24 or 16/44.1kHz? And now that converters are everywhere and Saracon can be paid for, why SBMD?
From various sources including Sony Music Tokyo and research papers, we know that K-1327 captures SDIF from Sonoma, and uses many boards to process each channel separately - adding up to FIR 30k taps filter for each channel, using 64 bit processing, and does 'SBM' noise shaping on top - it is uncertain whether this 'SBM' is truly different from the infamous 'SBM' from the 90s. SACD whitepapers also say that this is done as a 'single stage' process which makes it sound truer to the DSD. Unfortunately the 44.1kHz file that results is flatter and somewhat different from the DSD64. The question is whether 24/44.1kHz can retain enough information to recreate the original DSD if converted back to DSD64, using a process called DSD Direct, which is the SBMD algorithm implemented in reverse - included in VAIO PCs from 2006 or thereabouts. DSDD only accepts 24 or 16/44.1kHz and creates DSD64 or above (if hacked). If it can recreate the DSD by focusing on the audible spectrum and shaping the inaudible to sound like DSD then we know that DSD as a format creates its unique sound.
By analogy, if wav is compressed to high rate aac using FIR 30k taps filter or sth, and then we take that aac and FIR 30k tap it back to wav is it still sounding the same? This is not DSEE HX for heavy compressions, but simply a question about high rate audio.
Alternatively there are people who listen only to SBMD audio:
Mr. Otaki mentioned how he plays a CD.
What is that method···.
DA convert the CD (DAC-2000), convert the analog signal to a DSD file with SONY's commercial equipment ADA-7000R (ADA-7000R, a DSD-compatible AD / DA converter with a studio original rubidium clock - but by now surely external clocks are good enough not to cause jitter, so ADA is now obsolete?), and then down-convert it to a PCM sound source (wav file?) Using SBM (Super Bit Mapping) direct K--1327.
By doing this, you are freed from the conflicting composition of digital vs. analog in terms of sound quality, and you can now listen to CDs and vinyl records without distinction.
It goes without saying that CDs are no longer played directly, and surprisingly, everything is played in this way, including the playback of vinyl records (!), And Mr. Otaki says, "It's a body that doesn't accept anything else. I've done it "(laughs).
There was a comment that audio playback was "DA-AD-DA" in the 2011 New Year talk, but it was probably this playback method.
"The reason why I started listening to CDs in this way was that the sound of the waves recorded in the wav for radio broadcasting felt noisy, and when I consulted with an engineer, I was advised that converting the wav to DSD would eliminate that noisy feeling. That was the trigger.
It seems that wav's habit is that the sound of waves and the barking of insects can be heard noisily, but with DSD, that habit disappears.
For Mr. Otaki, it would be nice if the sound of DSD is a large-scale music such as a classical orchestra, but even if it has a large number of sounds, it basically does not match the music of the 4-rhythm organization. , When down-converted to wav, it seems that the sound became ideal for Mr. Otaki."
VAIO also included software called DSD Direct, which converts wav files to DSD, and DSD Direct Player, which converts CDs to DSD and plays them.
Mr. Otaki is using a commercial device prototyped by SONY for the studio, and there are only a few in the world, so I do not think that the sound of the file created with VAIO will be the same, but at least the work content You can do the same.
"The latest version of digital audio for sound creators" Keiji Kakizaki (Author) Byakuya Shobo
Dialogue: Eiichi Ohtaki x Keiji Kakizaki, from pages 170-206 of the book
Otaki: Generally, I don't really like the stereo method itself. I love things overwhelmingly. It feels like it's mixed up. I feel the heat there. If the separation is good, it's cold, or the heat is low.
After that, it is certain that the amount of information is large for the flowing and gorgeous sounds, but it sounds like something that is too much for me. I feel that the extra decoration is in the way of my liver. It's a comfortable sound for me that the liver is very straightforward and directly transmitted. i.e. passion rather than large information amounts?
Therefore, it's okay to use the 3-point principle, but then I don't think MP3 is the best, though it's an extreme theory. Strongly push the 3 points of the feature. However, if it is emphasized to that extent like MP3, the information is a little too missing, but it is not a mistake to say that it is such a direction. To put it broadly, it is. I'm sorry for the DSD group Kaki-chan. i.e. prefers MP3 equivalents over DSD for unexplainable reasons
Kakizaki: No, no....
Otaki Hey. Everything was a coincidence, but I finally met. Moreover, this is the last series of mastering on the 30th anniversary of Long Vacation, but at this time of year, he came out as if he had been waiting for it, a dream audio. The long journey from 35H is now perfectly over.
I AD with DSD and dropped the digital out to a PCM sound source with SBM direct, but I don't feel downgraded at all. Rather than the raw sound of DSD, it's my favorite sound.
The better the audio, the more information there is and the denser it becomes. In the movie, it's a cinerama. The screen is too wide and my eyes get tired. Even if there is a lot of information, my eyes will not get tired. If the standard screen isn't enough, expand it a little more.
In Kakizaki, it's Vista.
Otaki Vista. Vista is the thinned out DSD with SBM Direct. Then, this is DSD MP3.
Kakizaki: That's right. It's like compressing information.
Otaki Yes. that's interesting. This time, that's why the DSD MP3 was made into a CD.
Kakizaki: Not being able to go to SACD means that there is too much information.
Otaki: The basis of my sound is rock'n'roll four rhythm. Even if you can show 4 people at 70mm, 4 people are 4 people. But if you're just four people, the standard screen is fine, but my sound has a lot of instruments, so it's Vista. If you have a lot of characters and a lot of sounds, such as classical music and opera, I think 70mm cinerama, that is, a rich sound like DSD, is good, but Vista is enough for my sound.
Kakizaki: I want to reduce the amount of information.
In this book, Mr. Otaki denied the need for a higher format than the CD called SACD. This encouraged a Revolution in my thinking.
Some dude really wanted to emulate SBMD DAADDA, and is an Ohtaki fan. Seriously, DSD getting rid of waves and insects harshness doesn’t seem worth the trouble. ADA-7000R is a key item (as it captures analog rather than doing FIR filter digitally), and from Fujimoto’s articles we know K-1327 has better? SQ than SSMS SBM-D. We also know Sonoma has a SBM-D software option likely the same as SSMS’s. we know Nishio allowed MoFi to test out SBMD CD vs DSD and they felt difference – perhaps too exaggerated …amount of info relates to sound 'thickness' I guess! So if DSD also has softness of sound we haven’t explained it yet!
How is PCM harsh? Why? DSD gets rid of harshness? Is that only for analog capture or also includes FIR digital conversion? It seems digital is too accurate and keeps the harshness. One should record PCM audio of 'waves and insects harshness' and then convert to DSD to see if it is the same via analog and digital methods. One could try to broadcast PCM or DSD over radio and see which is less harsh. And one should see if SBMD really is better than any other conversion method out there, including Weiss Saracon. Given that Sony releases 24/44.1 hires files in this day and age, of Lang Lang, Glenn Gould, and others, many of which are implicitly SBMD derived from DSD Masters.
The content may be quite maniac, or, to put it bluntly, a completely unnecessary book for those who just like listening to music. However, a book with interesting content that is irresistible for those who like it:
In this volume, "Chapter 12 Jitter" and "Chapter 13 High Sound Quality Technology" were particularly interesting. For example, regarding "high-quality CDs" that have been elaborate on the latest trendy materials, "there is no one that objectively measures the played sound and proves that the sound quality is good," and the final conclusion is as follows. :
I think that "high-quality CD" should be seen as one of the products born from marketing strategy rather than high-quality technology
Not that all the SBMD is rubbish, but the best sound is not necessarily SBMD, it depends on source. Hong Kong cantopop using SBMD was quite hot, while Ohtaki is a proponent of SBMD in Japan, and Yo Yo Ma's Box set (among others) uses SBMD for all the CDs. SBMD remains a secret particularly in regard to its noise shaping method.
Perhaps someone has more experience with SBM Direct and can dismiss this post as a randomly written waste of space?
Attachments
Last edited:
One thing I do find fascinating is the apparently endless ability of the audiophile to find perfection in equipment so obsolescent that it is i) incredibly rare, II) incredibly expensive, iii) totally outmoded technologically, iv) unmaintainable in any sensible way and v) all the above, when combined, ensuring that any claims are entirely unprovable and seldom demonstrable.
I remember when Sony first offered their SBM converters. Subjectively, they did seem to have a noise floor a few dB quieter than most of their contemporaries but their opening gambit on price was hugely excessive. What Sony had brought to the market with SBM (Super Bit Mapping) was, essentially, what we now call noise shaping. That's it really.
I remember when Sony first offered their SBM converters. Subjectively, they did seem to have a noise floor a few dB quieter than most of their contemporaries but their opening gambit on price was hugely excessive. What Sony had brought to the market with SBM (Super Bit Mapping) was, essentially, what we now call noise shaping. That's it really.
andymok
Addicted to Fun and Learning
https://www.grammy.com/technical-guidelines
Technical Guidelines For Producers And Engineers
Recommendations for Hi-Res Music Production
https://www.grammy.com/sites/com/files/recommendations_for_hires_music_production_09_28_18_0.pdf
Ok, so that is an English translation of the Japanese site. But it is all just marketing bullshit. Eg, "How is PCM harsh? Why? DSD gets rid of harshness? Is that only for analog capture or also includes FIR digital conversion? It seems digital is too accurate and keeps the harshness." That is gold plated marketing rubbish. Please, this is ASR...
Last edited:
OP
Saidera
Senior Member
- Thread Starter
- #394
Sure it's marketing rubbish, but if this is ASR, it should be the right place to provide some objective answers to the basic question which was worded badly above: Is recording DSD from analog (analog from a PCM DAC) in most cases different or similar to simply converting PCM to DSD on a PC digitally? or: Is the purported harshness removed due to the internal processing of the DAC and DSD recorder or would the DSDD and SBMD process also do the same?
And indeed, why on earth is it that Japanese people and engineers are so attracted to DSD, even just as an 'intermediary' process to remove 'harshness' of PCM? Note that although VAIO DSD was marketing tricks-oriented, DA-AD-DA is little known but is a widely used process of Sony music labels and raises the question of whether taking a PCM CD, recording it as DSD, and then dropping it back to CD via SBMD improves sound - because surely Bernie Grundman wouldn't have done that SBMD to the 2004 Star Wars OST CD releases if it had no benefit at all? There is evidence that SBMD had no benefit: http://www.malonedigital.com/starwars.pdf
Not only is this all ridiculous, the worst thing is that it is near impossible to test, even if I share a virtualbox file of SBMD and others used various analysis software to painstakingly compare the PCM audio before and after SBMD. Just imagining it is unbelievable. Nevertheless I will share what results I can get regarding DSD-PCM-DSD and PCM-DSD-PCM processes.
If the resulting DSD can be analysed objectively - which I cannot do, we would find out more clearly whether it is the format's characteristics that create its sound. But surely the difference can be heard - that is, anybody should be able to tell in a studio environment that the DSD master and the DSD from after the SBMD (24 bit) process are different?
The PCM CD and the resulting SBMD PCM are not identical wave files (I confirmed using wavcompare jp software https://efu.jp.net/soft/wc/wc.html)
Common sense. How can the two be compared objectively?
And I think Eiichi Ohtaki was being honest about DA-AD-DA. https://en.wikipedia.org/wiki/Eiichi_Ohtaki
If not the harshness effect then at least the part about how DSD wasn't suitable to his style of pop music but SBMD was.
Last edited:
Remember, Sony is their pioneer and leader in sound: like Chord in the UK (PCM oversampling to million taps rather than oversampling PCM to octaDSD rate). If Sony says DSD is the future, these Japanese audiophiles will be as loyal to them as those Japanese warriors who were willing die/kamikaze for their country. And no, you cannot make a crappy master sound "smooth" with oversampling them to DSD unless you are comparing that DSD oversampled to NOS mode which we typically has roll-off in the audible treble frequency which is actually less harsh in itself than DSD due to treble roll-off
I answered that in post 2, I cannot understand the continuous rehashing of old and false claims.
OP
Saidera
Senior Member
- Thread Starter
- #397
I see. Thanks for the insights. I am only raising many obsolete Japanese sources since they don't seem to get enough exposure and critique from English forums. Before it was Ayataka Nishio about DSD, now it's Eiichi Ohtaki about SBMD. The reason I persevere with my questions is because I strongly believe there must be some basis for their claims, but I don't know what they are. Once I find conclusively that there is no basis I will be satisfied that both DSD and SBMD can go out with the garbage for me.
I bought many SBMD CDs over years, and continously told others that it is the best sound possible on CDs - compared to those other CD technologies uv22, K2, SBM, MQACD, HQCD whatever. Obviously it's not absolute and probably there is no answer to be found.
I used to think I knew SBMD but now I find I don't actually know.
Last edited:
- Joined
- Feb 23, 2016
- Messages
- 20,754
- Likes
- 37,580
Uh huh!
OP
Saidera
Senior Member
- Thread Starter
- #399
Could you clarify whether the now common noise shaping algorithms are on par with , or better than the Sony patented SBM (which I don't think is used anymore since no CDs credit it like they used to until the mid 2000s) which uses a loudness curve they developed themselves? I know many companies found ways around SBM to create their own variants but it is important to know whether, when a layperson is using any waveform editor available now, that the process of creating a CD uses something as 'good' as that 'SBM' (not that it is good).
- Joined
- Feb 23, 2016
- Messages
- 20,754
- Likes
- 37,580
SBM was simply noise shaping of 20 bits to 16 bits. The shape was to optimize a lower noise floor in the area where our hearing was more sensitive which is 3-5khz. So if it worked SBM would have given the ability to record and recover a -120 db signal over a 16 bit format. Can that be bettered or equaled now?
Yes, here is an FFT where I took Audacity and created a 44.1 khz 32 bit float file with 3 khz lying at -132 dbFS or the 22 bit level. Beyond what SBM would have done. I then saved it using shaped dither into a 44.1 khz 16 bit file. And you see the noise floor of Audacity's shaped dither (a common form of noise shaping in many if not most DAW software). You can see it has been recovered and that even lower levels approaching 24 bit are possible around the 4 khz region.
Triangular noise shaping can't quite do this. It would manage around -120 db FS signal recovery itself. Notice how the shaped noise floor follows the same general shape as Fletcher-Munson equal loudness curves above 1 khz.
Below is the original 32 bit float file prior to being saved as 44.1 khz 16 bit with shaped dither.
These shaped dither curves were developed very much from the same info about equal loudness curves Sony referenced when they made SBM. SBM was good for its time which is when much mastering software was 20 bit. It is old tech that has been surpassed. The shaped dither curve available for free in Audacity and much other software can do more.
So maybe the near religious fawning over SBM can slumber in peace finally.
Yes, here is an FFT where I took Audacity and created a 44.1 khz 32 bit float file with 3 khz lying at -132 dbFS or the 22 bit level. Beyond what SBM would have done. I then saved it using shaped dither into a 44.1 khz 16 bit file. And you see the noise floor of Audacity's shaped dither (a common form of noise shaping in many if not most DAW software). You can see it has been recovered and that even lower levels approaching 24 bit are possible around the 4 khz region.
Triangular noise shaping can't quite do this. It would manage around -120 db FS signal recovery itself. Notice how the shaped noise floor follows the same general shape as Fletcher-Munson equal loudness curves above 1 khz.
Below is the original 32 bit float file prior to being saved as 44.1 khz 16 bit with shaped dither.
These shaped dither curves were developed very much from the same info about equal loudness curves Sony referenced when they made SBM. SBM was good for its time which is when much mastering software was 20 bit. It is old tech that has been surpassed. The shaped dither curve available for free in Audacity and much other software can do more.
So maybe the near religious fawning over SBM can slumber in peace finally.
Last edited:
Similar threads
