• 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!

Remaining Considerations on DSD

OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
Listen to any of the Stones DSD CDs circa 2002. You will be converted. I promise.

http://theaudiophileman.com/stones-mono/

Once the master tapes were used to transfer the music, that music was then converted to DSD 2.8kHz for mastering. Why not PCM? “DSD most closely mirrors analogue in the digital world,” said Landi. “We really wanted to master in DSD, 24bit/192kHz just wasn’t close enough.” “DSD has come a long way in terms of tools, said Landi. “Bob [Ludwig] uses the Pyramix system that accommodates DSD and PCM. That has sophisticated editing tools in there. This is true DSD, incidentally, there is no inter-PCM conversion involved here.”

Here is from Bob Ludwig about preparation of 2002 re-issues from Pro Audio Review:
Facility: Gateway Mastering Mastering Engineer: Bob Ludwig
Project: Remastering 22 Rolling Stones Albums Record Label: ABKCO
Converters: EMM Labs Analog to DSD Converter, dCS 974 Digital to Digital Converter (DSD - PCM - DSD)
DSD Workstation: Sony Sonoma

PAR: Did you use much noise reduction?

BL: Very little. When there were spots that required its use, if it was possible in the analog world, I would do it there. If not, we went to the high resolution PCM world. If neither sounded that great, we just let the noise be. The only PCM in the whole series are those parts that were really necessary. Each corrected part was meticulously excerpted and edited back into the DSD master. One song had some significant sibilance problems, and I think 23 individual de-essed attacks of a word were edited back in by hand. A true pain in the neck, especially on the Sonoma DSD Workstation, but it kept the ultimate quality we were striving for.

PAR: The quality of the SACD layer is stunning. The CD layer also sounds much better than any previous releases I've heard.

BL: The downsampling from SACD to PCM used the latest Sony SBMD process, so the CD sound is a breakthrough as well. Being that everyone who buys a remastered Stones disc will have automatically have purchased a SACD disc, I hope that they will seek out a friend who owns a SACD player, hear the disc on that and get even more analog-like warmth and clarity than the CD layer affords.
 
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
What the impulse responses of DSD64 and 44.1k really look like:
wY0wzcW.png

The truth is if you want a shorter impulse response, then you need a slower DSD filter, and the result is more ultrasonic noise. If you want to get rid of the ultrasonic noise, then you need a steeper DSD filter, and results in a longer impulse response. You can't have both and that illustration is a lie and exaggerated the differences. For 44.1kHz PCM you can also get a shorter impulse response, at the expense of more imaging artifacts, so no free lunch here too. Of course SACD still has an advantage over CDDA in terms of ENOB and frequency bandwidth due to the sheer 4x data rate, but for data efficiency you only need about 3x the CDDA data rate to exceed what DSD64 can do, for example, 24-bit/88.2kHz, 22-bit/96kHz and so on.

So we want the shortest impulse response. We're ok with a slower DSD filter and more ultrasonic noise. So is the illustration a lie? Probably ultrasonic noise is to be preferred over imaging artifacts.
 
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
And that is simply because it is possible to go from DSD to PCM and back again a dozen times without it being noticeable. The fact is that, these days, the potential of both techniques is so far beyond our threshold of hearing that it really doesn't matter what you use.

But audiopiles tend to have a penchant for the obscure, the difficult, the complex and the expensive.

We will test DSD to PCM and back again a dozen times, unless it's already been done.
 
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
Something sound different is not a proof of whether PCM or DSD is better, since it is perfectly possible to have two CDDA releases with obvious differences as well.

Most of the DSD spectrum contains noise at a very high level. Any signal content above a few tens of kHz is completely drowned out. If anything, DSD is proof that ultrasonic content is completely irrelevant and thus that high sample rates are unnecessary.

DSD is 1-bit PCM with the quantisation noise pushed into ultrasonic frequencies.

It's all 1-bit PCM, just with or without noise shaping.
If something you want is only available as DSD, that's of course no reason not to buy it. It's not quite that terrible a format. What you shouldn't do is pay extra for DSD when the same music is available cheaper on another format. The most egregious pricing is probably that of Cookie Marenco who asks $15 for CD quality ($40 for a physical disc) and $50 for DSD256.
The precise pattern of bit values _is_ the noise shaping. Since it's not possible to completely fill a section of the spectrum with noise, there is indeed some dynamic range available at higher frequencies. In practice, that's not of much use, though, since the spectral content of music drops off as the frequency rises while the noise level increases. It doesn't take long for the noise to completely swamp whatever signal remains.
To elaborate on that, if you were to split the sample stream into blocks of 64 (assuming that is the oversampling factor) and collect all the 1 bits at the start of each such block, you'd get basically a PWM signal quantised (in time) with 6-bit precision. DSD can, if one is thus inclined, be seen as oversampled PWM quantised to two widths, zero and 100%, with noise shaping. Not that this interpretation is particularly useful. PCM is much more efficient since it gives each bit position within a block a different meaning, thus allowing the various combinations to express a wider range of values.
That is irrelevant. The great thing about science is that the results of a well-conducted experiment are valid for everybody, even if those who were not there. The limitations of human hearing are well-established, and 24-bit 48 kHz PCM is more than enough to capture everything anyone could possibly perceive. CD quality is cutting it quite close, and there may be corner cases where it isn't fully transparent, though I'd be surprised to find one in an acoustic recording.
It may be possible to hear a difference between a particular DSD implementation and a particular PCM implementation, just as it may be possible to hear a difference between two implementations of the same format. Studies suggesting audible differences between better-than-CD formats are invariably set up such that it's impossible to say with any certainty that the differences heard were not actually caused by something other than what is claimed.
If the contention is that DSD is superior and PCM detracts something, start with a DSD recording of your choosing. Convert it to 24-bit 48 kHz PCM and back to DSD using good software. Compare the resulting DSD file to the original using the same playback chain. If PCM is as horrid as some say, the difference should be readily audible.
Sure, a few good papers have made their way there. Still, the bulk of AES contributions come from people in the hi-fi industry whose livelihoods depend on the myth being perpetuated that some nebulous differences exist outside the ordinarily audible range. Even a solid result in favour of one inaudible thing over another could ruin business for the half of the industry touting the losing option. It is in the collective interest of the industry as a whole to keep cranking out poorly executed studies that hint at one thing or another without ever reaching a firm conclusion.
Different ADCs and DACs with associated analogue circuitry, for starters. Then there's whatever signal path additions are needed to match levels. On top of that, there's the possibility of downstream side effects of the residual ultrasonic noise after the DAC, which will differ. If you hear a difference, you have no way of knowing what actually caused it. It could be the result of a poorly executed conversion rather than an actual deficiency in the alternate format.
Too much misinformation. I'm speechless now.... To really tell the whole thing it will take a long long writing just to cover the surface. And I don't even know the whole thing.
Instead in all these pages, no one asked someone to do a measurement of the SAID equipment under the SAID condition to try to reproduce it?
We need this sort of common sense to prevail.
 
Last edited:

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
index.php

DXD is 24bit 352.8kHz. It's a lot like DSD.
The lie is the DSD impulse in the illustration has no noise, and the "Rel. level" as labeled in the illustration is much higher than all other PCM formats. It is 100% false and intentionally misleading.
https://www.audiosciencereview.com/...ds/the-sound-quality-of-dsd.14773/post-538904

So we want the shortest impulse response. We're ok with a slower DSD filter and more ultrasonic noise. So is the illustration a lie? Probably ultrasonic noise is to be preferred over imaging artifacts.
Even the lowest form of DSD (64) has 4x CDDA data rate and it is not a fair comparison. 24-bit/88.2kHz or 22-bit/96kHz PCM, as mentioned above, use no more than 3x CDDA bitrate and have better impulse response AND ultrasonic AND non-ultrasonic noise over DSD64:

https://www.audiosciencereview.com/forum/index.php?threads/digital-filter-game.23795/post-810355

Even with DSD, and especially DSD64, a relatively steep digital filter is still required to cut the noise. Also, the key issue is that with DSD, even when the signal only contain digital silence, the noise is still always present, but it is not the case with PCM. Download the Mahler Symphony example in the link above, the ultrasonic noise remain strong even in quiet portions of the music.

If one wants to have an apple to apple comparison, use DSD at the same data rate as CDDA, and that means 705.6kHz DSD16. PCM-DSD_Converter supports this format:
https://pcmdsd.com/Software/PCM-DSD_Converter_en.html
 
Last edited:
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
Thank you for further clarifying and proving beyond reasonable doubt one of the many misconceptions which I am sure still dupes the gullible out there!

DSD is marketing. So they're using the wrong units? https://www.merging.com/uploads/assets/Pyramix/dsdresponse_big.png

https://i109.photobucket.com/albums/n51/argentino_sf/DSD/Impulse_response_comparison_zpsd1e40551.jpg

Does Rel. level = amplitude in Reefman's diagram? He says the amplitude is 0.25 for all three?

If you convert 3μs to Hz, you can see how incredulous it is? 3 µs(p) = 333333.33333333 Hz? Or is that conversion wrong?

It was interesting that when discussing whether DSD or PCM is the better sounding format, Mr. Watt (CHORD) commented that PCM is by far the better format, with more resolution, spatial depth and separation than DSD. Although the sample rate is lower, the timing information can be made more accurate with better interpolation. On the other hand, DSD discards small signals due to the way it works, so minute information is easily lost.

DSD64 (or PCM (192Khz, 24Bit) conversion files with DSD as the source) playback contains abnormal high frequency noise components. This high frequency noise is in the audio amplifier frequency range, consumes unnecessary power, and may also cause intermodulation distortion. Worse, it may lead to damage of the tweeter. If you want to convert your records to high-resolution (Hi-Res), it is better to record in PCM because PCM files do not have the abnormal high frequency noise found in DSD and you can edit the waveform of the record to remove noise.

It's true that major music manufacturers can produce SACD without DSD format conversion, but it's a difficult situation for ordinary music production companies to maintain the DSD format all the way through the process.

There was once an interesting experiment in the US by an Israeli company called WAVES. They recorded a duo of piano and violin, one without any noise and the other with noise (a gate that opens when the sound is played, not all the time), and asked a lot of listeners which one they preferred. The result was that over 70% of the listeners chose the one with noise. Many people said that the sound was natural, realistic, etc. The other 30% said they couldn't tell the difference. This dithering noise has the effect of lowering the S/N, but softening the bandwidth peaks. From this result, WAVES has evolved its own technology in terms of how to mix the noise in a way that it is not noticeable when dithering, and has become the king of professional plug-ins. If you put it in the low frequency range, the mid and high frequency range will stand out, and if you put it in the mid frequency range, the high frequency range will sound extended. The sound is also changed by cables due to the generation of band noise (electromagnetic noise) caused by the cable structure. It can be measured with a precise measuring instrument. DSD is no exception, it just happens to be a noise that people find pleasant, but technically it is a degradation during playback (compared to the original waveform). In conclusion, DSD is clearly inferior to PCM, but just as there are people who prefer analog playback to digital, you can't determine the standard of comfort from the S/N and the data alone.

Another hypothesis here. The more realistic the recorded sound is, the more uncomfortable we feel the discrepancy with the real sound. It's just like how we feel uncomfortable when we see realistic wax figures or realistic CG. When noise is added to the sound, the sound becomes a little more distant from reality, but since it creates a sense of distance from the real sound, it becomes objectively perceived as quasi-real. The brain does not consider the sound with the noise to be real, so it can listen to it without confusion and without worry. In other words, the brain perceives sounds with a little noise as more real. The overtone load caused by cables is a good example of this. It's similar to how a wax doll or CG becomes deformed and becomes a cute character. Because it's obviously not real, you can accept it without confusion and think it's cute (e.g. SAPARi Cat avatar logo used on ASR). That's why, even if you try to make it sound more like real sound by using S/N, low distortion, etc your brain will look for things in the sound that cause discomfort (=not real). The more analytical the sound, the more the quirks of microphones and recording equipment (sounds that are not real) and sounds that are not being picked up (atmosphere) stand out. If the age of perfect recording and playback comes in the future, our brains will be tricked into recognizing the sound as real, and a new audio age will open up (although this may not be possible as long as we use a reproduction device like a speaker).

Q&A with Thorsten Loesch about PCM vs DSD Notes

When you make a recording, you convert it to DSD or PCM, but in both cases you lose some of the quality that only the original recording has. If a 352.8kHz/24bit DXD-PCM signal is converted to 1bit/2.822MHz DSD, about 99.96% of the amplitude information possible in PCM format must be discarded, while only 12.5% of the time domain information possible in DSD can be obtained. If you want to go from DSD to DXD, you will need to use a different format. If you are converting from DSD to DXD, i.e., from 1 bit / 2.822 MHz to a 352.8 kHz / 24 bit signal, you will have to discard 87.5% of the time domain information of the DSD source. Although it is theoretically possible to remap them all to the amplitude domain. So, in practice, you end up with the worst of both formats rather than the best.

These ADC/DAC components were generally developed for the PCM-centric market. There, the entire recording/editing/mastering/release process is done in PCM, ADCs generally output in PCM, and DACs are designed for PCM signal input. And they tend to be optimized for this operation.

DSD tends to be a supplemental feature that is added after the fact, often to "fit the trend". Many of these DACs, for example, have a complete PCM audio path with digital filtering and digital volume control. Inside, the DSD is first converted to PCM, digitally filtered (which adds all the negative effects of the PCM to DSD data stream conversion), and then finally subjected to a multi-bit Delta Sigma conversion. This means that the two undesirable conversion processes take place at the heart of a black box called the DAC chip. If you really want to get the best sound out of DSD, you must first convert it to PCM, then process it as PCM signal data and play it back as multi-bit Delta Sigma, which is actually almost the same as converting DSD directly to PCM and then playing it back as PCM. This is what happens in many products called "DSD-capable DACs".

The differences in sound that we can hear from converters that support conversion between PCM and DSD are strictly due to the conversion algorithm, and not the format itself. Compared to the original untouched DSD source, signal loss is unavoidable. Ideally, a true multi-bit DAC should be used to play PCM as PCM (without the losses associated with one stage of processing, whatever the original AD converted source was) and vice versa. Most of the so called latest flagship DACs mess up both PCM and DSD.

Daniel Weiss' White Paper on DSD Notes

1-bit sigma-delta A/D and D/A converters were very popular in their day, but converter technology has advanced. If you want to convert it to PCM in the middle of your work or at the end, you don't need to record in DSD. It makes a lot more sense to record at 88.2 kHz/24-bit instead of DSD.

DSD has been converted to PCM for editing and other purposes, so why bother listening to DSD at all? I'm not sure. However, if digital amplifiers become mainstream in the future, there is no format more convenient than DSD for realizing unconverted audio, from recording to the final amplifier. Sharp's 1-bit amplifier is like the amplifier part that also serves as the DAC. It is said that the DSD remains until just before the speaker.
http://www.sharp.co.jp/sc/eihon/smsx1/text/1bi.html
Digital amplifiers can be roughly divided into PDM and PWM. The PDM (Sharp) method is the only one that allows DSD signals to flow directly. Sharp's 1-bit amplifier had poor measured values such as distortion ratio. PDM amplifiers are not as efficient as PWM amplifiers. They have the disadvantage that they don't use the voltage of the power supply as efficiently. However, it is a natural output form of a 1-bit A/D converter. Yet if Sharp is sending DSD input data directly to the speaker, the block diagram shows that the DSD input should go to the digital driver circuit without going through the 1-bit signal generator circuit, and the feedback should be disconnected. The volume can't be adjusted except by changing the power supply voltage of the power switching circuit.

There are not many posts about sound quality in most DSD threads. However, once you listen to DSD native playback sound with clear characteristics from a 'proper' device, the opinions of the people who post in the thread will probably change. Hasn't happened yet!

Processing that can be done without PCM conversion in AudioGate is Cut & Merge, Fade in Fade out (of course the target part is converted to PCM). The processes that require PCM conversion are gain, DC cut, and LR balance adjustment. Processing that can be done without PCM conversion (including multi-bit ⊿Σ PDM) in SONOMA and current Pyramix is cut & merge, fade-in fade-out etc

The most common ADC/DACs are all multi-bit ∆Σ PDM, so most DSDs look like this
  • ADC: Analog input -> Multi-bit ⊿ΣPDM -> LPF -> PCM -> ⊿Σ -> DSD
  • Editing: DSD → decimation (LPF) → PCM → editing → interpolation → ⊿Σ → DSD
  • DAC: DSD → decimation (LPF) → PCM → (digital processing) → interpolation → ⊿Σ → multi-bit ⊿ΣPDM → DA section
Some people don't want to convert to DSD when editing, so they record once and use DAC-(analog processing)-ADC. However, the Waseda ADC does not use multi-bit ∆Σ PDM, so 1-bit Audio Consortium samples are DSD native in the true sense of the word, as long as the DAC is a DSD direct path.

1bit Audio Consortium http://1bitcons.acoust.ias.sci.waseda.ac.jp/

Sound quality is determined more by the recording and editing environment than the format. As you can imagine, compressed sound sources have poor sound quality. But for uncompressed sound sources, the difference between formats is too small to be of much concern.

Trust us, you can't tell the difference in any quick and simple blind test!
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
DSD is marketing. So they're using the wrong units? https://www.merging.com/uploads/assets/Pyramix/dsdresponse_big.png

https://i109.photobucket.com/albums/n51/argentino_sf/DSD/Impulse_response_comparison_zpsd1e40551.jpg

Does Rel. level = amplitude in Reefman's diagram? He says the amplitude is 0.25 for all three?

If you convert 3μs to Hz, you can see how incredulous it is? 3 µs(p) = 333333.33333333 Hz? Or is that conversion wrong?
In case you did not read carefully, I attached two DSD impulse files in one of the links in my previous reply. Follow the instructions carefully and you can 100% reproduce what I posted, so there is no false illustration, just fact. Don't just read, try to process the files yourself, because the whole idea of my thread is to encourage people to actually try these things out.

https://www.audiosciencereview.com/forum/index.php?threads/digital-filter-game.23795/post-810355

The first post in the thread above also involves PCM filter experiments, again, try to process the files yourself.
 
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
Those comments were only in response to the diagrams DSD supporters often use. I do not understand even the relationship between μs and Hz, or amplitude for that matter.

I did not immediately take up your guidance, thinking that I could come back after checking some texts on the basics. Once I do that I will attempt it and hopefully truly understand all about this.

After going through each post, reading only, not yet trying it out, I can instinctively grasp the fact regarding impulse responses. Before learning or trying it out though, I still do not get why Reefman wrote that the amplitude was equal for all three. Obviously it may have been, but the sizing of the impulses has probably been tweaked. I am just guessing. I also don’t get why DSD gets a black left screen until it gets filtered. I also hope that after looking into this area in detail, I will be able to interpret and understand those graphs that are shown in Cirrus Logic datasheets.

Thank you bennetng!

Oh man! How come I never revisited that site! https://pcmdsd.com/Software/PCM-DSD_Converter.html
I actually looked down upon that converter, I thought it was 'just another converter', but look at those graphs! This dude has compared every PCM to DSD converter on the planet! Like, I've got all of them except Saracon and XLD! Explained here: https://forum.audiophile.jp/t/目で見て楽しむpcm-dsd変換/19
And, he even mentions Mr Mans Rullgard, and my interest in getting to the bottom of all this, including FIR filters, the number 5 and 7, etc. has skyrocketed.
 
Last edited:
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
One wonders whether Sonoma+K-1326, or DAC-2000, ADA-7000R etc measure better than the standard DSD players today.

The extremes of audio - DSD = to do basically nothing to the signal (a lie) versus DSEE HX = to do as much DSP as possible to recreate what is lost and even 'remaster' or 'pad/edit with generic sound' (also a lie).

Both prove that Hires is a term to stay away from (for consumers anyway).

Here in this graph it seems CD trumps SACD:
7b09CDSACDnoislevel2.jpg
 
Last edited:
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
Even for re-releases of remastered material, Sony's bigger projects tend to search for newer master tapes, and NOT the Sony Music DSD archives.

Here's an example. I've already introduced it before too: https://www.audiosciencereview.com/...ound-quality-of-dsd.14773/page-20#post-516511

From the dialogue between Eiichi Otaki and Kageji Kakizaki, pp. 170-206 in Keiji Kakizaki (Author), "All you need to know about digital audio for sound creators", Byakuya Shobo 2011.
Otaki: Generally speaking, I don't even like the stereo system itself that much. I'm overwhelmingly fond of mono. It's the feeling of being mixed together. I feel the heat in that. If it's well-separated, it's kind of cold, or it's not very hot. The other thing is that although there is certainly a lot of information in a flowing or gorgeous sound, to me it sounds like too much is too little. I feel like the unnecessary embellishments are hindering the main point. For me, the most comfortable sound is one that conveys the heart of the matter very straightforwardly and directly. So, while the 3-point principle is fine, I don't think that MP3s are the best choice. However, if you emphasize them so much like MP3, it's a little too much information, but it's not wrong to say that it's like that in terms of direction. I'm sorry to say this to Kaki-chan who is a DSD fan.
Kakizaki: No, no.
Otaki: The long journey from the 35H is now complete. I'm using DSD AD, and then dropping the digital out to PCM using SBM direct, but it doesn't feel like a downgrade to me at all. In fact, the sound is more to my liking than the raw DSD sound.
The better the audio, the more information you get, and the more dense it becomes. In movies, it's like Cinerama. The screen is too wide and your eyes get tired. Even if there is a lot of information, your eyes will get tired. When a standard screen is not enough, you need to widen it a little more.
Kakizaki Then, Vista.
Otaki: So Vista is the result of thinning out DSD with SBM direct. So this is a DSD MP3, right?
Kakizaki: That's right. It's like compressing information, isn't it?
Otaki That's right. That's interesting. So this time, you've converted a DSD MP3 to CD.
Kakizaki: If you can't go to SACD, that means there is too much information.
Otaki: The basis of my sound is rock and roll folklore, and even if you show four people in 70mm, four people are four people. But if it was just four people, a standard screen would be fine, but my sound has a lot of instruments, so I decided to use Vista. If it's classical music or opera with a lot of characters and a lot of sounds, I think it would be better to use 70mm Cinerama, or in other words, richer sound like DSD, but for my sound, Vista is good enough.
Kakizaki: So you want to reduce the amount of information?
In this book, Mr. Otaki denies the necessity of SACD, a format higher than CD.

That was the 30th Anniversary release in 2011. Before, it was in 2001 and the 80s and 90s. This year in 2021 is the 40th Anniversary release.

Naitoh Tetsuya explains the history behind the 2021 40th Anniversary release.

https://story.mora-qualitas.com/story/477
https://story.mora-qualitas.com/story/480

For the 20th anniversary, the master sound source was digital, right?
That's right. In short, the original analog master tape had fallen apart and was not in a usable condition. I used the 3/4 U-matic digital tape, converted it back to analog using the DAC-2000 (a professional DA converter developed by Sony), and then converted it back to digital and maximized it. I was able to complete the project. But that was the end result, and it took me quite a while to arrive at that method.
I see. In the end, how long did it take you to remaster the 20th anniversary?
I think it took almost a year ...... more than half a year.
Do you ever spend that much time on a single remaster?
Not a single one (laughs).
That includes the time to choose the equipment?
That's right. Then there is the master tape selection.
To begin with, (his obsession) was not about specs. Even though he was very particular about his equipment, he wasn't so particular about the specifications. It was just a matter of whether or not the sound was to his liking. In that sense, I think he was particular about 44.1khz, and I think that's the reason why he didn't go in the direction of high sampling. I tried everything, DSD, 96kHz, everything, but in the end I think 44.1kHz digital was the best for me.
So, you were satisfied with the 20th anniversary remastering at that point, right?
I think so. If I hadn't been satisfied, I wouldn't have been able to finish the work. I remember that we were working until the very last minute. I remember I was working until the very last minute, until I had to deliver the product to the factory.
I've heard that remastering generally doesn't take that much effort, but did you feel that it was a natural part of Mr. Otaki's project?
Well, there was a sense of resignation on everyone's part (laughs). They said, "It can't be helped. That included our factory as well. We would have them do a lot of test cutting, and then send it back to us and repeat the process at .......
Mr. Otaki is outstanding in that he is uncompromising.
I think he is outstanding.
Since equipment evolves over a period of ten years, it is likely that the sound will change even if it is the same master. Is it safe to assume that the sound will get better with newer equipment?
Well, "better" doesn't mean higher specs. For the 30th anniversary, I completely changed the equipment I used, but it was all about how close I could get to Otaki-san's ideal sound.
I heard that for the 30th anniversary you used an analog master (second generation) made in 1984.
Yes, for the 20th anniversary, the theme was to go loud, but for the 30th anniversary, I had a policy of going as natural as possible and not applying any makeup.
Originally, Mr. Otaki said that the 30th anniversary remastering "feels similar to the atmosphere when listening to a lacquer disc.
Even if you say it's a good analog version, it won't be like that even if you print it as it is, unless you put some effort into it, so it needs some kind of direction or correction. In that sense, for the 30th anniversary edition, I used the ADA-7000R (professional AD/DA converter with rubidium clock) to convert the analog signal to DSD signal, and then converted it back to PCM signal using the SBM direct method.
Was that the first time you used DSD?
No, I had used DSD before that, but he wasn't satisfied with the sound of DSD alone, so I converted it back to PCM, and he was hooked. It's a lot of work, though, because I have to use almost every piece of equipment to try it all out (laughs).
While making the best use of the original analog sound, you also think about how to make it up so that it sounds the best on today's media.
That's right. I was only thinking about CDs for a long time.
So you are returning to the theme of how to reproduce analog sound naturally on CDs.
I think I was always thinking about how to improve the sound of CDs. The sound of a lacquer disc is ideal.
If you look at Mr. Otaki's words at the 30th anniversary, he said, "There's nothing more to do.
That's how thorough we were. I switched masters back and forth, and tried out all kinds of combinations of equipment. It took me more than half a year to do that.
At the same time, he also said that we don't need SACD because it has too much information.
Did he say that?
Yes, he did. So, does that mean we don't need high-resolution either?
I can't say that. It's not a matter of specs. It's a matter of whether the sound quality is right.
I see. At that time, Mr. Otaki said that he had done everything he could do and there was nothing more to do. ......
Hahaha!
I'd decided to do the 40th anniversary remaster. I participated in the 20th and 30th anniversaries, so I had a base to work from, which made things a little easier, but the biggest problem this time was the absence of Mr. Otaki (laughs). I was working on it, and I don't know if this is good or bad .......
There is no one to make the final judgment.
That was the most uncomfortable part of the project. However, when it comes to the theme, I still have some regrets about the 20th anniversary and the 30th anniversary. ......
What do you mean by "regrets"?
When I listen to the 20th anniversary album now, I feel like I'm forcing the level of the music into some parts. Also, because of the equipment used at the time, the sound was a bit harsh.
It has a loud sound, doesn't it?
Yes, the 30th anniversary album was a little too plain (laughs). It was well received by professionals, but at first listen, the 20th anniversary album was more popular. So, when commercial producers listen to the sound source used in commercials, they 100% choose the 20th anniversary version. I was wondering if it would be possible to integrate these two. I was wondering if I could make it a little brighter with the silky feeling that I think is unique to Niagara. That was the theme of this project. I mean, that's all I could think of.
We had always talked about releasing the 40th anniversary album in all the media we could. The only problem was that if we used the same system (via DSD) as the 30th anniversary album, we could only do 44.1 kHz. That's why we decided to go PCM-only without DSD this time. In other words, we went back to the 20th anniversary album.
I see. What kind of innovations did you make from there?
The next step was to choose a converter to convert from analog to PCM. But nowadays, in the DAW environment, there are very few professional AD converters. I digitally converted the data with several models and had Niagara's producers listen to them, and finally decided on the current model for AD conversion.
I was thinking about the kind of sound that Mr. Otaki would like. Of course, it's always in the back of my mind, thinking about whether he's angry or not (laughs)!

This means that the dCS DSD to 96kHz converter was excluded, so DSD was a pox for this project.

Do you change the mastering for high-resolution and CD?
No, we don't really change it much. I didn't change anything numerically, but I wanted high-resolution to sound like high-resolution. It's not a pinch of salt, but when I digitized at 96 kHz, I lowered the level by about 0.5 dB, not so much that I changed it, but I changed it a little.
What happens when you lower the level?
I thought that if I lowered the level, it would make the sound a little less "choked" and give it a wider feel. I wanted it to sound like high-resolution.
To be honest, I couldn't really tell the difference in my home playback environment, but I had the impression that the high-resolution version had a little more depth and breadth.
Yes, I think that's the extent of the difference. I hope you enjoy the subtle differences.

Hires is just no good value...

They went full blast and released every fan material, every studio tape, promotion recordings, a cassette tape, vinyl records cut by 2 studios, 4 CDs, 1 Blu-Ray, streaming etc. And ORICON rankings gave it a No 1 place.

Although Otaki certainly preferred vinyl sound, DSD didn't suit his music, and SBMD CD was the best they could do. For this project, the emphasis has been on CD, but with vinyl, the original purpose may have been achieved.

In terms of a project that really characterises Sony Music Studios, Otaki's Long Vacation is the signature album of Sony Music JP. From the 80s onwards this album has continuously been selling well. It's inconceivable why. And now in 360 Reality Audio as well.
 

Victor Martell

Active Member
Joined
Mar 9, 2018
Messages
178
Likes
217
Listen to any of the Stones DSD CDs circa 2002. You will be converted. I promise.

I think you just hit it right in the head - this thread have gone in the usual ASR mode - highly technical, both sides of the argument throwing more and more technical facts as proof... I say let's change tacks... As I member of ASR I fully understand the implications here - all that tech is running against the simply fact of what the human hearing is... and we know well its limitations and parameters... so that's my initial declaration - I fully declare myself unable to hear or prove that DSD is better, or not, or the same. Just being, you know, objective...

BUT

As a mostly classical fan still keep buying SACDs - where the format is very much alive and well - why? well - am I am not talking ultrasonics or stuff that we know are not human perceptible - basically is that those companies are great at recording and mastering so their recordings sound fantastic - their SACDs, but I suspect that if put on CD or Bluray Audio, they would be equally great.

Recording and mastering: that one thing that IMHO matters. So yes, although I haven't heard them, @Geoffkait is probably right that those Stones SACDs are fantastic. Not because of SACD per se, but because, knowing it was gonna be a SACDs release, a great, better done mastering was used.

Think about the 30 Anniversary DSOTM SACD - it's deemed as great but its CD layer is deemed as one of the worst - why? not because of the format, but because the CD layer used the brickwalled mastering used for, well, CDs, whereas the SACD layer had a different non-brickwalled mastering. Similar situation with the BMG Living Stereo hybrid SACDs - the CD layer used the same mastering as the old mid 90s CDs, instead of converting the new master used for the SACD layer. You were hearing different masters and depending on your taste, when comparing, it could lead you to the conclusion that SACD was better.

Analogue Productions, uses the same masters for both layers of their hybrid SACDs; they just convert them. And with the usual caveats - this it just my impression, with my hearing, etc. - it shows; it's obvious it is the same mastering, therefore there seems to be no difference between the SACD and CD layers... to my ears.

Just my $0.02 pesos...
v
 

andymok

Addicted to Fun and Learning
Joined
Sep 14, 2018
Messages
562
Likes
553
Location
Hong Kong
It's all about bandwidth and dynamics, also gives me a hint why DPA bothers to make high-voltage microphones

The slew rate of the PGA4311 and PGA2310 is high enough such that it has no measurable effect on linear audio performance. There is a common misconception that faster slew rate means less high frequency distortion in linear operation. Please keep in mind that a slew rate-limited response only occurs during completely non-linear operation of the amplifier and is not an indicator of how linear an op amp (specifically its input stage) is below the slew rate.

Consider that for a bandwidth limited system, the maximum bandwidth of the system also determines the fastest rise time it can have. For example, an audio system with a DAC sampling at 192 kHz has a maximum bandwidth of 96kHz. The relationship between bandwidth and rise time is: T(rise) = 0.35 / Bandwidth. So for a system with a 96kHz bandwidth, the fastest rise time possible is 0.35 / 96kHz = 3.646 microseconds. If the op amp is powered from +/-15V supplies, and had to step all the way from the negative supply to the positive supply (neglecting output voltage swing limitations) the fastest rate of change possible is this system is 30V / 3.646 microseconds = 8.23 Volts per microsecond. This is the fastest rate of change possible, limited by the system bandwidth and power supplies, not the op amp slew rate. This analysis also doesn’t include the bandwidth limitation of the human ear.

Despite the common misconception, most high frequency distortion in op amps is really determined by the op amp output stage linearity. This is a constant error source in the feedback loop of the op amp, and as the loop gain decreases this distortion rises above the noise floor of the measurement.
https://e2e.ti.com/support/audio-gr.../slew-rate-of-pga4311-pga2310/1667486#1667486
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,493
I think you just hit it right in the head - this thread have gone in the usual ASR mode - highly technical, both sides of the argument throwing more and more technical facts as proof... I say let's change tacks... As I member of ASR I fully understand the implications here - all that tech is running against the simply fact of what the human hearing is... and we know well its limitations and parameters... so that's my initial declaration - I fully declare myself unable to hear or prove that DSD is better, or not, or the same. Just being, you know, objective...

BUT

As a mostly classical fan still keep buying SACDs - where the format is very much alive and well - why? well - am I am not talking ultrasonics or stuff that we know are not human perceptible - basically is that those companies are great at recording and mastering so their recordings sound fantastic - their SACDs, but I suspect that if put on CD or Bluray Audio, they would be equally great.

Recording and mastering: that one thing that IMHO matters. So yes, although I haven't heard them, @Geoffkait is probably right that those Stones SACDs are fantastic. Not because of SACD per se, but because, knowing it was gonna be a SACDs release, a great, better done mastering was used.

Think about the 30 Anniversary DSOTM SACD - it's deemed as great but its CD layer is deemed as one of the worst - why? not because of the format, but because the CD layer used the brickwalled mastering used for, well, CDs, whereas the SACD layer had a different non-brickwalled mastering. Similar situation with the BMG Living Stereo hybrid SACDs - the CD layer used the same mastering as the old mid 90s CDs, instead of converting the new master used for the SACD layer. You were hearing different masters and depending on your taste, when comparing, it could lead you to the conclusion that SACD was better.

Analogue Productions, uses the same masters for both layers of their hybrid SACDs; they just convert them. And with the usual caveats - this it just my impression, with my hearing, etc. - it shows; it's obvious it is the same mastering, therefore there seems to be no difference between the SACD and CD layers... to my ears.

Just my $0.02 pesos...
v

Absolutely, very wise and spot-on. It is very unlikely that there's an audible difference between the formats - and even if there is, it's swamped by other factors, including the ones you cite:
  • Extra care during mastering: As you note, many SACDs have been mastered with more dynamics and less compression than a typical mass-market CD might be. So a lot of "SACD sounds great" commentary is about the good mastering, and in many cases folks haven't even bothered to compare it with the CD layer of the same disc - and of course if it's a single-layer disc they don't even have a CD version to compare it with, so there is no way in that case to differentiate between the format and the mastering.
  • Different mastering: This one is so frequent and it always surprises me (although I guess it shouldn't) how many audiophiles don't even bother or care to find out if the SACD and CD layers even have the same mastering. Another example to add to the ones you mention is the 2003 Talk Talk hybrid SACDs (Colour of Spring and Spirit of Eden). The SACD layers contain unique masterings that to my knowledge have never appeared anywhere else, while the redbook layers contain the standard 1997 Phill Brown mastering - not a bad mastering, but inferior to the SACD layer mastering.
  • Different mastered volume: This is infuriating - the CD layer contains the same mastering but compressed and/or brickwalled - or in some cases just flat-out clipped. I suppose it's because the SACD layer is supposed to be the "audiophile" layer and perhaps they want people to easily hear a different, superior sound from the SACD layer. Or perhaps it's simply because SACD has a different relationship with digital gain and some mastering engineers or disc-authoring techs don't adjust for that?
  • Different playback volume: Given SACD's relationship to digital gain, I have my doubts that if you simply switch between the SACD and CD layers on a disc, that you'll get playback at the exact same volume with each layer. Since you're playing both layers on the same equipment in the same room at the same time, it is of course natural to assume any audible difference must be because SACD sounds different than CD. But it could just be a level shift because of how your player plays each layer.
 
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
Great insights. Putting the imperfections and questionable technical details of DSD to one side, one can really give up and just enjoy the music. There are too many factors when it comes to consumption of commercially released audio. I definitely value likely unedited DSD releases.

On the other hand, if my intentions have not become absolutely clear yet, if we push commercially released audio to one side, and focus solely on DSD as a straightforward recording format, as an intermediary for PCM, and as a final upconversion before playback, then things have to get highly technical. The surface matters that cling to DSD, such as SACDs, hearing limits, implementation differences, mastering and so on must be brushed aside.

I noticed that Sony's Tokyo studios have expunged K-1327 from their gear list pdf, and overall ceased to use the DSD format as an intermediary for PCM, unless there is a small budget.

Of course, there's no real end to this. To amplify DSD signals straight to the sound you hear seems easily said, but sound quality is poor that way.

The technical reasons why PCM will better record the sound are swamped by ever bigger ideas such as using radar technology or lasers, a full wall of mics or speakers, and other mind-boggling efforts to capture sound in a space. Never mind that microphones aren't our ears. Much of the reasons for using DSD as an intermediary for PCM likely come from equipment sound characteristics and degree of mastering. It's difficult to give credit to DSD for the engineers' effort in DSD-based implementations and the inherent inability to edit. As for upconversion before playback, this depends on the system used, but in nearly all cases it is unnecessary. For some there may be a placebo effect, but some objective measurements should dispel that.

It's definitely important to understand where we're at in terms of progress in the audio industry. Is there a better way of doing things?
 

Victor Martell

Active Member
Joined
Mar 9, 2018
Messages
178
Likes
217
Great insights. Putting the imperfections and questionable technical details of DSD to one side, one can really give up and just enjoy the music. There are too many factors when it comes to consumption of commercially released audio. I definitely value likely unedited DSD releases.

On the other hand, if my intentions have not become absolutely clear yet, if we push commercially released audio to one side, and focus solely on DSD as a straightforward recording format, as an intermediary for PCM, and as a final upconversion before playback, then things have to get highly technical. The surface matters that cling to DSD, such as SACDs, hearing limits, implementation differences, mastering and so on must be brushed aside.

I noticed that Sony's Tokyo studios have expunged K-1327 from their gear list pdf, and overall ceased to use the DSD format as an intermediary for PCM, unless there is a small budget.

Of course, there's no real end to this. To amplify DSD signals straight to the sound you hear seems easily said, but sound quality is poor that way.

The technical reasons why PCM will better record the sound are swamped by ever bigger ideas such as using radar technology or lasers, a full wall of mics or speakers, and other mind-boggling efforts to capture sound in a space. Never mind that microphones aren't our ears. Much of the reasons for using DSD as an intermediary for PCM likely come from equipment sound characteristics and degree of mastering. It's difficult to give credit to DSD for the engineers' effort in DSD-based implementations and the inherent inability to edit. As for upconversion before playback, this depends on the system used, but in nearly all cases it is unnecessary. For some there may be a placebo effect, but some objective measurements should dispel that.

It's definitely important to understand where we're at in terms of progress in the audio industry. Is there a better way of doing things?

Totally understand - it does seem that our positions are complementary rather than in opposition in the true sense. Don't disagree with your points. I've heard some (mostly classical) labels address the issues you mention by recording in DSD, then mastering in analog, then cutting LPs from that. Whatever one thinks of the current love for the vinyl format, at least that seems more consistent, a least conceptually, doesn't it?

All moot in my case - although I like vinyl, decided that the best vehicle for Classical Music is digital. Period.
YMMV, of course - is just a preference.

v
 
OP
Saidera

Saidera

Senior Member
Joined
Jul 18, 2020
Messages
388
Likes
309
Location
Adelaide, South Australia, Australia
I used to be even simpler. DSD was a belief system easily accepted due to the prevalence of badly designed PCM systems. Somehow DSD would guarantee good performance (clearly NOT the case as proven time and again on ASR). Combined with a reluctance to DSP and optimise sound, my infatuation with DSD would have been rock solid if it hadn't been for ASR and the many DSD experts here. For the casual listener, no amount of upconversion and avoidance of more invasive DSP can beat the subjective benefits of crossfeed and EQing headphones. My knowledge on other areas of audio is still too little.
 

richard12511

Major Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
4,335
Likes
6,700
There will always be those who are influenced by market talk and trends to make irrational purchase decisions and there will always be those who are happy to take advantage of such people.

Given that, the only real matters are the skill of producers/engineers and honesty/transparency of the entire supply chain with regard to provenance.

Well said. None of these differences make an audible difference, so it's a waste of time to discuss imo. Different format releases may very well sound slightly different, but it's because of what you mention in your last sentence.

Given the formats themselves sound exactly the same, my opinion is that the best format is the one that takes up the least space and costs the least.
 
Top Bottom