Why do high up-sampling/ over-sampling rates (DSD,) kill PRAT and aliveness of music? Any ideas?

[Audiophile Bill]

I honestly don't know if we went that high... I was hanging out with my friend. And his digital audio software had a feature to upsample, doubling / nearly doubling from 44 to 88/ 96 to 176/ 192. The music turned more and more dull to me the higher we went... Going past 88, I would enjoy the music in my car more...

You realise that in all likelihood your dac is upsampling internally to 24/192 anyway, right?

 

[microstrip]

I honestly don't know if we went that high... I was hanging out with my friend. And his digital audio software had a feature to upsample, doubling / nearly doubling from 44 to 88/ 96 to 176/ 192. The music turned more and more dull to me the higher we went... Going past 88, I would enjoy the music in my car more...

What were the DAC and digital software being used?

 

[Legolas]

I honestly don't know if we went that high... I was hanging out with my friend. And his digital audio software had a feature to upsample, doubling / nearly doubling from 44 to 88/ 96 to 176/ 192. The music turned more and more dull to me the higher we went... Going past 88, I would enjoy the music in my car more...

What DAC was he using? Probably a delta-sigma oversampling? IMO a lot of the format argument is the reflection of the final filter in the DAC (or lack of) and how it handles the data stream. I am in the resident resolution camp TBH, and no digital filtering. Quite possibly DACs using a chip less engine / FPGA and DSD can avoid many of the pitfalls of the brick wall filter. One point about DSD in my mind though, it is stated it creates it's own modulation and that is a real problem at rates as low as DSD64, I am guessing it is reduced as you go higher, so possibly it is DSD256 at least or forget it, go back to PCM?

The other point is, how many studios use DSD as the master, or PCM? No idea, but I am thinking if the format was altered as well as downsampling to 96K or 44.1, then going to DSD 256 to whatever, we have 3 format changes right there. Something to consider maybe.

Going back to the original thread statement, I am thinking oversampling in software, to then feed a DS DAC that will upsample, well that is not going to be optimal IMO. And IMO oversampling outside of the multiples will only make things worse.

 

[christoph]

What were the DAC and digital software being used?

THAT is the correct question!

 

[opus112]

Nevertheless, the most revealing comparison was having the same vinyl record ripped (by a famous dude on $250K of equipment) to PCM 192/24 and to DSD, which eliminated the constant argument of whether the material came from the same master.

The difference is probably going to depend on the ADC, nowadays most ADCs are using high levels of oversampling with noise shaping internally. So your PCM 24/192 recording most likely isn't pure PCM.

With your experiment comparing the PRaT of DSD to PCM, what was the DAC in use? With the mainstream S-D type of DAC, oversampling on the PC normally improves SQ whereas with multibit DACs the reverse is true.

 

[caesar]

THAT is the correct question!

Not sure why - I can pick out DSD sonic signature as a 20 year spectral user can pick out a 1970s tube sound, and vice versa. But since folks are interested, we used a whole smorgasbord of "best digital", including Naim, Bricasti, Chord DAVE, and Berkeley Ref. I think the software was Roon, but we used other apps as well. And over the years I have heard DSD implemented by PS Audio, EMM, Playback Designs, etc., and it still lacks what I like, while giving DSD fans what they want.

 

[Al M.]

I get that pcm has a bite but to me that bite makes it less realistic.

The more I listen to unamplified live music the more realistic the 'bite' of PCM seems to me -- it simply reflects to a good degree the real thing, as heard in many but the smoothest sounding acoustics. DSD can also have realistic bite though. But in fact often the 'bite' of reproduced sound doesn't even go far enough, compared to the real thing.

 

[Fiddle Faddle]

Going past 88, I would enjoy the music in my car more...

I came across this article last night and I almost 100% agree with everything the author wrote (and 100% outright with what Dan Lavry believes). My only slight point of deviation is being more resolute in keeping sampling rates above 44.1 Khz but even the author appears to concede that point to some extent.

But Dan Lavry believes sample rates have a sweet spot around 50 - 60 Khz and I absolutely 100% agree with that based upon all my years of listening (and recording).

http://www.trustmeimascientist.com/...rates-when-higher-is-better-and-when-it-isnt/

 

[microstrip]

Personally, I enjoy rock, blues, and jazz, and can't stand how DSD makes the vast majority of these recordings so dull. DSD sure does make things smoother, but without the snappiness PRAT, DSD sounds very artificial to me.

Personal tastes aside, does anybody understand or have any theories why the high up-sampling/ oversampling rates of DSD, and such, kill the snappiness, aliveness, and PRAT of music?

Not sure why - I can pick out DSD sonic signature as a 20 year spectral user can pick out a 1970s tube sound, and vice versa. But since folks are interested, we used a whole smorgasbord of "best digital", including Naim, Bricasti, Chord DAVE, and Berkeley Ref. I think the software was Roon, but we used other apps as well. And over the years I have heard DSD implemented by PS Audio, EMM, Playback Designs, etc., and it still lacks what I like, while giving DSD fans what they want.

Re-reading your question and thread is becoming confusing to me, as you mix completely different things in a single pot. What are you exactly addressing - just DSD recordings? DACs that oversample ? What is mean by "and such"?

 

[Ken Newton]

Does anyone know whether or not anti-alias filtering is utilized by DSD encoding (A/D) units?

 

[caesar]

Re-reading your question and thread is becoming confusing to me, as you mix completely different things in a single pot. What are you exactly addressing - just DSD recordings? DACs that oversample ? What is mean by "and such"?

Microstrip,

Good Point. For context, let's take a step back. There are a lot of guys in this hobby who hate PCM. "It causes fatigue. It has glare. It lacks ease, palpability. It has harshness!" So they get this "newer" technology marketed to them called DSD. This technology takes away a lot of the digital nasties these folks complain about, and this new format also adds "spaciousness".

But, unfortunately, there is no free lunch. By working with a very high up-sampling rate, in addition to taking away these nasties, DSD also takes away "bite", "rhythmic dynamic drive", "PRAT", "muscular deep bass", or whatever you want to call it and replaces it with a "softness", "fake warmth", "softened attacks", and "lacking life" - in comparison to PCM.

My test for DSD gear is Hard Again by Muddy Waters. Unfortunately, DSD upsampling DACs make Muddy sound like a f*cking p8ssy, sucking all the intensity and energy out of his voice, whereas quality PCM DACs bring out the virility of his voice.

Interestingly, I find similar qualities pressing the upsample button using popular digital computer apps with PCM, with anything above 88/96 making Muddy sound like that friggin pussy instead of expressing raw emotion one listens to blues music for. (I don't really know which app, I think it's Roon, but why take up space in your brain with things that suck and you are not interested in?)

Ultimately, there are manufacturers catering to the PCM crowd and those targeting the DSD crowd. People are going to listen and make their trade-offs, accordingly. All I am simply asking in this thread is if anyone has discovered a technical rationale for this difference in presentation. That's all.

If not, engineers really need to get to work and figure this out, as Muddy must NEVER come across sounding like a pussy, no matter how hard the DSD guys claim it is the ultimate digital format.

 

[Lee]

This is an odd thread. Higher resolution digital does not harm PRAT. Must have something to do with your system Caesar.

 

[caesar]

This is an odd thread. Higher resolution digital does not harm PRAT. Must have something to do with your system Caesar.

I keep saying it's not my system - I wouldn't let DSD anywhere near my system - but every system I have heard DSD in

 

[Yuri Korzunov]

I honestly don't know if we went that high... I was hanging out with my friend. And his digital audio software had a feature to upsample, doubling / nearly doubling from 44 to 88/ 96 to 176/ 192. The music turned more and more dull to me the higher we went... Going past 88, I would enjoy the music in my car more...

DAC may have differeng distortions for different resolutions (DSD/PCM + sample rate + bit depth). Sample rate value mean nothings as itself. But, as rule, higher sample rates have lesser aliases.

One or several mode(s) may have minimal distortions/aliases ("best sounding" among available ones).

It depend on implementation only. As example, pro audio interface may have better sound on 44 than home unit on 192 kHz.

Sample rate conversion software (resolution conversion software) need for modifying resolution for one of the "best sounding" mode.

Sample rate conversion software aim is changing of resolution with minimal loses.

Resolution of audio file as itself don't impact to quality. Except objectively (measurable) higher quantization noise for lesser bit depth and lesser sample rate.

For DSD higher sample rate allow decrease noise in audible range. Higher sample rate - wider band with minimal noise.

If we upsample D64 to D256, target file have same (to original) noise level. But used DAC may work better in D256 mode, as example. Thus we can expect sound improving.

Resume:

1. Need found "best sounding" mode sample rate/bit depth/DSD/PCM.
2. And made up/down-sampling audio files to the sample rate/bit depth/DSD/PCM.
3. For sure need backup originals, because new further DAC may have other "best sounding" mode.

 

[christoph]

(I don't really know which app, I think it's Roon, but why take up space in your brain with things that suck and you are not interested in?)

And there we have it
Roons own up-resampling engine/music player really is nowhere near HQPlayer soundwise.
Even if you make absolutely no alterations to the music and play at native resolutions and formats, HQPlayer sounds way better than roon alone

Your claim is a bit like saying "driving offroad sucks ultimately, period!" but you're trying to do it in a Jaguar F Type and not in a Range Rover...

Roon is IMHO the best software for organizing and exploring music but soundwise I much prefer HQPlayer, even natively.
And the roon re- and upsampling is nowhere near the quality of HQPlayer.
Luckily for me (and others with open mind, willing to try), I can have both together

That is exactly why the player matters

 

[caesar]

And there we have it
Roons own up-resampling engine/music player really is nowhere near HQPlayer soundwise.
Even if you make absolutely no alterations to the music and play at native resolutions and formats, HQPlayer sounds way better than roon alone

Your claim is a bit like saying "driving offroad sucks ultimately, period!" but you're trying to do it in a Jaguar F Type and not in a Range Rover...

Roon is IMHO the best software for organizing and exploring music but soundwise I much prefer HQPlayer, even natively.
And the roon re- and upsampling is nowhere near the quality of HQPlayer.
Luckily for me (and others with open mind, willing to try), I can have both together

That is exactly why the player matters

Thanks, but the differences I am hearing are not just in that roon system, but every time I hear DSD. But this is a hobby for me, and I am open to new experiences. I am very curious to hear HQ Player. And also would like to get a better listen to the Lamplizator.

 

[bonzo75]

The more I listen to unamplified live music the more realistic the 'bite' of PCM seems to me -- it simply reflects to a good degree the real thing, as heard in many but the smoothest sounding acoustics. DSD can also have realistic bite though. But in fact often the 'bite' of reproduced sound doesn't even go far enough, compared to the real thing.

The bite of PCM vs DSD is different from the bite of Lyra vs Koetsu. The bite of Lyra is more towards real, the bite of PCM can be nothing more than a digital irritation. There are different types of bite in audio.

Btw, your above post was a reply to an ex-first trombonist. And just to give you a perspective of my week, I was listening to Richard Tognetti with his chamber orchestra on Monday and Tuesday, Aladdin musical on Wednesday, Schubert 9th and with Anne Sofie von Otter singing his leider stuff yesterday, and Schubert 8th and Brahms Requiem at Barbican today, starting in a couple of hours. Next week a led zep cover band, after that Joshua Bell, then Mahler 7, some baroque, piano, etc.

Apart from live music, you do need gear experience to make the connection. Listening to one thing at Goodwin's, one thing at home, another at Peter's, will not be sufficient. You need to get all in a room, in different set ups, to complete the connection.

 

[jfrech]

My test for DSD gear is Hard Again by Muddy Waters. Unfortunately, DSD upsampling DACs make Muddy sound like a f*cking p8ssy, sucking all the intensity and energy out of his voice, whereas quality PCM DACs bring out the virility of his voice.

Ceasar, a quick SA-CD.net search about your DSD "Hard Again by Muddy Waters" - says it's not out in SACD. What label are you using for your DSD testing of Hard Again? So I assume you're only talking about up-sampling from redbook?

 

[caesar]

Ceasar, a quick SA-CD.net search about your DSD "Hard Again by Muddy Waters" - says it's not out in SACD. What label are you using for your DSD testing of Hard Again? So I assume you're only talking about up-sampling from redbook?

Yes, I should have been clearer - it's "ancient" CD I really like. When I have it upsampled by DSD DACs, I don't like the outcome.

 

[caesar]

The bite of PCM vs DSD is different from the bite of Lyra vs Koetsu. The bite of Lyra is more towards real, the bite of PCM can be nothing more than a digital irritation. There are different types of bite in audio.

Btw, your above post was a reply to an ex-first trombonist. And just to give you a perspective of my week, I was listening to Richard Tognetti with his chamber orchestra on Monday and Tuesday, Aladdin musical on Wednesday, Schubert 9th and with Anne Sofie von Otter singing his leider stuff yesterday, and Schubert 8th and Brahms Requiem at Barbican today, starting in a couple of hours. Next week a led zep cover band, after that Joshua Bell, then Mahler 7, some baroque, piano, etc.

Apart from live music, you do need gear experience to make the connection. Listening to one thing at Goodwin's, one thing at home, another at Peter's, will not be sufficient. You need to get all in a room, in different set ups, to complete the connection.

Makes sense, but at the same time I know musicians and recording engineers, truly "golden ear guys" who like Al and me, think that "PCM bite" adds to realism. And I am sure we can find a bunch of other musicians and music teachers who would prefer DSD...