Octave Music Don Grusin High Resolution Music Analysis (Video)

[kongwee]

What does file size have to do with 11 MHz?

To contain the 11MHz 1bit signal. Mentions in the earlier messages parts of the decisions.

I advocate? I advocate 24 bit/48 kHz sampling or at best 88.2 kHz. Nothing remotely in Megahertz range let alone 11.

Isn't that imply for that earlier message? You don't need 192kHz PCM and above for 20-30kHz.

 

[amirm]

Isn't that imply for that earlier message? You don't need 192kHz PCM and above for 20-30kHz.

You don't. Most releases I test that are 192 kHz have junk in the ultrasonics. Natural music decays to nothing below 30 kHz or so.

The key is that the in-band spectrum is clean in 24 bit PCM. But not in DSD. Benefit of having 24 times more bits than DSD's 1 bit.

 

[kongwee]

You don't. Most releases I test that are 192 kHz have junk in the ultrasonics. Natural music decays to nothing below 30 kHz or so.

The key is that the in-band spectrum is clean in 24 bit PCM. But not in DSD. Benefit of having 24 times more bits than DSD's 1 bit.

I don't know about all these junk noise. I can't hear them. The real question I wanna to know is the removing the junk, doest it affect the audible part of data and even at DAC performance. No definite answer. Why not ask Octave record about the junk noise? Why is it there?

For me the more I read,(outside this tread) the more I toward 1bit. The mass DAC out. Isn't that SDM? Oversample at least 64 and send convert into PWM. 1 bit x128 sample cd player or CD walkman was very popular. I even can conclude my audio interface is SDM w/o looking at detail because of it low price. Why not DSD? a format that SDM DAC can read directly.

 

[Blumlein 88]

I don't know about all these junk noise. I can't hear them. The real question I wanna to know is the removing the junk, doest it affect the audible part of data and even at DAC performance. No definite answer. Why not ask Octave record about the junk noise? Why is it there?

For me the more I read,(outside this tread) the more I toward 1bit. The mass DAC out. Isn't that SDM? Oversample at least 64 and send convert into PWM. 1 bit x128 sample cd player or CD walkman was very popular. I even can conclude my audio interface is SDM w/o looking at detail because of it low price. Why not DSD? a format that SDM DAC can read directly.

Most likely your audio interface is 5 bit sdm .

The 192 format is quiet, but recordings pick up noise sometimes from other sources. It is not uncommon for microphones to have some response to 30 khz or so. Beyond that what's going on up there is not sound pickup from the microphones. With DSD some noise is part of the format even if the source is quiet.

 

[Miska]

This is the frustration I have with your arguments. Everything I posted here was to see if the more expensive and less convenient online and commercial DSD content is in some way higher resolution than the cheaper, smaller, and easier to manipulate PCM. That is the point of this whole thread.

Why not compare it from DAC output at that said 5 MHz range? DAC output is after all what matters.

Although for production side, using typical production chain these days where content is mixed and adjusted using analog mixing desk, not whole lot of difference between the two.

Then why are you arguing in this thread? It's not about HQPlayer and what can be achieved with a 20-core processor and a super-duper GPU along with custom algorithms.

My quad core i5-7600T does already amazingly good job.

It's about online sales of DSD content that is not better, and possibly worse than hi-res PCM. And yet, it costs more and is touted as being extremely hi-res while containing mostly noise in the ultrasonics.

I have no trouble finding typical PCM productions that are far worse than typical DSD productions. PCM is commonly compressed, driven to clipping, etc. And if something has been recorded for example in DSD256, why would you want to buy 44.1k PCM conversion of it instead of the original?

I just pay what is needed to get the original, even if it costs more than some lower resolution conversion.

 

[Miska]

It is not lucky. It is the opposite of what customers want to buy. This is why my video was such an eye opener for folks. They had no idea about this ugly truth in DSD encoding.

PCM performance is what I measure in DACs which have sunk many dBs below threshold of hearing. Here is a random example:

Distortion spikes at -135 dB. Combined noise+distotion is 5 dB better than threshold of hearing. This is stunning level of performance and is achieved for as little as $150.

So I suggest taking your misinformation campaign elsewhere. This dog don't hunt here....

And maybe that DAC would measure even better when you run it at DSD512?

Oh yes, I got two of the "stunning performance" DACs advertised here and both are performing very differently than the claims here. Out of three Topping DACs I have, only one is performing close to claimed values and that is D30Pro. And unsurprisingly also D30Pro performs better with DSD128/DSD256 than with PCM inputs.

For ES9038PRO based DACs, SMSL M500 mkII performs better than D90SE, costs about half. And it also performs better with DSD256/DSD512 inputs than with PCM inputs.

 

[pkane]

Why not compare it from DAC output at that said 5 MHz range? DAC output is after all what matters.

Because it's not what this discussion is about. But that appears not to matter to you.

 

[Geert]

I have no trouble finding typical PCM productions that are far worse than typical DSD productions. PCM is commonly compressed, driven to clipping, etc

No it's not, when you look at PCM recordings in the same genre as what's released in DSD. Audio production techniques have nothing to do with digital media format.

 

[Miska]

You don't. Most releases I test that are 192 kHz have junk in the ultrasonics. Natural music decays to nothing below 30 kHz or so.

You just said that PCM is clean yet you say 192 kHz releases have junk in the ultrasonics? Which one is it? If you want just 30 kHz bandwidth, DSD64 or DSD128 are fine. No steep anti-alias filters needed.

The key is that the in-band spectrum is clean in 24 bit PCM. But not in DSD. Benefit of having 24 times more bits than DSD's 1 bit.

DSD in-band spectrum is as clean or cleaner than 24-bit PCM.

DSD64 in-band spectrum is about 25 kHz
DSD128 50 kHz
DSD256 100 kHz
DSD512 200 kHz
DSD1024 400 kHz

Here's 0 - 22.05 kHz sweep, in 24-bit TPDF dithered PCM:

And same, but with DSD64, despite I've even run it through two conversions:

 

[Miska]

No it's not, when you look at PCM recordings in the same genre as what's released in DSD. Audio production techniques have nothing to do with digital media format.

It is still, I see much more digital clippings in PCM productions than in DSD productions, even of same genre.

 

[Miska]

Because it's not what this discussion is about. But that appears not to matter to you.

That is the point of DSD so it should be what this discussion is about.

 

[pkane]

That is the point of DSD so it should be what this discussion is about.

No, it's not. You're like a chatbot, as soon as the "DSD" keyword is mentioned anywhere, you start and keep spamming with the same content regardless of whether it's appropriate or not. Once more, it's about the marketing claims of content sellers about their DSD content being much more hires than PCM. It's not about different mastering, or about different DACs, not about different filters, or that PCM can be sometimes recorded with more noise and distortions than some other DSD.

But I'm sure your next post will ignore all this anyway. Despite Turing's halting problem proof, I can see that this particular loop will never halt

 

[kongwee]

Most likely your audio interface is 5 bit sdm .

.

Yup, SDM is everywhere, but DSD is being discriminated.

 

[Miska]

Once more, it's about the marketing claims of content sellers about their DSD content being much more hires than PCM.

I don't follow much what marketing claims content sellers make. But I don't really remember seeing such marketing claims. By all means provide me some links to such.

Or maybe some people have produced some non-existing marketing claims in their heads instead.

It's not about different mastering, or about different DACs, not about different filters, or that PCM can be sometimes recorded with more noise and distortions than some other DSD.

Let's not make claims then that PCM content would be better than DSD content. Many times "easier to manipulate" means "easier to screw up" as well. And to correctly play back PCM - without the ultrasonic distortions, you need much more DSP processing power than for DSD...

SMSL M400 playing 44.1k content:

SMSL M400 playing 705.6k content:

SMSL M400 playing DSD512 content:

 

[pkane]

I don't follow much what marketing claims content sellers make. But I don't really remember seeing such marketing claims. By all means provide me some links to such.

Or maybe some people have produced some non-existing marketing claims in their heads instead.

Let's not make claims then that PCM content would be better than DSD content. Many times "easier to manipulate" means "easier to screw up" as well.

Exit loop, Jussi. PCM content is not better, but DSD content that's been examined here is worse. You're talking about generalities while very specific examples were given, starting from the very first post in this thread.

 

[kongwee]

Yes they do, huge amount of imaging distortion, and you need a brickwall filter to remove it.

This is a snippet how a 0 - 22.05 kHz sweep looks like in RedBook:
View attachment 195012

Why zip zag line? Meaning at DC, a signal or noise produce at 45kHz and peak frequency?

 

[Miska]

Why zip zag line? Meaning at DC, a signal or noise produce at 45kHz and peak frequency?

It sweeps from DC to 22.05 kHz and back to DC. This removes the abrupt discontinuity point at the end of the sweep at 22.05 kHz. Source is 44.1k floating point data, so the DSD one is upsampled from this to DSD64 and then back down to 50 kHz sampling rate. You can see how the upsampling digital filter used cuts in a bit before 22.05 kHz.

So for both cases, source data is exactly same, PCM has been just dithered to 24-bit without any rate conversions.

This also the same source data used for the SMSL M400 measurement above.

 

[kongwee]

It sweeps from DC to 22.05 kHz and back to DC. This removes the abrupt discontinuity point at the end of the sweep at 22.05 kHz. Source is 44.1k floating point data, so the DSD one is upsampled from this to DSD64 and then back down to 50 kHz sampling rate. You can see how the upsampling digital filter used cuts in a bit before 22.05 kHz.

So for both cases, source data is exactly same, PCM has been just dithered to 24-bit without any rate conversions.

This also the same source data used for the SMSL M400 measurement above.

I still don't understand. The real signal is DC-22kHz in 10second. What is the line above that.

 

[Geert]

It is still, I see much more digital clippings in PCM productions than in DSD productions, even of same genre.

Because compared to the number of PCM release the number of native DSD releases is insignificant, especially if you focus on really good music and not on audiophile labels where sound quality is a higher prio than the music.

 

[Miska]

I still don't understand. The real signal is DC-22kHz in 10second. What is the line above that.

Ahh, sorry, that one, it is just one way sweep from DC to 22.05 kHz.

Those lines above 22.05 kHz are images. On PCM, the frequency spectrum keeps repeating itself. Next band above Nyquist is inverse frequency spectrum. And then again above sampling rate it is forward and then from 44.1k + 22.05k = 66.15k it is inverse again. And so on.

Or in other words, x * fs below sampling rate you have "negative frequencies" and inverse frequency spectrum. And above sampling rate you have "positive frequecies" and forward frequency spectrum. Here "x" can be any number from 0 to infinity.

In order to correctly reconstruct analog signal from PCM, everything above Nyquist frequency must be removed. So if we start with RedBook (CD), with 44.1 kHz sampling rate and 16-bit this means you need a very steep filter to attenuate frequencies above 22.05 kHz by at least 96 dB to correctly reconstruct 16-bit resolution. For 24-bit resolution you need at least 144 dB attenuation and for 32-bit you would need 192 dB attenuation.

Using digital filters we can achieve enough steepness to do this, as you can see from my above example. If we use digital filters to increase sampling rate to 352.8 kHz like DAC chips do, these images move be around multiples of 352.8 kHz. DAC chips don't have enough DSP processing power to do proper digital filters above this, so you can see images around multiples of 352.8 kHz. If you do this externally, many DAC chips allow you to use 705.6 kHz sampling rate instead, which means these images move to around multiples of 705.6 kHz. If we could use high enough sampling rate, eventually the analog output filter in DAC would be able to attenuate these images enough. Typical analog filter in a DAC is 2nd order Butterworth with -3 dB point at 100 kHz. This means it as 12 dB/oct attenuation. Thus, for 96 dB attenuation, we would need 8 octaves. And for 144 dB attenuation we would need 12 octaves. And for 192 dB attenuation we would need 16 octaves.

With DSD output and enough DSP power, we can do digital filters for example to 44.1 x256 rate = 11.2896 MHz or 44.1 x512 = 22.5792 MHz. With number of DACs on the market we can also use 44.1 x1024 rate = 45.1584 MHz digital filter output rates. Thus these images are not an issue anymore. OTOH, if we already record at for example DSD256, we don't need to employ these kind of oversampling filters at all.

This is just the output side.

At input side, the problem is even more severe. If we don't employ steep digital filter, all the information outside of fs/2 will fold down to the 0 - fs/2 band. Here's an example of 0 - 176.4 kHz sweep at 352.8 kHz sampling rate, converted to 44.1 kHz 16-bit, without digital filter:

Whenever you do band limiting with filters, it has both frequency and time domain implications, because the two are linked through 1/x relationship.