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

Modulators with more than just a single bit can be effectively dithered, single bit ones can't. Its the lack of ability to correctly dither which Lipshitz and Vanderkooy claim is the fly in the ointment for DSD.

Sigma delta modulator have as many bits as have converted PCM signal.
As rule these schemes is cascade. There used low bit modulators.

I can't understand why need dithering in sigma-delta modulator. May be we see in this phrase ditherent things.

For analog singal converted in ADC to sigma-delta happens same things like to PCM in sigma-delta. But there 1-bit modulation work with noise level about -148 dB even for D64. All depend on implementation of sigma-delta modulator.

Could you show scheme of the modulator what you describe (datasheet some ADC, DAC or other)?
 
I can't understand why need dithering in sigma-delta modulator. May be we see in this phrase ditherent things.

The modulator includes a quantizer. Quantization being a non-linear operation it generates quantization distortion. Dither's used as a means of decorrelating that distortion from the signal, effectively rendering it into quantization noise.


Could you show scheme of the modulator what you describe (datasheet some DAC or other)?

You could have a look at this paper - http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.701.963&rep=rep1&type=pdf
 
For me me the only thing which is obvious is that there is to much testosterone here around:)
Personally I do not like PCM 24/192 for its clinical sound. 24/96 is just enough for me with the exception of RR 24/176,4.
However the few DXD files I have are very good, may be the best I have.
I agree that PCM is better for rock and music based on PRAT.
DSD is better for jazz and classical for its space , 3D soundstage and generally better timbre.
Good vinyls are on the top but the main thing is quality of recording and mastering and not coding or source, of course imho.
All this theorerical papers are for constructors of hifi gear and software , me as an audiophile care more about what I hear because even specialists cannot come to the agreement.
I am very far from fighting which type of files is more appropriate but I am rather in favour of native lower resolution instead of upsampling to high PCM or DSD.
Why not listen to all of them i.e good quality recordings be it PCM, DSD or vinyls?
Just because of testosterone?:)
 
Last edited:
Yeah I also didn't see a reference to dithering the quantizer which in that paper is two bits. You asked about an example of a scheme of a modulator so that's what I found.

So are you interested in dither being applied to single bit quantizers in DACs? I'm not really clear on your question.
 
Yeah I also didn't see a reference to dithering the quantizer which in that paper is two bits. You asked about an example of a scheme of a modulator so that's what I found.

So are you interested in dither being applied to single bit quantizers in DACs? I'm not really clear on your question.


Modulators with more than just a single bit can be effectively dithered, single bit ones can't. Its the lack of ability to correctly dither which Lipshitz and Vanderkooy claim is the fly in the ointment for DSD.

I don’t understood this phrase.

I see there:

1. sigma delta modulator with several bits at output.

2. Mentioned dither effectiveness and ability to correct dithering.

Point #1 like to scheme that you gave.

In sigma-delta modulator (independent on output bits) used noise shaping. It is like noise-shaped dither.

But I can’t understand:
1. why need dithering in this scheme and where there it may be applied?
2. What is mean effectively? What is criteria of effectiveness?
 
My comments about dithering 1bit quantizers are based on the paper I've been referring to. Suggest you have a read of that to see if you agree with Lipshitz and Vanderkooy's math.

https://sjeng.org/ftp/SACD.pdf

Thank you for link.

In the article compared concepts of converters of analog to PCM via sigma-delta modulator: with 1 bit and multibit outputs.

When we optimize sigma-delta modulator we use 2 criteria:

1. Noise level.

2. Steadiness to overload (maximal amplitude at input the modulator).

If demodulator is overloaded it give at output pure sine(s) or silence.

After overload modulator don't return to normal work without forced reset.

May be there will other effects, but I don't observed it yet.

If we improve point #1, then point #2 is degraded. And contrary.

But we can improve point #1 with keeping point #2 via increasing number bits at output the modulator and/or wider band.

All these things very actually for DSD64. For DSD256 and above it is almost no matter, in my opinion.

Dither in the article is "native" noise of sigma-delta modulator.
 
You may note that L&V consider limit cycles and noise modulation effects important criteria too in assessing S-D modulators.
 
You may note that L&V consider limit cycles and noise modulation effects important criteria too in assessing S-D modulators.

I estimate the modulators via sweep sine at input and time-spectrum analyzis. It open many details that invisible in 2-D spectrum.

Because it is sibultaneously integrated look and look to details with scaling/changing resolution ability.

There we can see all aliases/oscillations and how noise is changed by frequency.
 
Yet I guess in some sense the discussion is moot. As mentioned on this thread, a lot of modern A/D converters operate in DSD anyway, so even if you listen to PCM the recording is often pre-processed as DSD right at the source.

opus 112 responded:

This is a common mis-conception, I think its one actively promoted by DSD enthusiasts too (I've seen claims of 'DSD-like' ADCs for example) but there is a difference between most modern ADCs and DSD. The difference is in the number of bits in the modulator. I'm aware of one (the Grimm) which uses a single bit modulator, there may well be others but many (likely most) use multi-bit modulators (typically up to 5 bits). Modulators with more than just a single bit can be effectively dithered, single bit ones can't. Its the lack of ability to correctly dither which Lipshitz and Vanderkooy claim is the fly in the ointment for DSD.

Thanks for clearing this up, Opus. I was actually not just misled by DSD enthusiasts, but also by your own post earlier in the thread. I thought you were on their side (emphases added to quote):

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.

You further say:

No, but what you say is correct, not all D-S is DSD. But all DSD is D-S.

That is exactly what Berkeley Audio says. If I remember correctly, there was a vigorous discussion on Computer Audiophile where Michael Ritter from Berkeley Audio countered the claim by Miska from HQPlayer (total DSD enthusiast) that Delta-Sigma (D-S) by definition is DSD by pointing out that the Berkeley DACs, using Delta-Sigma modulation, are multibit (5 or 6 bits) and thus PCM.

Miska of course wouldn't have any of it, but apparently he is wrong, also according to you, if I understand you correctly.
 
Thanks for clearing this up, Opus. I was actually not just misled by DSD enthusiasts, but also by your own post earlier in the thread. I thought you were on their side (emphases added to quote):

Ah, then thanks also to you for clarifying what you found misleading in what I wrote. The DSD supporters over on CA try to undermine multibit lovers by telling them that the vast bulk of their beloved recordings are already in effect made with DSD technology. But S-D and DSD aren't the same, L&V point out its the 1bit quantizer of DSD which is the nub the problem.

That is exactly what Berkeley Audio says. If I remember correctly, there was a vigorous discussion on Computer Audiophile where Michael Ritter from Berkeley Audio countered the claim by Miska from HQPlayer (total DSD enthusiast) that Delta-Sigma (D-S) by definition is DSD by pointing out that the Berkeley DACs, using Delta-Sigma modulation, are multibit (5 or 6 bits) and thus PCM.

If 1bit defines DSD then yes, any higher number of bits implies PCM. But I don't see it quite in such black and white terms myself, just my own view though.

Miska of course wouldn't have any of it, but apparently he is wrong, also according to you, if I understand you correctly.

Yep, I've already had that discussion with Miska on CA. He is of course heavily invested in the kool-aid of DSD - the saying of Upton Sinclair comes to mind 'Its hard to teach a man something when his salary depends on his not understanding it'.
 
Ah, then thanks also to you for clarifying what you found misleading in what I wrote. The DSD supporters over on CA try to undermine multibit lovers by telling them that the vast bulk of their beloved recordings are already in effect made with DSD technology. But S-D and DSD aren't the same, L&V point out its the 1bit quantizer of DSD which is the nub the problem.

You're welcome, and thanks again for your clarification. Given the technical confusion out there, I think it is important to be really precise technically (if you're a technical expert and have the ability), and perhaps even to qualify each statement against possible misconceptions from the start. I know, that's a pain, but it may be worth it.

Yep, I've already had that discussion with Miska on CA. He is of course heavily invested in the kool-aid of DSD - the saying of Upton Sinclair comes to mind 'Its hard to teach a man something when his salary depends on his not understanding it'.

Great quote, Opus. To extrapolate more generally, its hard to teach a man something when his convictions depend on his not understanding it. That's pertinent to discussions about politics and religion/atheism as well (that's why these discussions often are so thorny and futile, and passions run high).

Here's the full DSD kool-aid by Paul Gowan from PS Audio:

The world is DSD

I'd be interested in your response to that.
 
The world is DSD

I'd be interested in your response to that.

A more striking example of kool-aid thinking it would be hard to find. Let's unpack a little of what's being said there -

Well, the fact is that the principles which underpin DSD are hard at work in almost every digital audio device you own, from the cheapest DAC chipsets built into your mobile phone, to the most expensive stand-alone audio DACs.


They're referring to quantization with noise shaping (aka S-D) and oversampling so yes, they're quite correct that those principles underlie the vast majority of consumer DACs and ADCs in use today.

The leap of faith comes here, at the end of the quoted paragraph -

So, at some point, if you want to understand Digital Audio, you’re going to have to understand DSD.

Notice the logical fallacy in that? There's probably an official name for it but I'm too lazy to look it up. They're arguing the wrong way around - DSD is built on the foundation of S-D so you'd want to understand the foundation, not the superstructure would you not if your aim was greater understanding of digital audio in general?
 
A more striking example of kool-aid thinking it would be hard to find. Let's unpack a little of what's being said there -

Well, the fact is that the principles which underpin DSD are hard at work in almost every digital audio device you own, from the cheapest DAC chipsets built into your mobile phone, to the most expensive stand-alone audio DACs.


They're referring to quantization with noise shaping (aka S-D) and oversampling so yes, they're quite correct that those principles underlie the vast majority of consumer DACs and ADCs in use today.

The leap of faith comes here, at the end of the quoted paragraph -

So, at some point, if you want to understand Digital Audio, you’re going to have to understand DSD.

Notice the logical fallacy in that? There's probably an official name for it but I'm too lazy to look it up. They're arguing the wrong way around - DSD is built on the foundation of S-D so you'd want to understand the foundation, not the superstructure would you not if your aim was greater understanding of digital audio in general?

Thanks for the excellent analysis! It also makes me understand the topic better.
 
That is exactly what Berkeley Audio says. If I remember correctly, there was a vigorous discussion on Computer Audiophile where Michael Ritter from Berkeley Audio countered the claim by Miska from HQPlayer (total DSD enthusiast) that Delta-Sigma (D-S) by definition is DSD by pointing out that the Berkeley DACs, using Delta-Sigma modulation, are multibit (5 or 6 bits) and thus PCM.

If 1bit defines DSD then yes, any higher number of bits implies PCM. But I don't see it quite in such black and white terms myself, just my own view though.

The Computer Audiophile site is mucked up with the transition to a new design, so you can't find old posts, but fortunately I found Berkeley's statement in my saved old archives from that site.

Here is the post from 12-23-2013 by The Computer Audiophile, quoting Berkeley:

Hi Guys - I received one last response from Berkeley Audio Design. It addressed some items directly.

Berkeley Audio Design:
DSD versus PCM part 2.

?
Initial assertion:

Multibit DSD is not PCM.

PCM is typically 24 bit of the whole sample value while multi bit DSD is typically 5-6 bits of DIFFERENCE between adjacent samples. In 1 bit DSD this difference is binary (0 or 1) while in multi bit DSD it's the same difference between samples but quantized with 5-6 bits (Sabre has a 6 bit DAC). Multibit DSD is BETTER than 1 bit DSD, so there is nothing wrong with going 1 bit> 6bit which is what all these DAC chips do.

Our assertion in response:

Multi-bit delta-sigma audio very definitely is PCM, and represents coarsely quantized whole output sample values being sent to a linear PCM DAC.

We have discussed the fact that multi-bit delta-sigma modulation is not the difference between adjacent samples at length in our previous post.

Going to a discussion of specific DAC structures does not even address our assertion.

It is true that typical monolithic multi-bit delta-sigma DAC’s use what is known as a thermometer DAC structure in which a series of current-source elements of equal value are switched on or off and their currents are summed to produce the output current. It is also true that, because the current output of each element can never be exactly same as the others, the resulting errors are normally scrambled and noise shaped by various schemes of mapping the required number of on current sources to different elements for each successive sample.

The number of elements, typically current sources, is normally an odd number, but this is not inconsistent with binary numbers. There is an extra state in which all of the elements are off, giving an even number of output states. For example, a 4-bit DAC would have 15 elements whose output sum goes from 0 to 15, which is 16 states.

It is not even necessary that the number of elements be related to a power of two. The nature of the delta-sigma algorithm and the element scrambling algorithm can map a binary number to a larger number of elements than the largest number represented by the binary number.

The important point to take away from the above is that the DAC itself is linear. It is true that the input to the DAC elements is unary coded, but what is not even mentioned in all the discussion is the mapping of a binary 5-6 bit input to a 32 or 64 bit unary code that actually controls the DAC elements. It would be absurd to think that much processing is done on 64 bit unary coded data. That is just the end result.

The multi-bit delta-sigma data stream is a series of 5-6 bit binary values or in some cases 8 bit data, i.e. PCM with noise shaping. These binary values are then translated into scrambled unary code just before the output elements. That unary code determines which elements are turned on, and the sum of theoutput elements, typically in current form, is the output of the DAC. In most designs, that current is converted to a voltage and then low-pass filtered with a conventional filter to yield the audio.

The multi-bit delta-sigma data stream is generated by a delta- sigma modulator, which is typically used to re-quantize and noise shape an input with a different bit depth. Each multi-bit word is a binary number which represents an instantaneous value of the signal – in other words PCM.

At Pacific Microsonics, prior to starting Berkeley Audio Design, we were directly involved in the design of several multi-bit DSM DACS, two of which were commercially produced by Japanese companies. These DAC’s all used the principles described above.
 
I guess the crux of the above argument by Berkeley Audio why mutli-bit delta-sigma is PCM lies in this paragraph:

The multi-bit delta-sigma data stream is generated by a delta- sigma modulator, which is typically used to re-quantize and noise shape an input with a different bit depth. Each multi-bit word is a binary number which represents an instantaneous value of the signal – in other words PCM.

(Not that I can pretend to technically understand this, perhaps you can help, opus112.)
 
My Dac has R2R PCM and DSD512 chipless. I am extremely happy.

Upsampling to DSD512 is magical, but also to DXD, PCM 384 is wonderful too.

Percussion stuff may be better in PCM, but Jazz/Vocals is better in DSD as a general rule. I want it all!
 
My Dac has R2R PCM and DSD512 chipless. I am extremely happy.

Upsampling to DSD512 is magical, but also to DXD, PCM 384 is wonderful too.

Percussion stuff may be better in PCM, but Jazz/Vocals is better in DSD as a general rule. I want it all!

Wisnon

The DAC you speak of is the Lampi L4?

Kerry
 

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu