Time resolution of Redbook (16/44) PCM

[Stereo Skunk]

No, it creates the distortion when you push it through standard audio equipment in Analog after reconstruction.
That's why it has the big-*** chebychev filter at 50kHz that limits its bandwidth and response time anyhow.

Did you forget about that?

I was referring to his graphic that was titled PCM-DSD Conversion.

Or did you not read the post?

 

[Stereo Skunk]

Ok.

Look.

Fact #1: Time resolution is a function of both bandwidth AND SNR.
Fact #2: DSD has a sharpish filter at 50kHz to keep all that horrid ultrasonic noise out of your analog sections.
Fact #3: That filter is barely sharp enough to keep that noise out.
Fact #4: DSD is NOTHING BUT 1 BIT NOISE SHAPED PCM.
Fact #5: That's neither good nor bad, it just is.
Fact #6: You MUST convert DSD to some kind of multibit PCM if you EVER want to change the level digitally. This is a massive pain in the behind.
Fact #7: Even using DXD, the sampling rate is so high that typical filters require 64 to 128 bit mantissas.
Fact #8: Almost all modern ADC's and DAC's use something in the oversampling domain (usually 4 bit, not 1 bit, but same principle) that is processed inside the chip to PCM.

DSD is/was nothing but an attempt at security by obscurity.

I'm sorry but I can't find where someone asked...

 

[Music1969]

RTFM

Manual doesn't answer the question.

 

[Stereo Skunk]

@j_j has decades of experience in this field and extensive data to support his statements. Here's a short bio on the AES site (been posted many times before): https://www.aes.org/member/profile.cfm?ID=1800973364

He said "several orders of magnitude", not "single order of magnitude". Time resolution of CD-rate and resolution converters is on the order of ns to ps, well below (by several orders of magnitude) the 10 us or so ITD threshold.

I don't understand the need to insult (with no research into their backgrounds) the very scientists we want to attract to this site.

interesting note on the 10us for ITD.

I'd like to ask about a single ear's timing 'resolution'

here is a 'well designed' speaker (revel salon2)

What's amazing is that the mid and woofer drivers are not even close to 10us time aligned. Yet, Harman insists that this kind of distortion is benign.

I don't like this. I think it's audible.

 

[voodooless]

Manual doesn't answer the question.

Hmm, you are right, the manual does not mention this. In any case: putting the E30 in DAC mode enables DSD direct.

 

[Music1969]

Hmm, you are right, the manual does not mention this. In any case: putting the E30 in DAC mode enables DSD direct.

Hehe so you didn't even RTFM, a lesson learnt ;-)

I'm not sure how you can know DSD Direct is properly activated (or de-activated) if it's not in the manual or some other official word from Topping?

Not only how to active but also de-activate (if someone wants that).

It's crystal clear with RME ADI-2. Not at all clear with the Topping you mention...

 

[j_j]

I was referring to his graphic that was titled PCM-DSD Conversion.

Or did you not read the post?

So, you do realize that the remodulation via delta-sigma reintroduces the noise, necessarily, right? Yes? No? It's a byproduct of delta-sigma convertors.

Was there some point to your snark?

 

[voodooless]

I'm not sure how you can know DSD Direct is properly activated (or de-activated) if it's not in the manual or some other official word from Topping?

If the DSD proponents think the difference is so crystal clear, should it not be obvious ? The only way we even know of this feature is because of other marketing material from Topping:

Not only how to active but also de-activate (if someone wants that).

Put the E30 in pre-mode.

 

[j_j]

interesting note on the 10us for ITD.

I'd like to ask about a single ear's timing 'resolution'

here is a 'well designed' speaker (revel salon2)

View attachment 148328

What's amazing is that the mid and woofer drivers are not even close to 10us time aligned. Yet, Harman insists that this kind of distortion is benign.

I don't like this. I think it's audible.

You do understand, do you not, that the fact there's a highpass filter in any speaker means there will be some ringing, necessarily, right?

What's more, what matters for ITD is simple, "do the two speakers match". What you're addressing is intRAchannel time delay, not intERchannel time delay. ITD is IntERchannel time delay. You did know that, right? You can have massive group delay in a speaker, and have no, repeat NO ITD if the two speakers are matched. Now, I'm not saying that's a good thing, of course.

I must note here that you seem to be heading down the same path as some other posters who don't understand linear superposition, but who like to play dominance games. In fact, you are starting to sound identical.

 

[j_j]

I'm sorry but I can't find where someone asked...

Perhaps you should check the head of this thread as to what it's all about, and how this relates to DSD, which you did note a few posts ago is the current subject.

 

[DonH56]

interesting note on the 10us for ITD.

I'd like to ask about a single ear's timing 'resolution'

here is a 'well designed' speaker (revel salon2)

View attachment 148328

What's amazing is that the mid and woofer drivers are not even close to 10us time aligned. Yet, Harman insists that this kind of distortion is benign.

I don't like this. I think it's audible.

ITD relates to differences between two speakers. Our two ears resolve timing difference down to 10 us or to determine a location vector. If both speakers send the signal at the same time, and it arrives at the listener's ears coincidentally, then ITD is a non issue. You get that sort of response from any filter, be it a crossover in the AVR, speaker, or wherever. Look up group delay. Every filter, be it speaker crossovers, AM/FM radio IF filters, radar filters, video band filters, etc. do the same thing and yet all those things work just fine.

The NRC, Harman, AES, etc. etc. etc. have researched this sort of thing for decades, and that is just the audio version. The exact same principle applies to filters for radios and other telecommunication systems going back to the beginning of electronics.

The incorrect ideas that many people think, with no scientific background or basis for them, is one of the things ASR tries to correct, for those wiling to read, research a little on their own, and/or accept based upon the experience of those with many years of research in the field.

"I don't understand it, but think it is so, thus you must be wrong" is not a valid argument. But that is how many "win" on the 'net; it is not practical nor worthwhile to try to explain fundamental concepts, that may take hours or years to learn, in a few lines in a post to someone who has already made up their mind. Faced with an unwillingness to learn or accept the prior art (through their own research), it is simply not worth the time and effort to (try to) correct invalid assumptions and beliefs.

 

[pma]

it is simply not worth the time and effort to (try to) correct invalid assumptions and beliefs.

Like Hydra, for every head chopped off, the Hydra would regrow two heads

https://en.wikipedia.org/wiki/Lernaean_Hydra

 

[Blumlein 88]

Like Hydra, for every head chopped off, the Hydra would regrow two heads

https://en.wikipedia.org/wiki/Lernaean_Hydra

Or Brandolini's law.
https://en.wikipedia.org/wiki/Brandolini's_law
It takes an order of magnitude more effort to refute bad information as it does to spread such misinformation.

 

[mansr]

The first website that pops up when I Google Blue Coast Records is PS Audio.

Forgive me for my skepticism.

Blue Coast recordings are generally well made, even though Cookie believes in all manner of voodoo. Her choice of music to record is a different matter, and most of it doesn't appeal to me at all.

 

[AdamG]

Perhaps you should check the head of this thread as to what it's all about, and how this relates to DSD, which you did note a few posts ago is the current subject.

Thread reply ban issued.

 

[thorvat]

interesting note on the 10us for ITD.

I'd like to ask about a single ear's timing 'resolution'

here is a 'well designed' speaker (revel salon2)

View attachment 148328

What's amazing is that the mid and woofer drivers are not even close to 10us time aligned. Yet, Harman insists that this kind of distortion is benign.

I don't like this. I think it's audible.

What you are looking at is a consequence of the phase shifts introduced by passive crossovers between drivers.

I don't remember seeing a single blind test which proves audbility of such phase shift - on the contrary, I remember seeing quite a few of them which proved it is not audible in "normal" listening rooms. I suspect situation may be slightly different in the rooms with little or no reflections but I am not sure such tests have been performed.

Btw, this kind of phase shift can be easilly corrected via EQ so why not trying it for yourself?

 

[Music1969]

Hi @mansr

Apologies if this has been answered so far (I'm slow!).

But let's say you start with Redbook music content (with timing resolution of say ~24ns from your blog ) and then upsample that to 1 bit 11.2896 MHz SDM (DSD256 rate)

Of course no timing resolution can be gained but what is the mathematical loss with this particular upsampling?

Or is the loss completely negligible?

Using the same 1 kHz signal at -20 dB

Not asking about music recorded at DSD256 rate (I think some calcs were shown yesterday about that) but specifically what timing resolution is lost with upsampling to DSD256 rate

 

[mansr]

What you are looking at is a consequence of the phase shifts introduced by passive crossovers between drivers.

I don't remember seeing a single blind test which proves audbility of such phase shift - on the contrary, I remember seeing quite a few of them which proved it is not audible in "normal" listening rooms. I suspect situation may be slightly different in the rooms with little or no reflections but I am not sure such tests have been performed.

Btw, this kind of phase shift can be easilly corrected via EQ so why not trying it for yourself?

In that context, this might be relevant: https://ieeexplore.ieee.org/document/9450008

 

[thorvat]

In that context, this might be relevant: https://ieeexplore.ieee.org/document/9450008

Eh, not really. That paper discusses GD equalization audibility but equalizing crossover phase shift doesn't really affect GD as passive XO phase shifts cause phase deviations with mild slopes so GD should look quite good even without the EQ.

 

[PierreV]

Thread reply ban issued.

??? Not to @j_j I hope.