Understanding Sample Rate

[mansr]

9 minutes ago, adamdea said:

True and I agree about the physics of sound and the spectrum of real musical sounds. 

Either way, the knowledge required comes from outside the sampling theorem itself.

The difference is that this outside knowledge is easier to acquire and validate than that concerning our ability to perceive sounds.

[mansr]

47 minutes ago, adamdea said:

Possibly as regards bandwidth- but just as easy to ignore if one takes the fluctuations of one's listening experience and one's personal interpretation of the meaning of those experiences as the only critical data, alongside a general faith in progress.

The way I see it, the goal of audio engineering should be to accurately reproduce the stimulus a listener is subjected to, i.e. the sound waves entering his or her ears. The study of perception belongs to the domain of neuroscience (or something similar). The only way it relates to the engineering is when its findings allow engineers to take shortcuts. A prime example would be the discovery of two-channel stereo and its ability to create the illusion of arbitrary left/right positioning without using a speaker for every possible direction.

 

Quote

At one point it looked as though DXD might become a de facto "this far and no further" standard. But forget it. Limits are offensive and if you play someone a dxd file and a 32/ 768 file pretty much the same group of people who can hear the difference between 16/44 and 24/96 will claim to hear the improvement in the 32/768. 

Don't I know it.

 

Quote

When first came to this hobby I assumed that the 16/44 spec must have been cobbled together as a result of the limits of technology in the early 80s and that it made sense that it must be easy to improve on it now. After all I remembered what computers were like in 1982 and assumed that analogies with video were closely applicable. It took me a long time to "unlearn" that.

The process leading to the specific design of the CD is interesting. The rough requirements were determined by the already well known limits of human hearing, both in frequency and dynamic range. Sony and Philips argued a bit over whether to use 14 or 16 bits, mostly because one of them had a 14-bit DAC chip and the other had a 16-bit one. The choice of precisely 44.1 kHz as sample rate was made because existing equipment could easily be made to handle it. The physical design of the disc (laser wavelength, track spacing, etc.) was mostly based on what could be manufactured at a reasonable price. As for the size of the disc, one story has it that it was chosen such that the longest known recording of Beethoven's 9th symphony would fit.

[mansr]

12 minutes ago, sandyk said:

Not quite , at least in the case of some of Barry Diament's recordings with genuine musical content to 57kHZ

How do you know it's musical if you can't hear it?

[mansr]

23 minutes ago, sandyk said:

 So is the BS that several here appear to be pushing about 16/44.1 being all that's necessary for perfect stereo reproduction. 

 Incidentally, years later Elektor magazine published a design intended to correct the phase related problems of the original Sony players.

The early Sony players multiplexed a single DAC for both channels resulting in one channel being delayed by 11.3 μs. That doesn't mean the format itself is insufficient.

[mansr]

21 minutes ago, sandyk said:

 Bone conduction is one of the reported mechanisms for noticing HF above the normal hearing range.

Bone conduction doesn't work for airborne sounds.

[mansr]

4 minutes ago, Ralf11 said:

some grist for the mill...

 

http://jeb.biologists.org/content/215/2/ii

Note the sound levels involved, upwards of 80 dB.

[mansr]

13 minutes ago, esldude said:

And did they not have an analog bit to fix that delay at least mostly?  Or am I mis-remembering that?

The portable D-50 has no such feature. That's the one I have measured.

[mansr]

1 minute ago, esldude said:

I was thinking of the first full size Sony CD player. 

I've never got my hands on one of those. Feel free to send me one.

[mansr]

21 minutes ago, Don Hills said:

I have a review of the D-50 where Martin Colloms measured the phase shift between channels to be 71 degrees at 20 kHz. I would have thought it would be approaching 180 degrees if there was a full 11.3 us delay between channels.

Martin Colloms isn't exactly trustworthy, but I'll check it again.

[mansr]

8 hours ago, sandyk said:

He is at least as trustworthy as you are, and far more well known than an anonymous E.E. like yourself.

I am hardly anonymous.

[mansr]

9 hours ago, Don Hills said:

 

"... the usual phase response of a shared converter player was shown. This produced a negligible phase difference at low and middle frequencies, rising to a maximum of 71 degrees at 20 kHz. As yet this has not been confirmed as being of any audible importance, unless the channels are mono-ed. ..."

- HFN&RR, March 1985.

I don't have the rest of the issue, so I could be wrong about the tester being Martin.

I measured the D-50 at 10 kHz and 20 kHz.

 

 

 

The channel offset at 10 kHz is 11.3 μs or half a sample period. At 20 kHz, it is a little lower at 10.1 μs. The schematic shows nothing external to the DAC that would explain this, so it is probably a property of the chip itself.

[mansr]

7 hours ago, Em2016 said:

Any issues with increasing clock phase noise, as you go to DSD256 and then higher to DSD512?

The crystal oscillator is ~22 MHz (512*44100), and the required bit clock is divided from that. The absolute jitter level is thus constant.

 

7 hours ago, Em2016 said:

I guess it depends on the clock used of course but do we know about clock phase noise performance of the iFi micro DACs? @jabbr

The clocks seem pretty good to me. The weak part of the iFi DACs is the analogue output section.

[mansr]

31 minutes ago, Don Hills said:

Thanks, Mans. That is indeed about 72 degrees at 20 kHz. So Martin (or whoever) did measure correctly. The confusion was all mine, I mixed up sampling rate and frequency.

Well, it never hurts to double-check.

 

31 minutes ago, Don Hills said:

Edit:

Your D-50 still works? Last time I checked, so did mine.

It did today. Granted, it hasn't seen much use the last 30 years. This particular unit was manufactured in 1984. Good build quality.

[mansr]

57 minutes ago, Em2016 said:

 

Thanks jabbr and @mansr

 

The background to my question comes from Andreas Koch and Ted Smith.

 

Andreas Koch's comments here: https://positive-feedback.com/audio-discourse/raising-the-sample-rate-of-dsd-is-there-a-sweet-spot/

 

"Double DSD seemed to "fix" above mentioned compromise that was committed with single DSD by moving the noise shaper from 20kHz up to around 40kHz, well above our standard audio band of 20kHz. It also allowed for gentler and simpler output filters on the DAC. Life became a bit easier and every time that happens in audio we can expect better sonic performance. That is clearly the case with double DSD, and the price of double the data rate seems well worth it.

 

With a clear improvement from doubling the sample rate of single DSD, it seems natural and, of course, tempting to quadruple the sample rate or go even higher. We should expect the same or similar improvement as from simply doubling the sampling rate again, right?

 

Not so fast! It turns out there are physical limitations such as electronic component speeds, finite clock slopes, etc., that limit the amount of performance gain we could expect from raising the sample rate above a certain threshold. Such limitations are indeed starting to affect performance with quad DSD in D/A converters. Side effects in the form of audible noise and distortion are creeping into our audio band with quad DSD, greatly overshadowing the small benefit of the noise shaper curve starting at 80kHz. This, of course, is only apparent in D/A converters that convert the DSD signal directly into analog without any conversion to PCM or other digital filtering. To avoid this the quad DSD signal would have to be low-pass filtered and/or converted to PCM before converting it to analog."

 

Ted Smith seems to hint at similar here:

 

http://www.psaudio.com/forum/directstream-all-about-it/questions-for-ted-about-upsampling-and-fpga/page-2/#p71776

 

"Ultimately you need to actually do the conversion from digital to analog – unlike digital processing where you (more or less) get to define your own universe that final piece of hardware is restricted by physics, cost and other real world constraints.  In particular the higher frequency your clock the more noise from jitter.  As the clock frequency goes up the noise can go down because you do more noise shaping (trading noise floor in the audio band vs noise in the high frequencies).  You get 3dB more S/N over the audio band for each doubling of the sample rate.  On the other hand the noise from jitter grows as the sample rate increases as well.  The final noise floor ends up depending on the number of bits in each sample, the oversampling ratio and the jitter.  The optimum sample rate doesn’t depend much on the number of bits in each sample but for one bit audio it’s between the sample rates of double and quad rate DSD (closer to double rate DSD.)  If I go to buy (at any price) better clock crystals the jitter (in particular the low frequency phase noise) goes up with higher frequencies.  I’d like to use a 16 FS clock (45.1584MHz) but it has more noise that the 22.5792MHz clock in the DS.  Also the bandwidth of the digital switches I use is fixed and the third harmonic distortion goes up as I increase the sample rate…"

 

Is there something/s they are measuring that others haven't/can't (access to better equipment)?

What they say makes sense and agrees with my own experience. The optimal rate depends on the DAC. In the case of iFi, their weakness is a fairly high degree of intermodulation distortion in the 80-120 kHz range (roughly speaking). If the DSD noise isn't pushed clear beyond that region, the noise floor in the audible range rises. For this, DSD256 is required.

[mansr]

Just now, Em2016 said:

Thanks mansr but the way I read what both said, something between DSD128 and DSD256 , but closer to DSD128, is the sweetspot.

That may be, but it's not an option on any DAC I know of. 

 

Just now, Em2016 said:

My bigger question then is are we enjoying more noise/distortion with DSD512 upsampling?

You'd have to measure your DAC and check how it performs.

 

I do think AK and TS somewhat overstate the problems with higher clock rates. Modern electronics handle several gigahertz with ease, and here we're only dealing with 100 MHz at most. Phase noise (jitter) is far from the only relevant parameter. If moving the noise up in frequency makes the filters perform better, perhaps overall performance is improved even if there is slightly higher jitter.

[mansr]

1 minute ago, Em2016 said:

But you/we have the option to choose DSD128 (instead of DSD256 and DSD512).

And that's worse on iFi DACs.

[mansr]

2 minutes ago, sandyk said:

I am well aware of the Forum's requirements,

Yes, you manage to stay just on the right side of the line.

 

2 minutes ago, sandyk said:

however it appears to be quite O.K. to ridicule and insult both myself and M.C. at every opportunity.

No more often than you ridicule and insult all engineers. Seems fair.

 

2 minutes ago, sandyk said:

Is it any wonder why we rarely see informative posts any more by leading industry figures ,?

Depends on your definition of informative. And of leading. And industry. And figure.

[mansr]

7 minutes ago, Ralf11 said:

I think it is fine to ridicule and insult engineers as a group - same for the English.

It's tiresome nonetheless.

[mansr]

4 hours ago, Fokus said:

A bit off topic, but while MC has a bit of an ego, the man also has some achievements. He co-founded Monitor Audio, later had Colloms Electroacoustics, a consultancy company active in design and forensics, IIRC. MC was responsible for a number of speaker and amp designs in the 80s and 90s, including Celestion and Aura. These are just the things I happen to know, as a relative outsider.

None of those things make him an authority on digital data storage and transmission.

[mansr]

1 hour ago, Spacehound said:

I don't think people other than criminals  'deceive' deliberately.

Politicians are a subclass of criminals, right?

[mansr]

1 hour ago, Spacehound said:

I see you hve moved the goalposts  from "Who copied who?" to law, patents, and other irrelevant stuff.

Copying is permissible unless prohibited by copyright, patent, or trademark. That makes those things relevant.

[mansr]

4 hours ago, Spacehound said:

What British hifi industry?  At a quick 'offhand' look there is   only Arcam (partly), Linn, and Sugden.

Bowers & Wilkins?

[mansr]

11 minutes ago, Spacehound said:

Not to Linn Engineering (Linn Audio didn't exist)  copying  an earlier and still current  turntable they made some parts for under contract, which was the subject at the time.

Unless those parts were covered by patent or copyright, they did nothing legally wrong.

[mansr]

7 minutes ago, Spacehound said:

Since 2016 they are  a subsidiary of EVA Automation of California. 

AFAIK the engineering is still done in Britain, as is the manufacturing of the 800 series.

[mansr]

19 minutes ago, Summit said:

To understand the effect of different sample rates that @beerandmusic have been asking about, it’s not enough to quote Nyquist-Shannon Sampling Theorem and think that as long as the samples are above that rate nothing else are of importance. Just try to play DSD at those sample rate and see. Its more, much more premises that comes in to play if we are looking to reproduce the sound as close as possible in SQ to the master tapes.

The sampling theorem assumes infinite sample precision. With a limited number of bits, we get quantisation error which is a non-linear distortion. This happens at any sample rate and has nothing to do with aliasing. With TPDF dither, the quantisation error is (mostly) turned into white noise at a level determined by the bit depth. If the sample rate is increased, the quantisation energy is spread over a wider frequency range, thus reducing the level of dither noise at any specific frequency. Doubling the sample rate gives the same improvement as extending the sample precision by one bit. In other words, not very efficient. A high sample rate does, however, bring another benefit in that it enables the use of noise shaping. Instead of the quantisation noise being spread evenly over the full spectrum, it can be concentrated at high frequencies where there is no signal of interest. This is very useful since a high-rate flash ADC with a small number of bits, say 8 or less, is much easier to construct than a slower ADC with high precision. Low-resolution noise-shaped sampling at a high rate, 10 MHz or more, followed by a digital low-pass filter can thus be functionally equivalent to high-resolution sampling at a lower rate. In fact, it can be better since there is no need for an analogue anti-aliasing filter, and the digital filter can be designed with just about any characteristics we desire.