Subharmonics of ultra high frequency information...

amirm said:

That is not true as a matter of math Jack . As long as we have twice as many samples as we have bandwidth, we can fully represent the signal. More samples are redundant and don't contribute to proper reconstruction.

What we can do with more samples and hence bandwidth is to have more variations in reconstruction filter design. Fore example we could have a slow ramping filter close to hearing range and then a sharper one way above. That way the ringing for the latter won't impact the audio range. So in that sense, we can help the lower frequencies but it is not due to having more samples per-se .

Click to expand...

i've already written 2 other posts and decided to delete them trying to avoid anything too incendary.

but i guess i just need to be blunt. after reading all this fine theory my question would be why does this whole sampling stuff sound so 'non-wonderful'? that is when comparing it to it's analog source.

the idea that 'more samples are redundant' after listening to redbook compared to analog seems silly.

is this an issue of yet to be achieved execution of concept......that we just cannot quite do what in theory we should be able to? if so i guess i could then understand. but if we are saying that 'more samples are redundant' and we are already at optimal execution then Nyquist must be wrong. if there are enough samples then something else must be missing (or gets lost) in the ADC (and/or DAC) process.

maybe we can find it and put it back.

Higher sampling rates do have a benefit. It involves the filter.

Mike Lavigne said:

i've already written 2 other posts and decided to delete them trying to avoid anything too incendary.

but i guess i just need to be blunt. after reading all this fine theory my question would be why does this whole sampling stuff sound so 'non-wonderful'? that is when comparing it to it's analog source.

the idea that 'more samples are redundant' after listening to redbook compared to analog seems silly.

is this an issue of yet to be achieved execution of concept......that we just cannot quite do what in theory we should be able to? if so i guess i could then understand. but if we are saying that 'more samples are redundant' and we are already at optimal execution then Nyquist must be wrong. if there are enough samples then something else must be missing (or gets lost) in the ADC (and/or DAC) process.

maybe we can find it and put it back.

Click to expand...

Eloquent and fairly stated, Mike. The theory is solid, the execution is suspect in my book... certainly at times. When we have experienced high-level recording engineers (who are members on the forum) stating that higher-than-redbook sampling rates produce audibly different results, then we should at minimum be curious. When we had such a marvelous technology as the Cray supercomputer that allowed men to walk on the moon, why did we go further to develop the tiny laptops of today (with far superior performance)... wasn't it "good enough" to have that much ability? Wasn't sending a man to the moon a sufficiently impressive level of performance? Many would say it's good enough. This is the What's Best Forum, so shouldn't we all expect these discussions to be foundations of a "search"?

It's just that the money isn't in audio engineering. Only a handful of perfectionists exist in this specialized pursuit. If the market desired better audio, then it would have seen the same focus, development, and competition as the computer chip race.

Understand what we have, but always strive for better. That's my opinion. I love digital, but I want it to be better yet. High-res is a step toward the answer for many listeners.

Rant over.

Lee

Mike Lavigne said:

i've already written 2 other posts and decided to delete them trying to avoid anything too incendary.

Click to expand...

Appreciate that .

but i guess i just need to be blunt. after reading all this fine theory my question would be why does this whole sampling stuff sound so 'non-wonderful'? that is when comparing it to it's analog source.

Click to expand...

You are reading something into the discussion which does not exist. The discussion was not analog vs digital. Nor was it about whether digital is perfect or not.

What was said that having more samples than Nyquist helps us recover those low frequency signals better. As a matter of provable science, that is not correct. Digital has issues. But lack of samples is not it in the audible range. You can have 1000 times more samples and it would sound the same if your filter your source the same way.

the idea that 'more samples are redundant' after listening to redbook compared to analog seems silly.

Click to expand...

That is not the comparison. The comparison is that if you had 44,100 samples or 88,200 samples, in an *ideal system*, would it make a difference from reconstruction of said signals. The math simply and clearly answers this. There should not be any layman notion that somehow those extra samples fill the gaps when they do not.

This says nothing about the entire digital system being transparent.

but if we are saying that 'more samples are redundant' and we are already at optimal execution then Nyquist must be wrong. if there are enough samples then something else must be missing (or gets lost) in the ADC (and/or DAC) process.

maybe we can find it and put it back.

Click to expand...

Let me put it this way.

Lets say the ultimate fidelity can be described as F = X + Y + Z + W

My view of digital is that we know X & Y. And we can prove that to be the case (purpose of my posts in this thread). When mathematics can prove something, we cannot doubt it. We must believe it or we lose credibility . The measured frequency response of a digital system is superbly good. Its distortion superbly low. If the math was wrong, these measurements would show otherwise.

Then again, when mathematic cannot answer something, then it is open to debate. That get us to "Z & W." If you ask Ethan, he will say there is no Z & W.

If you ask me, I would say Z & W exist but their values are quite small. They exist as a minimum because we know that our implementations of digital cannot and by definition so, ever be perfect. For example, there is no way to perfectly filter the signal before and after digitization. We know the timing will never be perfect either. I also accept that there are some unknowns here although I cannot prove to others that they exist.

Going full circle, this latest discussion was about X & Y. In that regard, Ethan is right. This is a myth and should not be used as an argument point. If you want to argue digital vs analog, you owe it to yourself to use the right arguments. Needing more samples to reproduce the lower frequencies (in an ideal sense) is not true.

I think it is pretty clear now why I was confused Amir. I'm coming from the AD side and you are coming from the reconstruction side. So I'm back to square one in my suspicions that higher sampling rates will yield better resolution even within the hearing band.

Let me backtrack and say why I think it is the case. I used an AVP2 for many years that AD'ed everything. The AVP2 acting as a DA converter off of SPDIF was very good at the time, anything AD'ed did not. Fast forward to late 2009 and I did my first foray into DRC. Much higher sampling rates (both still 32bit but I don't know if the Proceed was actually configured to take advantage of the SHARC chip but actually used lower bit chips for AD) and this time whatever analog came in on a flat curve (that sounded weird) vs bypass was at least to me indistinguishable. Now why would that be? Why did the Proceed AD conversion sound like it made everything up of little square blocks of sound and the RP-1 didn't?

amirm said:

Our library has a ton of articles on this front guys . This research while disputed, does talk to possible cause and effect (i.e. it is the ear whose response changes due to hypersonics): http://www.linearaudio.nl/Documents/high freq inpact on brain.pdf

Hypothetical explanation of neuronal mechanisms of the hypersonic effect

Click to expand...

For those interested in this topic, I had a look at publications relating to ultrasonic hearing and the hypersonic effect and have prepared a write-up.

Klaus

Tim according to Amir-"That is not true as a matter of math Jack . As long as we have twice as many samples as we have bandwidth, we can fully represent the signal. More samples are redundant and don't contribute to proper reconstruction.

What we can do with more samples and hence bandwidth is to have more variations in reconstruction filter design. Fore example we could have a slow ramping filter close to hearing range and then a sharper one way above. That way the ringing for the latter won't impact the audio range. So in that sense, we can help the lower frequencies but it is not due to having more samples per-se ."

To make it work you have to control the bandwidth. That means you have to filter out any content above your chosen bandwidth.

Okay. Can someone please illuminate me here because I'm going nuts.

If the job of the reconstruction filter is to smooth step response, what is the determinant of the gap between steps themselves?

Jack I am already over my head. I just follow what Amir says.

Me too, and drowning. Good idea to wait for Amir and Don.

Lucky. I kept my notes form last semester. See the PDF file in the first post. http://www.whatsbestforum.com/showt...d-street-sense.-or-when-is-close-close-enough

amirm said:

If you want to argue digital vs analog, you owe it to yourself to use the right arguments. Needing more samples to reproduce the lower frequencies (in an ideal sense) is not true.

Click to expand...

Amir,
But audio is not ideal - ADCs have quantification errors, independently of the number of bits you have. Most recording systems add analog dither (or simply have analog noise that does a similar effect) when digitizing. If you spread your errors between more values, you do it in a less systematic way and (may be?) you can improve sound in the bass frequencies.

You are reading something into the discussion which does not exist. The discussion was not analog vs digital. Nor was it about whether digital is perfect or not.

What was said that having more samples than Nyquist helps us recover those low frequency signals better. As a matter of provable science, that is not correct. Digital has issues. But lack of samples is not it in the audible range. You can have 1000 times more samples and it would sound the same if your filter your source the same way.

Click to expand...

what set me off was Ethan's 'often debunked' comment referring to his viewpoint about people who say digital is incomplete. i addressed my comments to your more 'focused' post since i wanted a more 'Amir-like' response.

thanks.

i realized that my comment was not exactly on topic but throwing around the idea of digital having 'enough samples' without specifying that there was maybe more to music reproduction than that needed to be noted.

and your point about Z & W answers my concerns and the idea that Z & W is unfamiliar to some is obvious.

Gregadd said:

Lucky. I kept my notes form last semester. See the PDF file in the first post. http://www.whatsbestforum.com/showt...d-street-sense.-or-when-is-close-close-enough

Click to expand...

Someone must have been an honor student. Yeah, that's the thread Amir was looking for.

Post #16 Had the teasers. Time for the deeper dive.

Hi

I am still waiting for the answer about tone discrimination. Can we, reliably, distinguish between High Frequency tones above 10 KHz, with same level (SPL if you will), same frequency but different harmonics content? IOW do ALL 10 KHz regardless of wave shape sound the same to us?

Frantz,
as long as you have a computer, you can use this to determine for yourself:
http://www.nch.com.au/tonegen/index.html
It's shareware, which means that it's free, but they will try to get you to pay for it if you find it useful.

You can save the files to WAV and burn a CD, and listen to the various tones. I can very reliably hear the difference of the sine, square, triangle, sawtooth, etc. waves at 10kHz. Be careful if you play this on your computer because it may cause the speakers and built-in audio to distort if you use too high a frequency.

I will give a brief answer on the filter issue. Maybe Don can expand and/or write a new article on it .

A quick intro. Sampling theory only works if you filter any and all components above half the sampling frequency. Any components left there create "aliasing" which in simple language means extra frequencies that should not be there. In a texbook, we can draw filters that are absolute. In case of a CD playback at 44.1KHz, that filter can be straight up and down step function at 22.05 KHz. In reality, we can never make such a filter. Remember that since creating a squarewave requires infinite bandwidth and hence not doable, creating a filter that has that kind of cut off is also an impossibility.

If we take into account that we only need 20 KHz out of that 22 KHz worth of bandwidth, we can construct a filter that starts at 20Khz and starts doing its thing. That makes the filter more doable but still not quite possible. Such a filter in typical configuration in IC DACs still doesn't have zero output at 22 Khz. Some amount is likely to exist. If we further assume that we can't hear those components since they are > 20 Khz, it seems that life is well. Alas, that is not meant to be.

When you play sound in the above system, the "aliasing components" above 20 Khz mix with the high frequencies in your music. When you mix two signals, you get the original signals plus their sums and difference. If there is a component at 23 KHz, it will mix with something at 19 Khz and create a new component at the difference of the two which is 4 Khz!!! So it is possible to now hear such a distortion since its frequency is much lower. This distortion is called Aliasing Intermodulation Distortion or AID for short.

You can minimize AID by starting earlier and sacrificing some of the your 20 Khz bandwidth.

In addition to AID, there is an issue of "ringing." The sharper a filter, the more its response varies in amplitude in the spectrum it is not supposed to touch.

Now let's assume that you have a 96 Khz sampling rate but still go to the school that says we do not hear anything above 20 KHz. You can build a DAC with dual filters. The first one can be a special type of filter which does not ring. Such a filter will not have the slope that we need to cut off all the aliasing components (i.e. above the 48 Khz bandwidth of 96 Khz sampling). But since we don't care about preserving all of that extra bandwidth, we can cascade a sharper filter which starts way later than 20 Khz but still manages to cut off everything at 48 Khz. Put another way, this system has a response that is flat to 20 Khz, then gradually rolls off toward zero and at some point, sharply goes to zero (or very close to it).

Think of it as a car with two brakes. One that is gentle but not super effective. And one that is very effective but shakes your car like crazy. If you use the gentle brake most of the way and only at the last minute use the strong brake, you get the benefits of both. Same is true here.

Alas, when someone makes a chip DAC (what is used in typical AVRs, mass market CD players, etc.), they try to comply with the "spec." The spec for a 96 Khz sampling says it must have a flat response (or close to it) up to 48 Khz. So they attempt to preserve the full bandwidth and put the same sharp cut off filter at the end which causes ringing. Discrete/dedicated DACs sport custom filters that deploy the above techniques and hence can minimize the above factors.

Note that there is a trade off. As Don and I have noted in the past, when you run a DAC faster, it simply cannot do its job as well. The system will have more noise and accuracy drops. So there is no free lunch here and the world is full of trade offs of this sort .