Are we just kidding ourselves?

[wgscott]

Surprise, Surprise.

But you went into this, not expecting to hear any difference, didn't you ?

If someone else said they heard a difference, you would be likely to put down to the same reason ,(Expectation Bias) wouldn't you ?

 

When hearing differences that can't be measured, the burden of providing evidence falls on those who make the extraordinary claim.

[Jud]

 

Well it wasn't directed at John's comment. Just the topic in general. Seems all these worried about filter effects are ultrasonic. They are real enough, but inaudible. So my proposed procedure was just seeing what a lower frequency digital brick wall left vs one at 4 times the sample rate. I repeated it by running it through a different series of gentler filters 3 times. Same result with none of the residuals being audible.

 

I also generated some digital single sample impulses. The kind that cause the nasty looking graphs with pre and post echo covering in extreme cases a few milliseconds. I subtracted the pure digital impulse from the same signal which had undergone brickwall filtering. So all that was left was filter effects. You could see it in an FFT, you could see it in the waveform. You couldn't hear anything. I then slowed down the speed of playback. Same file just played at 50% and then 25% normal speed. When you do that you can hear the results of the filter as a high pitched momentary 'tick' sound. Again because it appears those effects of filtering are pushed to an ultrasonic region. Where in normal playback they are a non-issue.

 

Ah,, OK, got it. Yes, I wouldn't expect the difference file to contain anything but ultrasonics. Could you do your "bad filter" experiment with a short segment of something like cymbals, then do a quick ABX with Foobar on the ringing version and the original?

 

Edit: I'll likely have some time tomorrow if you want to send me the two files to see if I can hear any difference.

[sandyk]

When hearing differences that can't be measured, the burden of providing evidence falls on those who make the extraordinary claim.

 

Here we go again.

Expecting proof to be provided by those who are unable to provide it, due to not being suitably qualified in that area, or not having access to suitable test equipment, or the knowledge of how to use it, is plain arrogant, and holds back progress, no matter what the field of human endeavour is.

It's a bit like an Atheist such as yourself, demanding proof from a Christian ( or Muslim etc.) of "life after death".

[esldude]

Ah,, OK, got it. Yes, I wouldn't expect the difference file to contain anything but ultrasonics. Could you do your "bad filter" experiment with a short segment of something like cymbals, then do a quick ABX with Foobar on the ringing version and the original?

 

Edit: I'll likely have some time tomorrow if you want to send me the two files to see if I can hear any difference.

 

Yes, it could be done. Haven't done it. I have used cymbals at 44 khz and 176 khz sample rates. With the 44 upsampled to 176. Not discernible via Foobar ABX by me.

[wgscott]

Here we go again.

Expecting proof to be provided by those who are unable to provide it, due to not being suitably qualified in that area, or not having access to suitable test equipment, or the knowledge of how to use it, is plain arrogant, and holds back progress, no matter what the field of human endeavour is.

It's a bit like an Atheist such as yourself, demanding proof from a Christian ( or Muslim etc.) of "life after death".

 

I'm in total agreement. It is exactly like religious beliefs.

[Jud]

When hearing differences that can't be measured, the burden of providing evidence falls on those who make the extraordinary claim.

 

Here we go again.

Expecting proof to be provided by those who are unable to provide it, due to not being suitably qualified in that area, or not having access to suitable test equipment, or the knowledge of how to use it, is plain arrogant, and holds back progress, no matter what the field of human endeavour is.

It's a bit like an Atheist such as yourself, demanding proof from a Christian ( or Muslim etc.) of "life after death".

 

I suspect Bill would happily agree with you on that last point - are you sure that's what you want?

[beanbag]

Well, you could try this. Take your nicest 192khz recording. Duplicate it in Audacity or similar sound editor. Invert the second copy. EQ the first copy so it has a brickwall digital filter at 20 khz. Mix the two and see what is left. Of course everything above 20khz will be there along with any of the transient killing effects of filter ringing. Play the result and see if you can hear anything.

 

You can use a filter with many more taps than is typically found in hardware DACs so the transient smearing effects should result in extreme results.

 

Do be careful with the EQ in Audacity, it is easy to slide the line so you get a gain change you didn't intend. This will result in residual signals. There is a small trick to prevent that.

 

All the effects of the filter ringing is in the transition band of 20-24 khz, and were just as audible as those in the 80-96khz transition band which is to say I heard nothing.

 

This seems like a good first test to try, but I am not sure if the results will be conclusive.

 

For example, if you did hear something, then I could also claim that in the presence of real musical signals, it would get masked.

 

If you didn't hear anything, I could always claim that maybe the hearing mechanism works by your little receptors first needing to get charged or loosened up by some ultrasonic ringing, before the <20 kHz "main event", and then they will send a signal to the brain that some twig snapped somewhere. Whereas if you took away the "main event", the receptors might say "oh never mind" and then you don't hear anything.

 

Just making stuff up.

 

My current theory after an extensive literature search is that ultra sonic frequencies cause relaxation of the masseter muscles via bone conduction, leading to the commonly known effect of "jaw dropping". This physiological (as opposed to audible) response varies among different people.

[sandyk]

I suspect Bill would happily agree with you on that last point - are you sure that's what you want?

 

Hi Jud

Why not?

Without help from someone from "the other side" these things can't be verified .(grin)

 

Regards

Alex

[kumakuma]

My current theory after an extensive literature search is that ultra sonic frequencies cause relaxation of the masseter muscles via bone conduction, leading to the commonly known effect of "jaw dropping". This physiological (as opposed to audible) response varies among different people.

 

"jaw dropping" is better than what happened to Alex C. when he installed Maverick on an SD card:

 

The results of my experiment this afternoon compel me to start another thread.

 

… SNIP …

 

h) Played tracks;

i) Soiled my pants.

 

… SNIP …

 

[kumakuma]

Hi Jud

Why not?

Without help from someone from "the other side" these things can't be verified .(grin)

 

Regards

Alex

 

[sandyk]

I could always claim that maybe the hearing mechanism works by your little receptors first needing to get charged or loosened up by some ultrasonic ringing, before the <20 kHz "main event",

 

Funny you should say that. Years ago, after several hours listening to a friend's Ohm Walshe speakers with Plasma Tweeters", on the drive home my ears seemed overly sensitive to even coughing.

[esldude]

This seems like a good first test to try, but I am not sure if the results will be conclusive.

 

For example, if you did hear something, then I could also claim that in the presence of real musical signals, it would get masked.

 

If you didn't hear anything, I could always claim that maybe the hearing mechanism works by your little receptors first needing to get charged or loosened up by some ultrasonic ringing, before the <20 kHz "main event", and then they will send a signal to the brain that some twig snapped somewhere. Whereas if you took away the "main event", the receptors might say "oh never mind" and then you don't hear anything.

 

Just making stuff up.

 

My current theory after an extensive literature search is that ultra sonic frequencies cause relaxation of the masseter muscles via bone conduction, leading to the commonly known effect of "jaw dropping". This physiological (as opposed to audible) response varies among different people.

 

For sure I didn't think this test is conclusive. I didn't hear anything. As to regarding masking if someone heard something post echo almost surely is well masked. Pre-echo is not well masked, and according to some data with a bit of experience it may not be masked at all. But thus far I don't hear anything.

 

The supposition that ringing at ultrasonics only gets registered if it is followed by a larger lower frequency sound, well one might suppose anything. Is there some known mechanism in the ear for that to happen? The ear doesn't physically respond to ultrasonics. There are ways to test for that. Don't know of much research to support the idea.

 

And one can decide to chase rabbits down a hole as long as they wish. But the significance of such even were it to occur vs the size of known problems in current loudspeakers and headphones is simply dwarfed by the speaker and headphone issues Mayhem13 has mentioned.

[beanbag]

"jaw dropping" is better than what happened to Alex C. when he installed Maverick on an SD card:

 

I thought "brown note" effects were sub-sonic?

[Fokus]

Pre-echo is not well masked, and according to some data with a bit of experience it may not be masked at all.

 

Pre-ringing can be clearly audible if it is positioned in the mid-range, for instance with a low-pass at 4kHz, or an all-pass (although in that case one wonders if anything heared is the result of the ringing, or of the phase shift; one could even wonder if the two are not actually one thing), and if the filter slope is steep enough, so that the ringing's duration exceeds a certain limit.

 

This limit can even be calculated. As you know the cochlea can be seen as a continuum of band-pass filters, with width (ERB) of about 1/3 octave. The inverse of an ERB's width has the dimension of time, and is a measure for the time a particular ERB needs to fully react to a stimulus (you see where I am heading?).

The ringing has to be confined to less than this time.

 

Further, if the experiment is redone with increasingly higher filter cutoff frequency, it is known that the ringing audibility disappears at a certain (listener dependent) frequency. For most living and breathing people this is below 20kHz. Even then one can always widen the filter's transition band to about one ERB, in which case filter and cochlea are equally fast or slow.

 

 

Flying high above all of this is Miska, who simply insists that a filter response be confined to one cycle of 20kHz. That's another approach. One would call it overoveroverkill ;-)

[Fokus]

Of course, you are simplifying here more than a little. Something like a simple transport delay is essentially linear, and delays all frequencies equally. That is not the case with PCM reconstruction,

 

That is exactly the case with orthodox, linear-phase PCM reconstruction.

[Fokus]

Quote from this article:

 

"In lab tests, people can distinguish between sounds as little as five milliseconds apart".

 

Depends on what is meant by 'distinguish between sounds'.

 

Anyway, the temporal acuity of the auditory system is more like 5 to 10 microseconds. That is not a secret. It also does not pose a problem to CD.

[Fokus]

Or Fourier and harmonics...

 

OK, obviously something I need to understand more about. What's all the stuff in the various journal papers about the great difficulty of modeling non-sinusoidal sounds?

 

The papers are about modelling sounds. As said above, this is taking a limited number of relatively simple base functions, combining them in a way that approximates an original, arbitrarily complex, function. The example with the mountains is a good one: given ten triangles, can you replicate the K2 simply by scaling, shifting, and adding these triangles?

 

Sound recording is not modelling. Back to the mountains: if you want a representation of the K2, simply take a picture of it with a suitable camera.

 

I hope you see the difference?

 

 

As for Fourier: what matters in our context is the transform, which 1:1 maps any integrable function (which encompasses just about everything bar the big-bang) onto another function based on the integration of a continuum of sines. Integration and continuum are key here. This is emphatically not the same as a finite sum of sines.

In the case of audio the first function is amplitude versus time, and the second one is then complex amplitude versus frequency. The transform is used for reasoning about signals. It is generally not executed when recording or playing sound, and there is no need ever to calculate it for recording and playing sound. Knowing that it exists, and knowing its limits (i.e. 'bandwidth') is sufficient.

 

This is very fundamental to all of this here. And yes, many are still confused. The underlying reason is that for periodic signals the Fourier transform simplifies to a sum of sines, which is called the F series and tought in school.

 

 

The Fourier limit or Gabor limit is a mathematical rule (i.e. a natural law) that puts a ceiling to the accuracy of concurrent knowledge one can have about frequency and time of a given, arbitrary, unknown signal. It is equivalent to Heisenberg's uncertainty in quantum mechanics.

 

There is an article that titles that the auditory system breaks the Fourier limit. But then this limit does not apply here at all: the subjects had to order two known sonic pips in time, and they knew that had to do this before the test started. The title of that article is, depending on you attitude to such things, either sensationalist or tongue-in-cheek.

[Fokus]

So is (to use the description from the IEEE article abstract) "highly nonstationary inharmonic behavior" getting close to "pathological"?

 

No

 

I think I'm correct in saying at least a good portion is non-periodic.

 

Of course. Most interesting sounds are non-periodic. And the more interesting portions of seemingly-periodic sounds are non-periodic. This is not a problem at all.

 

Take an infinitely fast infinitely narrow positive going impulse. What is its Fourier transform?

[Ned Kelly]

I can only offer this - the Hugo gives me almost as much joy listending to MP3s, iTunes downloads and YT vids (ok - some YT vids) as it does listening to the few 24/96 albums I've managed to download. I did download a single DSD sample track, but at over 250MB for a track that simply isnt going to work for me. I could buy more SACDs and play them with my Oppo, but thus far I've yet to hear a single SACD that has blown my mind to the extent that I'm prepared to spend the money - Redbook is as good as it's going to get for the vast majority of my music and - horror of horrors - I'll stick with 256K AAC from iTunes for the one-off tracks I take a fancy to periodically. How many of us seriously want to buy an Eminem album just to hear the single ?

 

Finally, I find this a very interesting discussion given the pasting that a certain engineer from Ayre received when he came here to publicly state that he felt DSD was another scam from Sony and others with a vested interest in pushing a 'solution' to a problem that doesnt exist. He still went ahead and built DSD-capability into the DAC, but one of the first things I saw in reviews of the new toy was 'I cant believe how good this sounds with Redbook !'. Interesting times ahead.

[Jud]

 

Sound recording is not modelling. Back to the mountains: if you want a representation of the K2, simply take a picture of it with a suitable camera.

 

 

I agree recording is not modeling. But isn't *reconstruction* of the signal one has recorded (by sampling it) analogous to modeling?

[Jud]

 

Take an infinitely fast infinitely narrow positive going impulse. What is its Fourier transform?

 

I thought instinctively for some reason it might be 1 (which, I flatter myself, is not bad for someone whose last math course was in high school more than forty years ago - but then I must also apologize, since I know that the math is the reality of all this, and my not knowing any makes it difficult to explain or discuss concepts of interest).

[Jud]

The inverse of an ERB's width has the dimension of time, and is a measure for the time a particular ERB needs to fully react to a stimulus (you see where I am heading?).

The ringing has to be confined to less than this time.

 

Further, if the experiment is redone with increasingly higher filter cutoff frequency, it is known that the ringing audibility disappears at a certain (listener dependent) frequency. For most living and breathing people this is below 20kHz. Even then one can always widen the filter's transition band to about one ERB, in which case filter and cochlea are equally fast or slow.

 

Flying high above all of this is Miska, who simply insists that a filter response be confined to one cycle of 20kHz. That's another approach. One would call it overoveroverkill ;-)

 

Before getting back to ringing, would I be correct in assuming you agree there is better and worse ADC and DAC filtering available to studios and consumers? Can we further extend this agreement to audibly better and worse? If so, what parameters/characteristics of the filtering do you consider primarily responsible for giving audibly different results?

[Jud]

When hearing differences that can't be measured, the burden of providing evidence falls on those who make the extraordinary claim.

 

With regard to at least some of the subjects being discussed, the differences can easily be measured, but the question is whether we can hear them - so almost the reverse of "hearing differences that can't be measured."

[Jud]

Wat is exactly the difference of time and frequency domain? Is frequency not just a derivative of time?

 

I don't know about "exactly the difference," and you're correct that time and frequency domain are two ways of looking at the same basic phenomena. But for what it is worth, coming from a non-expert, it seems to me that certain characteristics of the behavior of filters used in digital audio tend to get lumped into "time domain" or "frequency domain." For example, ringing is often discussed as a time domain phenomenon, and aliasing is often discussed as a frequency domain phenomenon.

[wgscott]

With regard to at least some of the subjects being discussed, the differences can easily be measured, but the question is whether we can hear them - so almost the reverse of "hearing differences that can't be measured."

 

In what Dennis described, the differences do not exist in the audible region.