High resolution audio. The science, or lack of...?

[shadders]

So, what would happen if we were to repeat fr0gs experiment, but using a cassette deck instead of 16/44 files?

IE take a 24/96 file, make a cassette recording of it. Play that recording back into a decent ADC recording at 24/96, invert the phase and then compare that file against the original 24/96 file using the same methodology fr0g has used?

Would we again only have some differences at -70 odd dbs in the upper frequencies, or would there more difference than that?

Hi,

At a guess the cassette will reduce the higher frequency content, include tape and other noises and the difference will be much greater than the computer Audacity experiment.

I have just taken an audio file, copied, inverted the second, added to the first - obtained a completely zero valued sampling time domain file. Took a spectral plot and the graph seems have a minimum of -90dB.

As such, fr0g may need to examine his results as a text file to see the actual values. (export button on the graph)

The reason i state this is that all my values despite the graph minimum being -90dB, were all "negative infinity". So the graph may not be displaying the actual calculated values.

Regards,

Shadders.

That's possible.

However, it does demonstrate that the difference is negligable and almost certainly not audible in a normal listening room. I think I linked the difference file somewhere, which as far as Audacity is concerned is a flat, null waveform. Only the spectral plot reveals anything.

Hi,

I would have to disagree. The difference if at -90dB is potentially audible - as before the issue of dither at the -96dB level, if not applied has been proven to be audible.

The -90dB is a difference value, not absolute.

OK- to take another approach - how are you going to prove your statement "and almost certainly not audible in a normal listening room" ?

Regads,

Shadders.

 

[lindsayt]

So, what would happen if we were to repeat fr0gs experiment, but using a cassette deck instead of 16/44 files?

IE take a 24/96 file, make a cassette recording of it. Play that recording back into a decent ADC recording at 24/96, invert the phase and then compare that file against the original 24/96 file using the same methodology fr0g has used?

Would we again only have some differences at -70 odd dbs in the upper frequencies, or would there more difference than that?

As soon as you introduce analogue to digital conversion you will introduce some kind of noise. It's only the other way around where the signal in the audible spectrum can be perfectly recreated and tested this way.

Whether we agree that that is at a 44.1 KHz sample rate or a bit higher, that's a different argument.

Yeah, but how much noise, using the methodology you've used?

What I'm looking for is a variation of your experiment that might shed some light on how useful and valid an experiment your original one is.

Because I'm still not sure exactly what it is that you've measured and graphed and how?

 

[manicm]

As soon as you introduce analogue to digital conversion you will introduce some kind of noise. It's only the other way around where the signal in the audible spectrum can be perfectly recreated and tested this way.

With their new Exakt system Linn would disagree with you.

 

[busb]

You can denigrate them as 'geeks' if you like (although I understand that word is now used somewhat more positively than you may have intended it), but why don't *you* let them get on with what they want to get on with.

let the nice geeks get on with swapping their graphs

Just remember Norman that geeks design the means for you & the rest of us to listen to recorded music in the 1st place.

 

[jcbrum]

I would have to disagree. The difference if at -90dB is potentially audible - as before the issue of dither at the -96dB level, if not applied has been proven to be audible.

The -90dB is a difference value, not absolute.

OK- to take another approach - how are you going to prove your statement "and almost certainly not audible in a normal listening room" ?

Regads,

Shadders.

I take your point about it being a difference value, and not absolute, but I think fr0g is right, - anything 90dB down is inaudible, whether it's a difference, or absolute.

You ask us to take a different approach to prove it.

Well most power amplifiers have around 40dB of gain at best. A level of -90dB is like turning your amplifier down to minimum volume and listening. In fact it's more than twice as quiet as that, even. It's totally inaudible, but it might be measurable.

JC

 

[shadders]

I would have to disagree. The difference if at -90dB is potentially audible - as before the issue of dither at the -96dB level, if not applied has been proven to be audible.

The -90dB is a difference value, not absolute.

OK- to take another approach - how are you going to prove your statement "and almost certainly not audible in a normal listening room" ?

Regads,

Shadders.

I take your point about it being a difference value, and not absolute, but I think fr0g is right, - anything 90dB down is inaudible, whether it's a difference, or absolute.

You ask us to take a different approach to prove it.

Well most power amplifiers have around 40dB of gain at best. A level of -90dB is like turning your amplifier down to minimum volume and listening. In fact it's more than twice as quiet as that, even. It's totally inaudible, but it might be measurable.

JC

Hi,

OK - given your response - should we remove dither ?

If the reason for implementing dither is based on audible artefacts at the same threshold of -90dB if you do not implement dither, then the entire audio industry must be wrong ?

If a non-dithered audio file causes audible issues at the -90dB to -96dB threshold for a 16bit recording, then surely you can hear the -90dB differences for the downsampled audio file.

You are basing your response on absolute levels as per fr0g - you have made statements regarding amplifier gain etc., but this is not the issue.

My statement to prove it was based on fr0g's statement that it is inaudible - but only to his ears.

Regards,

Shadders.

 

[jcbrum]

Shadders, dither is applied to nullify quantisation error, which is audible at all levels, not just -90dB.

Dither does nullify quantisation error, and therefore restores 'perfect' sound quality, but at the expense of an increased noise floor, in the region below -90dB, which is not audible, but is measurable.

That is why it is entirely satisfactory, to human ears, in a domestic environment, which is what fr0g is alluding to.

JC

 

[shadders]

Shadders, dither is applied to nullify quantisation error, which is audible at all levels, not just -90dB.

Dither does nullify quantisation error, and therefore restores 'perfect' sound quality, but at the expense of an increased noise floor, in the region below -90dB, which is not audible, but is measurable.

That is why it is entirely satisfactory, to human ears, in a domestic environment, which is what fr0g is alluding to.

JC

Hi,

This is not correct. Dither is applied to the Least Significant Bit ONLY. Quantisation error ONLY applies to the Least Significant Bit.

Dither does NOT nullify quantisation error, it is the modulation of the Least Sginificant Bit to spread the error with a specific probability density function that is less abrasive to the ears. The audio industry use this (dither) because quantisation error is audible for 16bits word depth.

The issue is that fr0g has used dB's to represent the difference - so yourself and fr0g are interpreting this as an absolute level of the audio.

fr0g has taken the spectrum of the difference of the audio signal. All fr0g has shown is that the difference between the two spectrums is 0.003% across the 0Hz to 17.5kHz bandwidth.

Again, i have remained with the -90dB aspect since this is what is presented, although it is a spectrum difference only.

You cannot get back to the audio file from the spectrum - it is a one way function. All this infers is that the spectrum of the 16bit/44.1kHz is 0.003% different to the 24bit/96kHz file across the audio band 0Hz to 17.5kHz.

Regards,

Shadders. .

 

[Jota180]

Talking about 24 bit and 16 bit but is there any proper double blind scientific test or experiment that shows people can tell the difference between 320Kb MP3 and CD quality? Or even 192 VBR and CD?

Are there any properly conducted double blind scientific studies into cables, cheap and expensive. HIFI racks, cheap and expensive. Power supply, 'clean' or standard?

If there were wouldn't we have seen them posted by now and also, if the companies behind all these expensive accoutrements are so sure of their products and so confident of their ability, why no independently commissioned double blind tests?

Taking it further, why does this magazine, Whathifi, not commission these tests? They'd only need to be done propely, once.

Surely the readership would be most interested in the results and it would put to bed the topic once and for all.

 

[professorhat]

If only someone had thought of that before...

 

[davedotco]

Talking about 24 bit and 16 bit but is there any proper double blind scientific test or experiment that shows people can tell the difference between 320Kb MP3 and CD quality? Or even 192 VBR and CD?

Are there any properly conducted double blind scientific studies into cables, cheap and expensive. HIFI racks, cheap and expensive. Power supply, 'clean' or standard?

If there were wouldn't we have seen them posted by now and also, if the companies behind all these expensive accoutrements are so sure of their products and so confident of their ability, why no independently commissioned double blind tests?

Taking it further, why does this magazine, Whathifi, not commission these tests? They'd only need to be done propely, once.

Surely the readership would be most interested in the results and it would put to bed the topic once and for all.

Two answers to that.

The cynical view is that it is not in the industry's intrest to do such tests, the industry, in all it's aspects, relies on subjective evaluations to drive sales.

Less cynical is the view that not all differences can be evaluated by double blind testing. There is at least a possibility that some aspects of performance impact on a subconcious level and are not shown up in relatively short duration tests.

Not for one instance knocking blind testing, I have taken part on a couple of occasions and find them enlightening. I have also lived with fine sounding systems that fail to satisfy in the long term, difficult to prove as it is so subjective, so I am just trying to give a considered view.

 

[Phileas]

I'm reposting this link to a very clear explanation of dither for the benefit of those who may have missed it.

http://www.audiomisc.co.uk/inadither/Page1.html

(It contradicts lindsayt and Oldric)

 

[altruistic.lemon]

I'm reposting this link to a very clear explanation of dither for the benefit of those who may have missed it.

http://www.audiomisc.co.uk/inadither/Page1.html

(It contradicts lindsayt and Oldric)

Thank you so much! It's an honour to have people like you pointing out the error of our ways, mate.

 

[jcbrum]

Hi,

This is not correct . . // . . Dither does NOT nullify quantisation error . . // . . . Regards,

Shadders. .

Ah well, . . . It seems we cannot agree, and probably should leave it there, for the time being. I fear to go on with such a technical argument may well be tedious for many readers who might not wish to study the theory in such detail.

It's a good job that Claude Shannon was such a brilliant mathematician, and working with Harry Nyquist's theorem, laid out all the rules for digital audio so precisely, so that we don't have to.

JC

 

[busb]

I'm reposting this link to a very clear explanation of dither for the benefit of those who may have missed it.

http://www.audiomisc.co.uk/inadither/Page1.html

(It contradicts lindsayt and Oldric)

What the hell would this guy know about music, electronics or anything else? Anyone would think he's one of those damn geeks who think they no everything! Yeh, Yeh. Ignorance? Bring it on. Just remember that these geeky gits are a bunch of shite-stirring misfits with personality disorders. Geeks & other engineering types should not be allowed to design anything, nothin', nothin' Geeks are even worst than them arty farts!

 

[busb]

Talking about 24 bit and 16 bit but is there any proper double blind scientific test or experiment that shows people can tell the difference between 320Kb MP3 and CD quality? Or even 192 VBR and CD?

Are there any properly conducted double blind scientific studies into cables, cheap and expensive. HIFI racks, cheap and expensive. Power supply, 'clean' or standard?

If there were wouldn't we have seen them posted by now and also, if the companies behind all these expensive accoutrements are so sure of their products and so confident of their ability, why no independently commissioned double blind tests?

Taking it further, why does this magazine, Whathifi, not commission these tests? They'd only need to be done propely, once.

Surely the readership would be most interested in the results and it would put to bed the topic once and for all.

Maybe they want to hold onto their jobs? The last thing any commercial economy wants is potential purchasers questioning the big picture. They are there to consume, not think. Thinking is subversive, the sort of **** engineers do.

On a slightly more serious note, I love the way some folk bandy about the term "blind testing" without seeming to question their effectiveness or understand what's involved or how long it takes to do them properly. If someone could prove the non-existance of God, do you think the entire human race would stop believing in heaven? If this particular thread has taught me anything is that many people will believe anything they damn well please with total disregard for evidence or for the fact that engineers design the means for us to listen to recorded music to start with. Many people can't stand the idea that others are less ignorant than they are. Damn smartarses!

 

[shadders]

Hi,

This is not correct . . // . . Dither does NOT nullify quantisation error . . // . . . Regards,

Shadders. .

Ah well, . . . It seems we cannot agree, and probably should leave it there, for the time being. I fear to go on with such a technical argument may well be tedious for many readers who might not wish to study the theory in such detail.

It's a good job that Claude Shannon was such a brilliant mathematician, and working with Harry Nyquist's theorem, laid out all the rules for digital audio so precisely, so that we don't have to.

JC

Hi,

I introduced the dither argument since :

1. Dither operates on the LSB of a 16bit word for CD and hence is at the bottom end of the maximum Signal to Noise ratio, that is -96dB to -90dB.

2. fr0g has stated and others agreed that you cannot hear signals at the -90dB level.

If fr0g et al are correct, then why is dither applied to 16bit audio samples ?

Dither is applied to remove the artefacts at the -96dB to -90dB level since they can be heard. (laymans terms used)

fr0g's analysis is that you cannot hear -90dB. This is in error based on fr0gs statement that the -90dB value is an absolute signal level in his plot.

It is not, it is a difference value, which equates to 0.003% difference.

I could construct a waveform - frequency sweep that will be within 0.003% of the 16bit/44.1kHz spectrum fr0g produced. The two waveforms, fr0gs audio track, my frequency sweep, differ by only 0.003%, and hence have the same sprectral signature. Obviously they are completely different in the time domain. For fr0g to state they are not different - then his test must be true for all cases.

They are not true for all cases - hence fr0g's conclusion is not correct.

This is a case of misunderstanding the process results to obtain the wrong conclusion.

Regards,

Shadders

 

[busb]

Hi,

This is not correct . . // . . Dither does NOT nullify quantisation error . . // . . . Regards,

Shadders. .

Ah well, . . . It seems we cannot agree, and probably should leave it there, for the time being. I fear to go on with such a technical argument may well be tedious for many readers who might not wish to study the theory in such detail.

It's a good job that Claude Shannon was such a brilliant mathematician, and working with Harry Nyquist's theorem, laid out all the rules for digital audio so precisely, so that we don't have to.

JC

Hi,

I introduced the dither argument since :

1. Dither operates on the LSB of a 16bit word for CD and hence is at the bottom end of the maximum Signal to Noise ratio, that is -96dB to -90dB.

2. fr0g has stated and others agreed that you cannot hear signals at the -90dB level.

If fr0g et al are correct, then why is dither applied to 16bit audio samples ?

Dither is applied to remove the artefacts at the -96dB to -90dB level since they can be heard. (laymans terms used)

fr0g's analysis is that you cannot hear -90dB. This is in error based on fr0gs statement that the -90dB value is an absolute signal level in his plot.

It is not, it is a difference value, which equates to 0.003% difference.

I could construct a waveform - frequency sweep that will be within 0.003% of the 16bit/44.1kHz spectrum fr0g produced. The two waveforms, fr0gs audio track, my frequency sweep, differ by only 0.003%, and hence have the same sprectral signature. Obviously they are completely different in the time domain. For fr0g to state they are not different - then his test must be true for all cases.

They are not true for all cases - hence fr0g's conclusion is not correct.

This is a case of misunderstanding the process results to obtain the wrong conclusion.

Regards,

Shadders

?? Dither is used to remove otherwise audible low level distortion (well above -90 to -96dB) at the expense of wideband but inaudible noise. This noise can be fairly broadband or shoved up towards the HF. It's amazing what randomising the LSBs can achieve. It took me ages to work out what the hell dither was for! It's also important to dither the conversion from 24 to 16 bit appropriately to avoid truncation artefacts. Any effective (downward) conversion of bit rate will be inexact. Applied dither may well reside with the LSB but much analogue noise will impinge on higher bits during initial conversion that will effectively do the same thing as pseudo random/Gausian/triangular dither does.

 

[jcbrum]

// . . fr0g's analysis is that you cannot hear -90dB. This is in error based on fr0gs statement that the -90dB value is an absolute signal level in his plot.

It is not, it is a difference value, which equates to 0.003% difference . . //

Regards,

Shadders

Thanks for coming back, in reply, again, Shadders, but I'm sorry, I don't agree, as I said.

It doesn't matter what you call it, quantisation error, dither artifact, noise, distortion, difference error, whatever. If it's -90dB or distortion of 0.003%, whatever, I'm satisfied that it's below human perception for hearing in domestic audio replay, even though it might be measurable.

If it's inaudible, it doesn't matter whether it's present or absent.

Frog couldn't hear it, and I don't think I could, and I don't think anyone could.

That's it as far as I'm concerned.

Regards, JC

 

[fr0g]

As soon as you introduce analogue to digital conversion you will introduce some kind of noise. It's only the other way around where the signal in the audible spectrum can be perfectly recreated and tested this way.

With their new Exakt system Linn would disagree with you.

I doubt it.

 

[fr0g]

You cannot get back to the audio file from the spectrum - it is a one way function. All this infers is that the spectrum of the 16bit/44.1kHz is 0.003% different to the 24bit/96kHz file across the audio band 0Hz to 17.5kHz.

Regards,

Shadders. .

No.

Up to around 17 KHz, the inverted lower res file COMPLETELY cancelled out the HD file. NULL, zero, NADA. There is no information there, it is a deceased sound wave, this sound wave has ceased to be....

Over that you can argue absolutes. Rather pointlessly and missing the point imo, but still.

In any case, it's "more than likely" inaudible to anyone in a normal, quiet listening room, not least because the level is probably below which your amp has reached its mute point.

 

[shadders]

Hi,

This is not correct . . // . . Dither does NOT nullify quantisation error . . // . . . Regards,

Shadders. .

Ah well, . . . It seems we cannot agree, and probably should leave it there, for the time being. I fear to go on with such a technical argument may well be tedious for many readers who might not wish to study the theory in such detail.

It's a good job that Claude Shannon was such a brilliant mathematician, and working with Harry Nyquist's theorem, laid out all the rules for digital audio so precisely, so that we don't have to.

JC

Hi,

I introduced the dither argument since :

1. Dither operates on the LSB of a 16bit word for CD and hence is at the bottom end of the maximum Signal to Noise ratio, that is -96dB to -90dB.

2. fr0g has stated and others agreed that you cannot hear signals at the -90dB level.

If fr0g et al are correct, then why is dither applied to 16bit audio samples ?

Dither is applied to remove the artefacts at the -96dB to -90dB level since they can be heard. (laymans terms used)

fr0g's analysis is that you cannot hear -90dB. This is in error based on fr0gs statement that the -90dB value is an absolute signal level in his plot.

It is not, it is a difference value, which equates to 0.003% difference.

I could construct a waveform - frequency sweep that will be within 0.003% of the 16bit/44.1kHz spectrum fr0g produced. The two waveforms, fr0gs audio track, my frequency sweep, differ by only 0.003%, and hence have the same sprectral signature. Obviously they are completely different in the time domain. For fr0g to state they are not different - then his test must be true for all cases.

They are not true for all cases - hence fr0g's conclusion is not correct.

This is a case of misunderstanding the process results to obtain the wrong conclusion.

Regards,

Shadders

?? Dither is used to remove otherwise audible low level distortion (well above -90 to -96dB) at the expense of wideband but inaudible noise. This noise can be fairly broadband or shoved up towards the HF. It's amazing what randomising the LSBs can achieve. It took me ages to work out what the hell dither was for! It's also important to dither the conversion from 24 to 16 bit appropriately to avoid truncation artefacts. Any effective (downward) conversion of bit rate will be inexact. Applied dither may well reside with the LSB but much analogue noise will impinge on higher bits during initial conversion that will effectively do the same thing as pseudo random/Gausian/triangular dither does.

Hi,

Agree that if the signal is not full scale deflection, that effectively the number of bits used is less than 16bits, which then reduces the S/N ratio, and hence the dither acting upon the LSB, will in effect be higher with regards to the noise floor it operates in.

I do not see how it impinges upon the higher bits, if only the LSB is modulated, and the DAC is monotonic.

If the signal is not full scale then the -90dB to -96dB is not the range of the dither noise.

Regards,

Shadders.

 

[shadders]

// . . fr0g's analysis is that you cannot hear -90dB. This is in error based on fr0gs statement that the -90dB value is an absolute signal level in his plot.

It is not, it is a difference value, which equates to 0.003% difference . . //

Regards,

Shadders

Thanks for coming back, in reply, again, Shadders, but I'm sorry, I don't agree, as I said.

It doesn't matter what you call it, quantisation error, dither artifact, noise, distortion, difference error, whatever. If it's -90dB or distortion of 0.003%, whatever, I'm satisfied that it's below human perception for hearing in domestic audio replay, even though it might be measurable.

If it's inaudible, it doesn't matter whether it's present or absent.

Frog couldn't hear it, and I don't think I could, and I don't think anyone could.

That's it as far as I'm concerned.

Regards, JC

Hi,

Ok - so you cannot hear at -90dB even if this is a difference error/distortion which equates to 0.003%.

At what value of S/N ratio (number of bits used) do you perceive (hear) the quantisation error, where dither removes this effect ?.

Is it 12bits, 14bits, 11bits ??.

Regards,

Shadders.

 

[shadders]

You cannot get back to the audio file from the spectrum - it is a one way function. All this infers is that the spectrum of the 16bit/44.1kHz is 0.003% different to the 24bit/96kHz file across the audio band 0Hz to 17.5kHz.

Regards,

Shadders. .

No.

Up to around 17 KHz, the inverted lower res file COMPLETELY cancelled out the HD file. NULL, zero, NADA. There is no information there, it is a deceased sound wave, this sound wave has ceased to be....

Over that you can argue absolutes. Rather pointlessly and missing the point imo, but still.

In any case, it's "more than likely" inaudible to anyone in a normal, quiet listening room, not least because the level is probably below which your amp has reached its mute point.

Hi,

I did ask you to examine the table of the plot (export button on the graph) to state what the values were for the 0Hz to 17.5kHz bandwidth.

If these values are negative infinity, then yes - they have cancelled out. The value -90dB is not zero, null or otherwise.

Your statement :

In any case, it's "more than likely" inaudible to anyone in a normal, quiet listening room, not least because the level is probably below which your amp has reached its mute point.

An amplifier can resolve to a much lower resolution than -90dB.

The -90dB value you have in your graph needs to be validated - can you provide the table rather than the graph - just a few points ?. Thanks.

Regards,

Shadders.

 

[jcbrum]

An amplifier can resolve to a much lower resolution than -90dB.

In theory, a perfect mathematical model, of a perfect amplifier, in a perfect environment, can . . .

But in practice, in a real environment, a real amplifier will be limited by it's self noise and induced noise. That will be around the same level as -90 or probably worse. So the resolution would be lost in the noise, but more importantly you couldn't hear it anyway, because it's inaudible, because it's below your auditory threshold with human ears in a real environment.

JC