Generational fidelity loss with repeated DAC/ADC loops.

[RayDunzl]

Oh no, more stuff to worry about.

 

[Blumlein 88]

Such tests results should be humbling and make one revisit their assumptions about audio and their listening ability. Instead he responds thusly:

"One hypothesis is that with the others high frequency components were exciting resonances in the ADCs digital filter, since the phone doesn't have anything above 15K it may not be exciting said resonances. "

Resonances in the ADC (capture device)? Really?

Well I did not respond to that as I am not a big name digital designer and no one would believe me.

 

[fas42]

Well, the idea is to start with two versions of some music, which nominally are the same - if they sound identical to you that's probably as far as it needs to go ... but if they sound different then there should be some reason for that - you know, that ol' cause and effect thing. That's where things can become interesting - in fact I might even use a scary world for that, "science" - of course many people work on the basis that Authority Figures already know everything that needs to be known, for some reason this is also called science ...

Started on examining those files, and so far the resampled version does not match the original, at about 55dB down - need another round of perusing to get a handle on it ...

 

[Blumlein 88]

Well, the idea is to start with two versions of some music, which nominally are the same - if they sound identical to you that's probably as far as it needs to go ... but if they sound different then there should be some reason for that - you know, that ol' cause and effect thing. That's where things can become interesting - in fact I might even use a scary world for that, "science" - of course many people work on the basis that Authority Figures already know everything that needs to be known, for some reason this is also called science ...

Started on examining those files, and so far the resampled version does not match the original, at about 55dB down - need another round of perusing to get a handle on it ...

If you used the originals from the other test, I chopped a few samples off the front so checksums would differ. If you look at the other forum I think I posted the difference. Other than a tiny ultrasonic bit differences are below -100 dB.

 

[Phelonious Ponk]

Such tests results should be humbling and make one revisit their assumptions about audio and their listening ability. Instead he responds thusly:

"One hypothesis is that with the others high frequency components were exciting resonances in the ADCs digital filter, since the phone doesn't have anything above 15K it may not be exciting said resonances. "

Resonances in the ADC (capture device)? Really?

Did he do anything to test the hypothesis? Do we even know that the phone was limited to 15K? Ken Rockwell measured the iPhone 5 at down by just half a db at 20k back in 2013. Looking at his measurements, there's no reason to expect the iPhone to sound anything but excellent. Sometimes, listening to audiophiles, I begin to wonder if I just can't hear, or if there are just bad things going on in inexpensive consumer audio that I don't understand. Then I read opinions and findings from folks who really know what they're talking about and think no, maybe my ears are just fine and it's audiophiles who just don't understand..

http://www.kenrockwell.com/apple/iphone-5/audio-quality.htm

 

[amirm]

Do we even know that the phone was limited to 15K?

I think it is in this case and hence the argument he is making.

 

[amirm]

Did he do anything to test the hypothesis?

That's like asking if anyone needs to test the hypothesis and photon torpedos exist. It "makes sense" on Star Trek so we go on with it.

 

[Blumlein 88]

Yes, I measured the smartphone being referred to. It is flat to 15 khz and then down -10 db at 20 khz.

 

[fas42]

If you used the originals from the other test, I chopped a few samples off the front so checksums would differ. If you look at the other forum I think I posted the difference. Other than a tiny ultrasonic bit differences are below -100 dB.

Yes, that difference I found was a false lead - resampling causes the waveforms to very slightly misalign, after adjusting the waveform envelopes to match to a bit accuracy; even though I upsampled to an extremely high whole number multiple of the 48k. I was caught once before by this - I refined the method yesterday sufficiently to be certain that the two waveforms did in fact match, to any sane level.

 

[Phelonious Ponk]

Yes, I measured the smartphone being referred to. It is flat to 15 khz and then down -10 db at 20 khz.

Well, that's quite a drop off. Not that it would make any difference to my ears.

Tim