Phil Leigh;381160 Wrote: 
> The link doesn't work ... is it this one?
> http://en.wikipedia.org/wiki/Information_theory
> 
> If so - yes I have read it.

Sorry. Nope, I mean the "Practical Implications" section of this one:

http://en.wikipedia.org/wiki/Sampling_(information_theory)


In short, I agree with you wholeheartedly that 44.1 captures all the
salient information of a band-limited (20-20khz) signal. No doubt about
it.

The key point that I'm trying to make is that it is not possible to
implement the necessary anti-aliasing and reconstruction filters in a
sonically-transparent way. We cannot build "clean" filters steep enough
- 96dB down immediately for 16-bit data (24-bit is worse). Phase effects
alone are large.

These filters are very possible at 88Khz or above.


Phil Leigh;381160 Wrote: 
> what "intersample effects"? Do you think there is information (different
> sample values) missing. There isn't. If there were, the Shannon et al
> are completely wrong.

They are not wrong. We just can't engineer the filters necessary to
make their math "real."

Phil Leigh;381160 Wrote: 
> This is easily proven. Record something with no or very little energy
> above 20kHz at 96/24 and downsample it properly in software to 48/24.
> Compare the two versions in a blind test on good equipment. They sound
> the same. Even better, record an analogue replay of the two versions
> using a 96/24 ADC and compare the captured files...hmm - identical.
> Why? - because there is no extra information captured at the higher
> sampling rate.

Fair enough. I do not believe that the two files would be bit-perfect
correct, but - as I have said before - the blind listening test may be
challenging! <g>

Phil Leigh;381160 Wrote: 
> I wouldn't dream of telling you or anyone what you are and aren't
> hearing! Only your brain can do that.
> 
> I would however urge you to try the tests with an open mind. I have
> several friends who confidently told me that they could detect a
> difference between some 96/24 Linn Masters and the same downsampled to
> 48/24... until we did the blind test and they failed miserably :o). I
> failed too...

Believe me, my mind is open, and you make excellent points. It would be
quite nice for me and everyone else if 16/44 sounded as good as 24/88
(or above). But in my opinion, it doesn't.

I will also point out that another key variable is in play besides
sampling rate - bit depth! I must confess that I have not done tests
at, say, 16/96. But one day I shall.


-- 
teros
------------------------------------------------------------------------
teros's Profile: http://forums.slimdevices.com/member.php?userid=18601
View this thread: http://forums.slimdevices.com/showthread.php?t=57631

_______________________________________________
audiophiles mailing list
[email protected]
http://lists.slimdevices.com/lists/listinfo/audiophiles

Reply via email to