This is a very good question. Hopefully Teddy Ray may be able to help
out with his very high powered friends.

I have been trying really hard to grasp the temporal resolution issue
and have found it really difficult. It seems to have been kicked off in
audiophile circles by the NOS dac bollocks and (separately and
indirectly) by an article by Craven for AES about the desirability of
apodising/ minimum phase filters [not quite the same thing I know].
These filters are somewhat in vogue although neither weiss nor
briscasti seem to use them. 

Annoyingly I think that this is an incredibly difficult area and the
basic digital electronics text books i have read have very little on
it.

Just to get things started here are a couple of thoughts [APOLOGIES I
NOW REALISE I HAVE GONE OFF ON A BIT OF A RAMBLE BUT HERE GOES
ANYWAY}-

a. AFAIK there really isn't any evidence that we can hear above 20Khz
(make that 15 for most of us)

b. in order to be nyquist compliant before sampling at 44.1kHz (or
downsampling to 44.1), the signal has to be band limited to 22.05kHz or
lower.

c. there is an issue about whether the effect of this band limitation
could be audible.

d. we were talking here about the band limation prior to the ADC (or
resampling) NOT in your DAC.

e. but wait! the band limitation at the ADC has been created by a
number of processes starting with the mic and then (maybe the mixer etc
and the tape if its an old analog recording. So in order to analyse out
the effect of band limitation you have to look at this chain as a
whole

f. time domain and frequency domain analysis are basically two sides of
the same coin -if you limit the max frequency you must have an impact in
the time domain.      

g. but this expression temporal resolution is a real bugger. AFAIK it
has no fixed meaning. It is bandied around by certain audiophiles as a
possible peg o which to hang their belief that "digital" especially
properly filtered digital doesn't sound good.

h if you don't believe me check out the kunchur/ JJ debate in which
Kunchur made an utter arse of himself, despite being a genuine
physicist, when he suggested that "temporal resolution" was limited to
the inter sample period.

i strictly speaking looking at the OP it seems to me that the answer is
that if there is no information in the original signal about 20Khz there
cannot be any time resolution issue with 20kHz band limitation. This is
because a proper filter should have no impact below the stop band. It
is therefore trivial that there could not be any information loss.

j. that answer might be a bit to smart arse though- the real question
is whether, assuming there was information above 20KhZ in the original
signal, the band limitation prior to sampling might still have some
detectable effect on the signal.

k Any band limitation must affect the signal in the time domain. In
particular the maximum frequency determines the maximum steepness of
the slope in the time domain. So the band limited signal has to be
spread out in the time domain if higher frequencies are filtered out. 
BUT CAN ANYONE ACTUALLY HEAR THIS? 
IF THEY CAN HEAR THE DIFFERENCE BETWEEN FIKLTERED AND UNFILTERED, the
IS ONE Sensible)  FILTER BETTER THAN ANOTHER

l.  this is where we get into pre ringng an post ringing. IN order to
make a filter which perfectly complies with nyquist it should in
principle be a linear phase filter. This will spread out the signal
equally in both directions (earlier and later) . But according to
auipophile magazines this apparently means pre ringing and this is
apparently bad.

m I would love to see 
1) a demonstration of how a real world signal is affected by a linear
phase filter with a passband up to 20hZ or so and proper attenuation at
nyquist. Weirdly I have not actually seen one
2) some evidence that people can hear the difference between the
unfiltered and filtered signal and the difference between MP and LP
filters..


-- 
adamdea
------------------------------------------------------------------------
adamdea's Profile: http://forums.slimdevices.com/member.php?userid=37603
View this thread: http://forums.slimdevices.com/showthread.php?t=93483

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

Reply via email to