Sorry. Need more coffee here. That cloud of quantization noise is not
above 96 kHz, because that is where a recording sampled at 192 cuts off. I
meant between 48 kHz (where a recording sampled at 96 kHz cuts off) and 96
You have to keep in mind that these large numbers way up there are really
small differences in pitch, as the numbers double per octave. Several
thousand kHz of frequency way up there is a very small pitch spread, where
in the lower end of the spectrum, several Hz is a large pitch spread.
On Tue, Mar 25, 2014 at 9:59 AM, John Haley <[log in to unmask]> wrote:
> I am just reporting what you see in a .WAV file (in spectral view), since
> you can't directly hear it. Noise does not look at all like music most of
> the time. What you see extending up above 22 kHz in the spectral view of a
> hi-def .WAV file is the extension of certain tones that have a lot of upper
> frequency content right up into the stratosphere. I looks just like music
> looks, not random noise. Could it be "ringing" set off by that note?
> Perhaps. But I would expect that kind of corruption of particular notes
> to have some kind of audible effect in the range I can hear, or to be
> visible as in increase in energy at upper levels, which would not happen
> with musical overtones (not what you see). All I can say I how it looks.
> Many LPs do appear to have content up there above 22 kHz, and good
> cartridges can capture that. What happens in the rest of an audio system
> chain is up for grabs.
> Since I started doing restoration work at 96/24, I have noticed some
> things. I can also do it at 192 sampling rate, but when you do that you
> get a .WAV file having a very visible layer of quantization noise (an upper
> thick cloud of noise blanketing the top of the "picture") above 96 kHz.
> It's not audible, but why put the equipment and the media thru all the
> trouble to produce/reproduce that, when (because it is noise) it cannot
> possibly add anything to the musical signal? At 96/24, all of this noise
> is eliminated, and the audio signal at 96/24 is audibly indistinguishable,
> to me, from the same signal recorded at 192. At that point, whatever
> benefit might be gained by using 192 has become insignificant in the real
> world--i.e., not audible, and as Tom points out, possibly damaging to
> equipment. But that is not true at all for the comparison between 44/16
> and 96/24, which is very much audible.
> I think a lot of early CD's had stinky upper frequency sound because of
> phasing errors caused by the way upper frequencies above 22 kHz were
> filtered out, causing "side effects" in the audible signal. Not to
> mention, human beings were doing the audio work, and not everyone really
> knows what they are doing, or cares, then and now. I don't think any of
> the bad rap that early CD's got were the fault of the medium itself. But I
> agree that many of the earliest CD releases do not sound right, having a
> "hard," unnatural treble. I think that situation improved drastically as
> time wore on, and generally a CD issue would sound better than a prior LP
> issue, because (1) we got rid of the groove noise, and (2) the CD's were
> often the result of a return to the master tape. But these days, I don't
> assume anything. I listen to a lot of CD transfers that do not sound as
> good as the materials they were created from. That's the human factor at
> work again, not to mention variations in equipment used, as pointed out by
> On Tue, Mar 25, 2014 at 9:27 AM, Mark Durenberger <[log in to unmask]>wrote:
>> One refers to Rupert Neve's thoughts on why coherent super-audible
>> response is useful if not necessary.
>> Someone else on the list may know where his paper was delivered...I have
>> the audio of his remarks for anyone interested.
>> Mark Durenberger, CPBE
>> -----Original Message----- From: Gray, Mike
>> Sent: Tuesday, March 25, 2014 8:03 AM
>> To: [log in to unmask]
>> Subject: Re: [ARSCLIST] Fwd: [ARSCLIST] "Why Vinyl Is the Only Worthwhile
>> Way to Own Music"
>> Seconding Tom's comments - what exactly *is* that energy above 20kHz?