Olympus-OM
[Top] [All Lists]

Re: Scanning (Was: Re: [OM] HCB - OM eqivalent to Leica M + 35/2?)

Subject: Re: Scanning (Was: Re: [OM] HCB - OM eqivalent to Leica M + 35/2?)
From: "Ken Norton" <image66@xxxxxxx>
Date: Thu, 13 Aug 1998 09:29:06 -0500
Denton wrote:
>The theoretical limit of human hearing in the high frequency range is say
>20,000-22,000 hz. Therefore under Nyquist theoretically in order to capture
>all information in this upper range, which is in analog wave form, the wave
>must be sampled twice.
>
>However, Nyquist or no, it has become clear that 44khz is not sufficient to
>capture all the harmonic information of a musical recording, which is way a
>suprising amount of music is recorded to 2" analog tape, and why records
>sound better than CDs. In the near future it appears the sampling rate
>might be bumped up to 96khz (requiring the purchase of new CD players,
>etc.) Perhaps that will be sufficient to finally capture true musical
sound.


Audio data is only slightly different the visual data.  On the subject of
audio, it is true that 44.1 kHz is barely sufficient for quality audio.  The
recording industry has been switching over from 48kHz tracking (later to
convert to 44.1 for CD mastering), to 96 kHz tracking.  Not only that but
they've been converting from 16, 18 and 20 bit resolution to 24 bit
resolution.  24/96 is a huge amount of data to be churning on/off of your
storage devices, (the industry is also converting from tape based storage to
HD based storage.

Can the ear actually hear 30 or 40 kHz?  Nope, but you can hear the affects
of it.  But most importantly, it isn't the frequency response that makes the
difference, but the timing accuracy.  If you can place a sonic nuance
precisely in the time domain you can capture and recreate the
3-dimensionality to sound.  With a standard CD recording of an orchastra,
you can visualize where the various instrument sections are located, but
with a 24/96 recording you can actually hear and identify individual
instruments within the sections.

The bit-depth, ie 16-bit (cd), 18, 20 or 24, is similar to the bit depth in
images.  The more bits, the greater the number of possible colors to make up
the picture.  In audio, 18 and 20 bit samples have slightly better S/N
(signal to noise) ratio than 16bit, but 24 bit acually doesn't stretch the
S/N ratio very much, but places a lot more (approximately 40%) bits in the
dynamic spectrum that is most important.

Oh, Denton, I've been testing with the new Earthworks directionals.  My pair
are flat (within +/- 2dB) from 5Hz up to 30kHz and down about 6 dB at 40kHz.
I've recorded these directly (via extremely high quality mic-amps) into a
Sadie running 24/96.  The results are absolutely amazing.  There is so much
"placement" information that it sounds more real than real.  Even when
listening to the studio monitors you can hear things happening behind you.
Converting the results to 44.1 kHz lost the placement information, but just
band-passing at 20kHz didn't.

Ken Norton


< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >


<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz