[linux-audio-user] Re: 192kHz

Hans Fugal hans at fugal.net
Sat Jan 28 11:10:04 EST 2006

On Sat, 28 Jan 2006 at 13:21 +0100, Carlo Capocasa wrote:
> > Only two values are enough to mathematically reproduce
> > an exact waveform; even more precise than you can sample it.
> Like vector graphics! So if that's the case say, why do we still have
> MP3? Why don't we just convert whatever sound files we have into
> mathematical formulae and have players to convert them to sound at any
> sampling rate?

I know you're being facetious, but whenever you go digital, you go
mathematical. A bunch of samples may not be what's traditionally thought
of as a formula, but in actuality they are. They're called vectors by
the math-heads. You do mathematical and physical operations on vectors
to get the sound back out (DAC). 

When you're talking MP3 there's even more math involved.

Might as well weigh in on the thread in general while I'm here. What I'm
about to say is not necessarily in direct response to your post.

I'm no expert of all this, but I am a student of it. The sampling
theorem says you can reproduce any signal with a maximum frequency of
sr/2 with perfect accuracy. It's been shown a million times over. It's

Humans hear up to 20khz. I don't understand when someone says
higher frequencies may modulate the lower ones - if the lower
frequencies were modulated then they would already be modulated when
captured, no? Please fill me in. Some say we can sense the lower and
higher frequencies with our bodies but not our ears. I'm no biologist,
but that sounds possible enough.

Part of sampling is filtering out anything above sr/2 to avoid aliasing,
but a brick wall filter or one with a steep slope can be troublesome,
so we need headroom. I'm no DSP hardware designer, so I don't know
whether 44.1 or 48 or 96 is necessary for doing this well. It's my
guess, however, that low-quality 44.1 gear may just sound worse than 48
or 96 because of cheap filters.

I think headroom is always a good idea, balanced with resource needs.
I've pretty much been convinced that 96khz isn't a bad idea, if I ever
have the cash for the hardware and gear to match. I'm not yet convinced
that 192 is necessary.

And I completely agree about the point to worry less about the gear and
make good music, if making music is what you do. Many people including
myself still enjoy old jazz recordings that sound like they were
recorded in a tin can. I've tolerated listening to some music once or
twice that was so badly damaged by mp3 that it was painful, but the
music was good so I listened anyway.

Hans Fugal ; http://hans.fugal.net
There's nothing remarkable about it. All one has to do is hit the 
right keys at the right time and the instrument plays itself.
    -- Johann Sebastian Bach
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://music.columbia.edu/pipermail/linux-audio-user/attachments/20060128/4abf1fa8/attachment.bin

More information about the linux-audio-user mailing list