[Doc]

Kind of an esoteric question. Fooling around with rebalancing the mix
of someone on a karaoke site where they're performing to a stereo
track - I use the same track and invert it. Now of course by the time
I get their performance besides reverb that's added on the site it's
gone through conversion to .flv and then reconversion to .wav so it's
not a sample perfect clone of the original, nor is the track I'm
using, though they're close.

What I find when I get these close but not identical tracks matched -
i.e. where moving it one sample in either direction makes an audible
difference and adjusting gain to the nearest .1 db where it makes the
most sound go away - is that what mostly remains audible from the
track that was mixed with their original performance are the highs and
lows. Particularly obvious in spots where it's just instrumental.

Wondering why this is.

[Peter Larsen[_3_]]

Doc wrote:

Kind of an esoteric question. Fooling around with rebalancing the mix
of someone on a karaoke site where they're performing to a stereo
track - I use the same track and invert it. Now of course by the time
I get their performance besides reverb that's added on the site it's
gone through conversion to .flv and then reconversion to .wav so it's
not a sample perfect clone of the original, nor is the track I'm
using, though they're close.

What I find when I get these close but not identical tracks matched -
i.e. where moving it one sample in either direction makes an audible
difference and adjusting gain to the nearest .1 db where it makes the
most sound go away - is that what mostly remains audible from the
track that was mixed with their original performance are the highs and
lows. Particularly obvious in spots where it's just instrumental.

Wondering why this is.

"Almost identical" seems to get voided by the perceptual encoding.

Kind regards

Peter Larsen

[Les Cargill[_4_]]

Doc wrote:
Kind of an esoteric question. Fooling around with rebalancing the mix
of someone on a karaoke site where they're performing to a stereo
track - I use the same track and invert it. Now of course by the time
I get their performance besides reverb that's added on the site it's
gone through conversion to .flv and then reconversion to .wav so it's
not a sample perfect clone of the original, nor is the track I'm
using, though they're close.

What I find when I get these close but not identical tracks matched -
i.e. where moving it one sample in either direction makes an audible
difference and adjusting gain to the nearest .1 db where it makes the
most sound go away - is that what mostly remains audible from the
track that was mixed with their original performance are the highs and
lows. Particularly obvious in spots where it's just instrumental.

Wondering why this is.


You're listening to what's been discarded or added by the rate reduction
process. "Why" is because ... that's what that process does.

"I got the digital blues - my soul is just a number" - JJ Cale...

--
Les Cargill