pops and crackles are very brief transients and easy to filter out, ie
softwares that do just that. they dont impact the accuracy of something liek
this. Im sure some fennesz or noto might f with it.. :p
the other stuff you mentioned is moot because most people arent altering the
dynamics (outside of normalizing/noise reduc) audio that was professionally
produced. anything amatuer doesnt qualify as there is no copyright to be
concerned with.
the purpose was to prove I could do it. that was before I realized that
you can do just abotu anything you want in max/msp if you outline your
process carefully and have the attention span to realize your idea.
ok enough OT for me, sorry all
----- Original Message -----
From: "Thomas D. Cox, Jr." <[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Thursday, July 19, 2007 8:27 PM
Subject: Re: (313) track identifier future gadget machine
On 7/19/07, /0 <[EMAIL PROTECTED]> wrote:
reminds me of a project I did in max/msp to find copyrighted audio.
it worked by extracting and creating transient profiles, and flagging
anything with an 80+ percent match to a transient profile on record
fun stuff
what was the purpose for doing that? i mean, it seems like that would
only work when comparing a CD to a sample of that CD that hasnt really
been altered by compression, limiting, or any other dynamic
processing. it also seems like it wouldnt work if the recording was
sufficiently noisy, the transients from a vinyl recording with lots of
pops and crackles should be wildly different from a pristine CD
version.
tom
