Right now, we are using an attack detection that compares energies of 9
high-passed subblock, and this time-domain high passing is relatively
time consuming.
Idea:
Considering that most of the time we're going to do a long block fft,
and considering that a time domain transcient implies a spread into the
frequency domain, we could use this latter feature to speed up the
transcient detection:
*long fft
*check how much energy is there on high freqs
*if ratio of high freq energy is low, then there is no transcient => no
need to do the regular transcient detection
*else, do the usual transcient detection.
I think that this would save some computing time (especially considering
that transcient detection is an important part of our cpu time)
Only theoretical, as it's likely that no one will ever have time to try
it...
--
Gabriel Bouvigne
www.mp3-tech.org
personal page:
http://gabriel.mp3-tech.org-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/_______________________________________________
Lame-dev mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/lame-dev
