David Balazic wrote:
> From: Mark Taylor <[EMAIL PROTECTED]>
>
> >Just a note on the highpass filter: at 44.1khz sampling rate,
> >we only have 576 MDCT coefficients on which the filter acts and thus
> >a frequency resolution of only 22050/576 = 38Hz. So the accuracy of the first
> >few coefficients is questionable, and a highpass filter at 50Hz would
> >only effect the first 2 MDCT coefficients. I dont know how big a problem
> >this is, but a true 50Hz filter would need to be done before
> >calling lame, using a much larger fft/mdct window.
>
> What is the purpose of this high-pass filtering ?
> You said that it would affect only 2 MDCT coeficients, that is
> less than a percent of them all, so what gain do you/we expect from it ?
>
> --
> David Balazic , student
> E-mail : [EMAIL PROTECTED] | living in sLOVEnija
> home page: http://surf.to/stein
> Computer: Amiga 1200 + Quantum LPS-340AT
> --
Well, I plugged it in. Consider the following situation:
At 16 kHz sampling rate one MDCT coefficient represents 8000/576 = 13.9 Hz.
For PHONE quality at 16 kbits we need only the frequency range
from around 300 Hz upto 3400 Hz. So a "highpass filter" with a sharp cutoff
at 300 Hz can drop around 21 MDCT coefficients, ~3.7%
I don't know how many bits we can save this way, maybe more than 3.7%?
A 10 Hz highpass filter was suggested in the ISO docs too.
And by the way, it's a feature, if you don't want it, don't use it ;)
There is no "highpass filter" activated by default, you have to
turn it on explicitly.
Robert
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
