Hi...
Thanx for the reply..
Actually block-switching is required to detect the transients, or
portions where the audio is differing from the neighbouring area.
If 2 neighbouring areas differ in characteristics, then the short
block is used. High pass filtering doesnot work I guess. This is
because the high pass filtering will give somewhat same energy
content for the following 2 case:
1. 2 areas with 2 different frequency but both are are high frequencies.
>
> 2. Single area with the high frequency.
>
> with regards
> > I need a algorithm which does block-switching decision based upon samples in
> > Time -domain only.
> > Please give me some reference, papers etc for this.
> > Any useful comment/idea will be very helpful to me.
>
> LAME with nspsytune does it.
>
> The algorithm is simple.
> fs/4 High-pass filter
> -> check if there's big energy (of time domain) change
> -> if yes, use short blocks. if no, use long blocks.
>
>
> And, LAME4 (CVS tag "takehiro-2002_05_07-experimental") does it, too.
> But it uses subband filter output instead of fs/4 high-pass filter.
> --
> _______________________________________________
> Surf the Web in a faster, safer and easier way:
> Download Opera 8 at http://www.opera.com
>
> Powered by Outblaze
--
_______________________________________________
Surf the Web in a faster, safer and easier way:
Download Opera 8 at http://www.opera.com
Powered by Outblaze
_______________________________________________
mp3encoder mailing list
[email protected]
http://minnie.tuhs.org/mailman/listinfo/mp3encoder
