[vorbis-dev] New LSP code committed

[Segher Boessenkool]

> 
> > The cos itself isn't that heavy (on pc processors; embedded is different,
> > of course).
> > 
> > Oh, you can calculate the cos'es by CORDIC; no need for a look up table
> > or actual cos() calls. Expect a patch soon :-)
> 
> (describe CORDIC.  I think I know what you mean, but I want to be sure).

Actually, I wouldn't use CORDIC for this. It was a thinko. A simple iteration
will do:

w_0 = 1.0;
w_1 = cos(theta);
w_{n+1} = 2 * cos(theta) * w_n - w_{n-1}

So keeping the last two w's allows us the calculate a new one.

> 
> The p/q lspcos loop needs very fine angle resolution... but the same lookup 
> for that would also work for w.  These are the only cos() calls outside the 
> iMDCT.

Do the lspcos's or the w's need the resolution? If the lspcos's do, yes,
use the same table for w lookup. If it is fast, of course.

Ciao,

Segher

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.