[tremor] [PATCH] another additional 15% performance gain

[Nicolas Pitre]

On Thu, 10 Oct 2002, timmy brolin wrote:

> May I suggest that you also include the linear interpolation of the large
> 4096 and 8192 size sin and cos mdct tables under the _LOW_ACCURACY_
> define, and keep the full tables by default? After all, linear
> interpolation does result in lower accuracy.

Well from the testing I made the current linear interpolation are really 
accurate, especially if your output samples are 16 bit wide, since it 
happens in the last part of mdct_backward.  There is actually nothing to 
gain by keeping the full table at that point.

<p>Nicolas

--- >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 'tremor-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.