[FFmpeg-devel] [PATCH 10/11] dca: factorize scaling in inverse ADPCM

Christophe Gisquet christophe.gisquet at gmail.com
Fri Feb 7 20:20:25 CET 2014


2014-02-07 Michael Niedermayer <michaelni at gmx.at>:
> like the first iteration could be done with sum = instead of +=

That's the only thing that causes a change (haven't checked
before/after to spot the modification).

> the adpcm_vb[s->prediction_vq[k][l]] and &subband_samples[k][l][m] and
> &s->subband_samples_hist[k][l][m+4]
> could be factored out

The loops are actually completely unrolled, whether for win32 or
win64. No jmp/call at all.

> the / 8192 could be changed to * (1.0 / 8192)

In fact this gets properly compiled already. It was a local test
program (fast-math?) that didn't have the proper change. But I agree.

So, what about only the first and last changes?


More information about the ffmpeg-devel mailing list