Hi, On Thu, Aug 4, 2011 at 7:00 AM, Stefano Sabatini <[email protected]> wrote: > --- > libswscale/utils.c | 9 +++------ > 1 files changed, 3 insertions(+), 6 deletions(-) > > diff --git a/libswscale/utils.c b/libswscale/utils.c > index f5ef1c6..163ed3a 100644 > --- a/libswscale/utils.c > +++ b/libswscale/utils.c > @@ -1079,12 +1079,9 @@ int sws_init_context(SwsContext *c, SwsFilter > *srcFilter, SwsFilter *dstFilter) > flags & SWS_FAST_BILINEAR ? "FAST_BILINEAR C" : "C"); > } > if (isPlanarYUV(dstFormat)) { > - if (c->vLumFilterSize==1) > - av_log(c, AV_LOG_VERBOSE, "using 1-tap %s \"scaler\" for > vertical scaling (YV12 like)\n", > - (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) ? "MMX" : > "C"); > - else > - av_log(c, AV_LOG_VERBOSE, "using n-tap %s scaler for > vertical scaling (YV12 like)\n", > - (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) ? "MMX" : > "C"); > + av_log(c, AV_LOG_VERBOSE, "using %s-tap %s \"scaler\" for > vertical scaling (YV12 like)\n", > + c->vLumFilterSize == 1 ? "1" : "n", > + HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX ? "MMX" : "C"); > } else { > if (c->vLumFilterSize==1 && c->vChrFilterSize==2) > av_log(c, AV_LOG_VERBOSE, "using 1-tap %s \"scaler\" for > vertical luminance scaling (BGR)\n"
My comment here is the same as for the others: I don't think we should show what optimization levels are used. A) the C may be wrong for e.g. PPC (Altivec) B) it will be outdated whenever I implement e.g. a SSE optimization C) the "mmx" is incorrect for anything not 8bit (that goes for the 1 vs. n-tap filter also) These things tend to get old and be wrong. The only reason to list it is pride, and let's be honest, it's 2011, mmx isn't that new anymore. Ronald _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
