On Mon, May 16, 2016 at 12:26 PM, Kieran Kunhya <kier...@obe.tv> wrote: >> Testcase is fate-aac-pred-encode >> >> Signed-off-by: Michael Niedermayer <mich...@niedermayer.cc> >> --- >> libavcodec/aacenc_is.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/libavcodec/aacenc_is.c b/libavcodec/aacenc_is.c >> index 473897b..e5cfa14 100644 >> --- a/libavcodec/aacenc_is.c >> +++ b/libavcodec/aacenc_is.c >> @@ -64,6 +64,9 @@ struct AACISError ff_aac_is_encoding_err(AACEncContext >> *s, ChannelElement *cpe, >> abs_pow34_v(I34, IS, sce0->ics.swb_sizes[g]); >> maxval = find_max_val(1, sce0->ics.swb_sizes[g], I34); >> is_band_type = find_min_book(maxval, is_sf_idx); >> + >> + av_assert0(minthr != 0.0); >> + >> dist1 += quantize_band_cost(s, &L[start + (w+w2)*128], L34, >> sce0->ics.swb_sizes[g], >> sce0->sf_idx[w*16+g], >> -- >> 1.7.9.5 >> >> >> > Does this assert on actual input data?
A threshold of 0 would in theory cause a zeroed band (zeroes[i] == 1), and those should be skipped. I think the proper fix would be figuring out why those aren't being skipped, if that is the case. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
