Kostya <kostya.shish...@gmail.com> writes: > @@ -1019,11 +1022,11 @@ static int bink_decode_plane(BinkContext *c, > GetBitContext *gb, int plane_idx, > ublock[*scan++] = get_value(c, BINK_SRC_COLORS); > break; > case INTRA_BLOCK: > - c->dsp.clear_block(block); > - block[0] = get_value(c, BINK_SRC_INTRA_DC); > - read_dct_coeffs(gb, block, c->scantable.permutated, > bink_intra_quant, -1); > - c->dsp.idct(block); > - c->dsp.put_pixels_nonclamped(block, ublock, 8); > + memset(dctblock, 0, sizeof(*dctblock) * 64); > + dctblock[0] = get_value(c, BINK_SRC_INTRA_DC); > + read_dct_coeffs(gb, dctblock, bink_scan, > bink_intra_quant, -1); > + c->bdsp.idct(dctblock); > + c->bdsp.put_block(dctblock, ublock);
Why doesn't this use idct_put() with an output stride of 8? Seems to me would do the same thing with less code. -- Måns Rullgård m...@mansr.com _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel
