On 10/15/26 16:28, Jaka Bac via ffmpeg-devel wrote:
CID 1271 (DNxHR HQX) uses incorrect luma and chroma weight tables.
Currently it uses tables from CID 1241 (marked D.4 in the standard)
which is for interlaced media.
It should use tables for CID 1235 (marked D.1).
Signed-off-by: Jaka Bac <[email protected]>
---
libavcodec/dnxhddata.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/dnxhddata.c b/libavcodec/dnxhddata.c
index ce06d77f18..3d113abfdb 100644
--- a/libavcodec/dnxhddata.c
+++ b/libavcodec/dnxhddata.c
@@ -1049,7 +1049,7 @@ static const CIDEntry dnxhd_cid_table[] = {
{ 0 }, { 57344, 255} },
{ 1271, DNXHD_VARIABLE, DNXHD_VARIABLE, DNXHD_VARIABLE, DNXHD_VARIABLE,
0, 6, DNXHD_VARIABLE, 4,
- dnxhd_1241_luma_weight, dnxhd_1241_chroma_weight,
+ dnxhd_1235_luma_weight, dnxhd_1235_chroma_weight,
dnxhd_1235_dc_codes, dnxhd_1235_dc_bits,
dnxhd_1235_ac_codes, dnxhd_1235_ac_bits, dnxhd_1235_ac_info,
dnxhd_1235_run_codes, dnxhd_1235_run_bits, dnxhd_1235_run,
--
2.51.0
I believe this is correct, but needs to update some fate tests that
exercise this code.
Can you open a PR on code.ffmpeg.org so it can run through automated tests?
Timo
_______________________________________________
ffmpeg-devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]
