On Sun, Jan 28, 2018 at 4:54 PM Marton Balint <c...@passwd.hu> wrote:
> > On Thu, 25 Jan 2018, Ray Tiley wrote: > > > On Tue, Jan 23, 2018 at 9:07 AM Devin Heitmueller < > > dheitmuel...@ltnglobal.com> wrote: > > > >> Hi Ray, > >> > >>> On Jan 22, 2018, at 11:20 PM, Ray Tiley <rayti...@gmail.com> wrote: > >>> > >>> I'm reading 334-1:2017 Sec 4 > >>> "When the ANC packets defined in this standard are carried in a high > >>> definition signal, they shall be carried in the Y stream." > >>> > >>> I couldn't find anywhere in the document where it calls out standard > >>> definition > >> > >> Right, so my understanding was that allowing the ability to use both > luma > >> and chroma was a result of having less space in the VANC to hold data, > >> compared to HD resolutions where that is much less likely to be an > issue. > >> That said, I’ve never seen an implementation that actually puts it in > the > >> chroma. > >> > >>> > >>> Conversation is here: > >>> https://video-dev.slack.com/archives/C0XKDDH5Y/p1516141555000074 < > >> https://video-dev.slack.com/archives/C0XKDDH5Y/p1516141555000074> with > >>> kier...@obe.tv <mailto:kier...@obe.tv> who I believe is in the > >> ffmpeg-devel IRC. > >>> > >> > >> Ah, ok. That’s Kieran. He’s really knowledgable in this area, > although I > >> cannot see the conversation you’ve linked to as it seems that room is > only > >> accessible by people with email accounts from certain domains. > >> > >> Devin > >> _______________________________________________ > >> ffmpeg-devel mailing list > >> ffmpeg-devel@ffmpeg.org > >> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > > > > > Please find updated patch attatched. I reverted the vanc lines changes > and > > found that all my tests worked as expected, so not sure what was wrong w/ > > my original test. The need to extract vanc from the entire line vs just > the > > luma in NTSC is still required. > > Is interleaved VANC data is always the case for SD NTSC, or it depends on > the equipment? Maybe we should add some auto detection, and check for the > first VANC header. E.g. > > if (ctx->bmd_mode == bmdModeNTSC && > (*(uint32_t*)buf & 0x3ffffffff == 0x3fffffc00)) > unpack_v210 > else > unpack_luma_v210 > Sounds good to me. Will send patch shortly. > > Also, later on in the code, you only parse half of the decoded VANC. > If you decode from both the luma and chroma, get_metadata should be called > with frame width*2, no? > Good catch. I was testing with very small caption packets so I did not see it cut off. Will send patch shortly. > > Regards, > Marton > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel