Open ZmGorynych opened 5 months ago
In CMAF BoG at MPEG145 it was noted that if the file undergoes any conversions this information would be necessary.
Should a profile be created to address this?
On Wed, Jan 24, 2024, 15:49 Dimitri Podborski @.***> wrote:
In CMAF BoG at MPEG145 it was noted that if the file undergoes any conversions this information would be necessary.
— Reply to this email directly, view it on GitHub https://github.com/MPEGGroup/CMAF/issues/45#issuecomment-1909048894, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGEYZJ7TLOFI6P3VFMD77DYQGFZBAVCNFSM6AAAAABB5CITN2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBZGA2DQOBZGQ . You are receiving this because you authored the thread.Message ID: @.***>
I would simply add a recommendation ("should") that the clli
box should be present. This remains compatible with old players and old files remain compatible.
On top of what @cconcolato says, we may add defaults in case they information is not present.
In the industry HDR10 implies PQ colorimetry, 10 bits, and presence of static metadata SEIs, namely SMPTE 2086, MaxFALL, and MaxCLL, the latter two often considered optional as they don't work well in live linear.
The only mention of static metadata in the CMAF spec is in footnote c of table B.1 where all of these are optional.
Should we consider a profile where SMPTE 2086 is mandated, along with the recommended MaxFALL / MaxCLL where possible?