Full/Video workflow

Hi all,

I'm trying to get a sense of the most efficient Resolve workflow when exporting for broadcast TVC. Up until now I've graded between 64-940 and exported as full range and that's been fine, but I wondered if there was a benefit to grading between 0-1023 and exporting as video range to compress for broadcast.

So I tried this, and this is where my questions start to arise. Regardless of whether I selected Full or Video in export, the exported file still displayed as 0-1023 in other NLEs and even Photoshop. The weird thing is, if I reimport the two test exports into Resolve, the full range clip stays full range regardless of whether its set to Full or Video in the clip attributes, whereas the video clip will display as 64-940 when set to Full, and then will happily export in that range and will display as such in other NLEs and Photoshop.

There must be an easier way to go from 0-1023 to 64-940 without exporting, reimporting and exporting again? Is it a metadata thing rather than hard baked?

I'm always amazed how much people overthink this stuff. Stop worrying about 0-1023.

Just leave all the Resolve settings in default -- DaVinci YRGB, Rec709, 2.4 gamma -- and work entirely in Video mode. This is what you want for a standard broadcast/cable/streaming deliverable show.

Resolve will automatically scale (not compress per se) the bits for whatever ultimate format and bit-depth you're going to. If you're worried about it, include a second or so of SMPTE color bars in the "leader" prior to the show, like before the slate (assuming you do a slate), and check the slate on scopes. Worst case, upload it to YouTube or Vimeo and see how it looks.

I look on the waveform monitor in old-school terms, with "0" at the bottom and "100" at the top. If the signal still goes from 0 to 100 and nothing is clipped or crushed, and the color is in the right place, all is well. If it's screwed up, you'll be able to tell when you compare the render to the original.

I'm always amazed how much people overthink this stuff. Stop worrying about 0-1023.

Just leave all the Resolve settings in default -- DaVinci YRGB, Rec709, 2.4 gamma -- and work entirely in Video mode. This is what you want for a standard broadcast/cable/streaming deliverable show.

Resolve will automatically scale (not compress per se) the bits for whatever ultimate format and bit-depth you're going to. If you're worried about it, include a second or so of SMPTE color bars in the "leader" prior to the show, like before the slate (assuming you do a slate), and check the slate on scopes. Worst case, upload it to YouTube or Vimeo and see how it looks.

I look on the waveform monitor in old-school terms, with "0" at the bottom and "100" at the top. If the signal still goes from 0 to 100 and nothing is clipped or crushed, and the color is in the right place, all is well. If it's screwed up, you'll be able to tell when you compare the render to the original.

Ok, thank you, so grade between the 0-1023 lines and leave delivery setting as auto for whatever format the house wants. The only reason I’m overthinking it is because with the previous method the ad delivery lot would flag anything and everything to get a buck out of us (including audio deviations within the 0.5 lufs margin). Oh well, I’ll try it like this and see what they say!

Fill scopes as needed in Resolve. Trying to stay within 64-940 and then exporting full range is wrong approach.
Resolve will convert your data during export to limited range whenever codec requires it- your 0 will became 64 for YUV based codecs etc. It doesn't matter if your work for broadcast, youtube etc. - 0 is your black in Resolve not 64. Resolve doesn't really "care" about YUV range until you get to export stage.
You will get some overshoots in black or white (or gamut) after export, but those are related to compression. Those overshoots should still be not a problem as about everyone (who requires normalisation) uses EBU-R103 spec, which has thresholds to handle those possible compression "problems".
Same applies to audio- there is always some allowed error threshold. Working to perfection is nothing good or sometimes simply impossible.
Also- don't try to be broadcast safe as per Resolve feature. This is based on legacy analog signals and it's about useless these days.

There is one bug (at leats 1 known) though- exporting to DNxHD MXF won't properly scale Resolve internal full range data to video.