-
Floating Point vs. Brooadcast Safe?
Hi,
I have another question about Floating Point that is puzzling me.
For broadcast, it’s great that you can have these ‘whiter than whites’ and ‘blacker than blacks’but, in the end, aren’t you going to have to scale all the values back anyway to make them broadcast safe?
… or am I just misunderstanding the whole issue?
If anyone could post a simple workflow for using Floating Point that is ‘broadcast safe’ that would be great.
Thanks in advance,
ALEX
-
5 Replies
-
[Alex Kuzelicki] “For broadcast, it’s great that you can have these ‘whiter than whites’ and ‘blacker than blacks’but, in the end, aren’t you going to have to scale all the values back anyway to make them broadcast safe?”
You’re correct in that the 32bpc/floating point color won’t be carried into the rendered file. However, you’re never explicitly seeing the overbrights in the composition anyway, anything above 1.0 (white) gets clipped to 1.0 for display. The difference is that if a pixel has a value over 1.0 it will change the way that pixel is composited in the scene. A bright light (say a value of 5.0) when blurred or motion blurred will give a much ‘hotter’ more realistic result, all of the pixels displayed will still reside between 0.0 (black) and 1.0 (white), but the compositing will be more natural. To summarize, 32bpc is color range used for compositing, not display.
The final render will be an 8bpc file (or maybe slightly more depending on the codec you use) with all of the values ranging between 0 and 255. If you need the final render at 16-235, I believe you can get away with dropping your 32bpc comp into an 8bpc render comp with a Broadcast Safe (or Levels) filter, I can’t test that right now as I don’t have access to AE.
Darby Edelen
DVD Menu Artist
Left Coast Digital
Aptos, CA -
Indeed. 32bpc and float gives you more “headroom” or margin for error when compositing or applying effects.
-
try applying 20 drastic curves adjustments to an image, then view it in 8 bit and then in float. You should see a pretty big difference.
Pete O’Connell
-
andrew kramer has two turorials that demonstrate benefits of working in 32bpc…
https://www.videocopilot.net/videotutorials/earthquakehdr/index.htm
https://www.creativecow.net/articles/kramer_andrew/hdr/index.htm
Kevin Camp
Designer – KCPQ, KMYQ & KRCW -
But what, if you want HD/16bpc output (f.i. in Avid DNxHD codec)?
Which values are safe? Can’t find those values anywhere. Even not in the ITU/EBU documents.
Thanks!
Jan
Reply to this Discussion! Login or Sign Up