Blackmagic Shuttle and DNxHD

P

paulcurtis

Veteran
A very quick one as i'm up to my eyeballs in shooting right now. Blackmagic claims that the shuttle is recording 0-255 via DNxHD, all of my recorders seem to be using just 16-235. I'm hoping it's something with my workflow.

Has anyone experience of using DNxHD rather than uncompressed? I know the uncompressed stuff is fine.

thanks
Paul
 
Paul
i can confirm that the DNxHD175x codec does carry full range (0-1023) intact, you might want to take a look at the options for injest on your system and make sure that it allows full range at that end

d
 
Dermot,

Thanks, yes i can see the codec can carry that range, but i'm having some weirdness with whatever the blackmagic shuttle is doing. Perhaps you can sanity check what's going on. Blackmagic themselves say that it is recording the full range.

I've a side by side of avchd, uncompressed and dnxhd. (I ran the out of one shuttle into another, one dnxhd and one uncompressed)

Take 3 files into after effects. The AVCHD and Uncompressed come in fine and very similar colourwise. The DNxHD comes in with the blacks lifted. If i apply a profile, a rec709 16-235 then the blacks look similar to the others (because at the end of the day this stuff is all REC709 out of these cameras). My overall concern is whether that container has clipped anything when normalising the 0-255 range into the 10 bit equivalent of 16-235... I hope not. In AE i apply a 16-235 profile and it very much depends how it handles a 10 bit file.

Here's where it gets weird:

If i set my working space to 32 bit float (as normal) and REC 709, i will get the same effect (as expected). The DNxHD is lifted, the others look the same.

If it set the working space to P3, then the uncompressed/avchd look the same but the dnxhd looks correct from the black point but the colours and saturation are really very different (actually look better). I don't understand why the working space change isn't affecting all the versions relative to each other though. How are these DNxHD being tagged so that a P3 working space picks them up so differently from the other files. They should all be REC709 natively?

cheers
paul
 
alot depends upon the decodeing even if the material is ful range, there is a pile of choices that software makes... all Qt decoders are (really really) not equal
Check out this excelent rant on the subject;
http://vitrolite.wordpress.com/2010/12/31/quicktime_gamma_bug/

I don't use PP or any Adobe software, i get expected resualts when i use Avid's implementation - be it capturing or linking in Avid|DS

end game is that what you see depends upon the de-compression, can you try it on the desktop in VLC with YCbCr set to 709, input levels set to 0-1023, and output levels set to what ever you want to work in... seems like the Qt decoder in Adobe auto-selects "standard", hence the elevated blacks... on the good side that bodes well for it not clipping anything, instead scaleing full to video levels, then sending that out the door

I would look at Adobe's options at input, i know DS & MC/Symp pretty well.. they give you options to select.

If there is no choices in Adobe, then i would log a request with them to implement that at their end

d
edited to add;

Oh, and P3 is full range, i'm not atall sure how (or even why) it's implemented in Adobe, on most software it's not an option as they are all full range already, and P3 is ONLY your choice of FL, gamma and XY on your monitor

So the idea there's even a "P3" option leads me to think the Adobe is doing something i would not expect with "non-P3"... i have no idea what they are doing, or why... but it seems pretty strange way of dealing with something so simple

Kinda reminds me why i don't use Adobe really ;-)
 
Last edited:
You must log in or register to reply here.
Back
Top