[Tig] existing digital cinema cameras
Sat May 6 22:46:51 BST 2006
Thanks to oktobor for supporting the TIG
Ok, let's talk about some cameras that are actually shipping right now.
Reel Stream makes a modification of the DVX100 ($4,000MSRP) that lets you
capture uncompressed off the camera's CCDs, before it hits the DSPs. It
does about ~9.5stops, depending on how you count. An entire system (camera,
computer) should run under $20,000.
Dalsa makes the Viper of course. ?They claim 10-12 stops.?
You can see what the cameras are capable of at the following webpage (scroll
to the bottom for the gist of it):
I have shown the output from the cameras in conjunction with a prototype
color enhancement filter I am working on. The reason of this is clear if
you look at the images. The output from these sorts of cameras is
'linearized' (i.e. similar to lowering gain)... this leads to a flat-looking
image lacking in contrast and saturation. There are a few ways to increase
A- RGB curves (these are inherent to film). What I dislike about this
method is that the color shifts are very noticeable... most easily seen in
B- Use levels / pedestal + gain. This throws away dynamic range, which is
not necessarily a bad thing.
C- My algorithms. I'm not saying that they're the best way of doing things
(they're not), but the look is a little more naturalistic than that of RGB
curves. You could still argue that there are some things about my
algorithms that looks unnatural.
Anyways, the point is that video origination can look at least as good as
what's demonstrated on my webpage. In my opinion, these new extended
dynamic range cameras are a leap forward in quality from traditional video.
For lower-budget production, I think it is becoming a more compelling
alternative to film (cheaper, arguably faster, and getting closer in
2- Problems with video:
A- Film still records greater amounts of information to begin with (i.e.
greater exposure latitude).
B- Grayscale tracking: At the upper ends of exposure, grayscale tracking
becomes progressively worse. This limits the amount of useful dynamic
range. In the Andromeda capture, you can see that the window looks
magenta-ish. I would expect that the camera manufacturers try to figure out
the transfer functions of their cameras to create 1-D LUTs to counteract
I expect a similar technique could be used on film to undo its s-shaped
transfer function... film would have similar color accuracy to video.
Tig list - http://tig.oktobor.com/mailman/listinfo/tig
TIG wiki: http://tig.colorist.org/wiki
More information about the Tig