All of the recent posts about workflows, workarounds, and discoveries using the RED One camera brings back memories of a warning I issued when DV cameras first became popular among independent shooters and moviemakers: Be careful what you wish for; you might get more than you expected.
Many hold up film as the gold standard of footage acquisition. Here’s the funny thing about film: A lot of it looks terrible straight out of the can. Quite often, the goal of film is to capture as much image information as possible, so that a lab, colorist, or other magician can then mold this ample lump of clay into something that looks truly gorgeous. This is why colorists make so much money (with press releases announcing their movement between facilities), and why a Da Vinci color station is a machine of modern mythology.
DV format cameras really juiced the documentary film field, allowing shooters to capture their footage for a lot less cost. However, when we started loading this footage straight into our computers, we skipped that all-important colorist stage where mere images were turned into magic. If you were on a budget and the lighting wasn’t quite up to snuff either, this new reality was even more challenging. Suddenly, you - not the lab - were responsible for adding that special something to your shots. (Fortunately, there are several books on the subject; I particularly like Color Correction for Non-Linear Editors
The good thing about DV cameras was that they were at least point, shoot, import, and use. In other words, the lump of clay was already molded into generally the right shape; you just had to polish it a bit. The RED One with its special “raw” format puts you even closer to film in that you’re getting an unshaped lump of clay which you must now “develop” as well as color correct to taste. (For example, read Stu Maschwitz’s excellent article on Digital Cinema Dynamic Range.)
This realization is causing a bit of a backlash among those who expected the RED to be point and shoot magic. But the fault isn’t completely with the RED; it’s more with some of our expectations. You wanted digital film? You got it. Oops - now you have to learn how to handle “film” (in addition to learning how to be a colorist). And the RED tools are still in their infancy.
This is going to cause some pain for awhile. Pioneers are the ones with arrows in their backs, and all that. Fortunately, some (including writers on PVC such as Adam Wilt and Art Adams) are sharing information on where they got shot, how they dressed the wound, and ways to turn those arrows into advantages. Don’t shoot the messengers; be patient and learn. No one has all the answers yet. But groping together, we can eventually find them.
Cameras • Production •
(Page 1 of 1 pages for this article )
Well put indeed Chris. Yes, the advantage of capturing ‘Raw’ data is great..but as you say, it looks terrible to the un-trained eye, just like most film negative, prior to grading etc...I just loved shooting on the SI-2K camera since it allows the DOP/Director to “play’ with the look on set, arrive at a ‘look’ they both like....but when finish, since the incoming Raw data is ‘indistructable’ and therefore separate from the meta data graded look dialed in....one can always start again and go back to the raw digital negative/unmanipulated footage...Nice. Is this possible on the RED as well?...I have not had the pleasure of trying it out yet...Will soooon.
tai krige sasc
Posted by on 03/24 at 01:43 PM
Thank you Chris & Trish for putting us onto the fab, interesting article on Stu’s Dynamic Range of the RED.- This is always my problem with HD capture devices...Dynamic Range, or in most cases the lack thereof. As stated, I hav’nt had the pleasure of trying the RED ONE but I shot tests on the SI-2K and fully zoomed-into the pitch black, (to the eye), behind the front seats of a white truck, I then zoomed fully out with direct sun on curve of the roof flaring into the lens, ( my spot meter was at f92 on the roof )....one could make out detail in the black behind the seats, see detail in the white roof and white clouds in the sky. I was very pleasantly surprised...at least 10 to 11 stops. ( camera rated at 320asa ) And ‘they’ said HD could not match film......its pretty damn close.
Posted by on 03/24 at 02:03 PM
It might be daunting to have to “develop” your RAW images, but it’s not too bad, and in some ways easier than dealing with in-camera processed images - let me explain....
Say you shot on your DV or HD camera, and the white balance is wrong. Now you’ve got to fix that. Well, fixing a wrong white balance and making it look good is not too hard, but it can be a bit of a pain. But in RAW, you just either find another shot in the same lighting that has the correct white balance and copy it’s numbers over, or use a picker, or type in a good number for that light and you’re done, and there’s no compromise to the image - it’s as if you’d got it right first time. We all know that white balance should be performed on linear light data, but if you shoot on a video camera, with all it’s knee and gamma processing, you cannot really re-white balance that without screwing up the image somewhere else in the tonal range, because there’s now this curve on the data. With RAW, the white balance control will work correctly on the linear light data, and all is good, and you’ll get great greyscale tracking up and down the range.
Digital stills photographers have been working with RAW a lot longer than digital cinema people. They’ve certainly got the hang of it, know the use and understand the advantages.
You can play with the look of RED video by altering the metadata using the camera controls. Soon you’ll be able to upload a file to the camera from our apps that will let you round-trip the metadta.
Posted by Graeme Nattress on 03/24 at 02:24 PM
Thanks for the info and quick reply Graeme. good to know that about the RED....But forgive me, I am really just a ‘babe in the digital woods’ and am trying to get to grips with it all with very little experience...am really new to this, my background is mostly film. What is meant by; “… let you round-trip the metadata.”?
Posted by on 03/24 at 02:37 PM
Thanks! What I mean by “round trip” is to be able to alter metadata for the raw image decode in camera, have that follow through with the file to Redcine, for you to be alter the image, there, save out the new metadata and put it back into camera for next time you shoot.
Graeme
Posted by Graeme Nattress on 03/24 at 04:00 PM
“And ‘they’ said HD could not match film”
Well, they’re right - a normal HD camera (as in, a Rec.709 high def spec video camera) cannot match film - it does not have the dynamic range, clips too soon, etc. But the SI-2K you used (along with the RED) is not a normal HD video camera.
If you want to match film, you need a helacious amount of dynamic range - either something in Cineon format (which a lot of film is scanned into), or Raw like the RED. But then - it becomes your responsibility to decide what section of that dynamic range to take and squash into video space.
Some of the worst sins I’ve seen committed, for example, are by those trying to squeeze film editing into an HD workflow. On one indie film I worked on, they took Cineon scans of film and used the same default white point and highlight rolloff for all the shots, posterizing the hell out of some shots of light sources. I had to go back in and re-tweak the rolloffs on a shot by shot basis to recover the details for the HD video workflow - and remember the setting for each scene so we could expand the highlights back out and “round trip” when we printed to film. The post house - who had done a lot of HD work, but never film - didn’t even know detail was being lost. That’s the sort of things we need to become more aware of now.
- Chris
Posted by Chris Meyer on 03/24 at 07:31 PM
Thanks again to you all....lots to learn, very exciting.- Will certainly keep track of this website and all your many fine posts.....
Posted by on 03/24 at 10:13 PM
When the smoke clears a bit (crazy times right now), I’ll write up a short article on the importance of highlight preservation and place it over in our CMG Keyframes “channel.”
Posted by Chris Meyer on 03/25 at 08:42 AM
Chris Meyer,
Dynamic range hasn’t much to do with the Rec 709 space. Increased dynamic range will only mean that the sizes of the “atoms” of the Chromacity diagram become smaller, in any space.
One mismatch between film and Rec 709 that is at least admissible is the fact that the gamuts of Rec 709 and film on Chromacity diagrams don’t overlap everywhere.
So if there is a difference between the dynamic ranges of (any) two systems, care should always be exercised when talking about a particular manufacturer’s implementation of Rec 709; and based upon that inferring properties of Rec 709 space itself.
Posted by on 03/25 at 02:46 PM