CG and Super16 Workflow...?

Matt Irwin · January 15, 2008

I am shooting a low budget short on super16 soon, and there are a few scenes that involve CG elements-- glowing plasma figures of different colors, dark matter, etc.

I've got some ideas about how to light for those effects, but I am a bit in the dark regarding post workflow... This is a bit of a first for me.

As of now, we are transferring to and finishing on either HDCAM or D5. I'm wondering at what stage the CG elements should be added--- is it done on a cut-by-cut basis, are they added in the online? After grading? I am concerned about when the image and the CG are baked together in case we need to adjust color on one to mesh it better with the other.

The effects are being done by an individual, not a post house, so resources are limited. I'm looking for (of course) the most economical way to do this while maintaining a good-quality blending. Most involved in the production seem to be looking at me for answers.

Any information about workflow for this helps.

Many thanks,

Scott Fritzshall · January 15, 2008

As of now, we are transferring to and finishing on either HDCAM or D5.

If HDCamSR is out of the question, I would recommend D5 over HDCam; its got less information loss due to subsampling.

I'm wondering at what stage the CG elements should be added--- is it done on a cut-by-cut basis, are they added in the online? After grading? I am concerned about when the image and the CG are baked together in case we need to adjust color on one to mesh it better with the other.

Your general post workflow will look something like this:

1. Process your negative and get your offline transfer

2. Offline edit. While doing the edit, you can give the effects shots to your VFX artist and have him do previz effects so that your editor can see how their timing will affect the cut. You can fine-tune exactly what you want to be seeing at this point, but it's just low res so your artist can crank out a bunch of versions.

3. HD Transfer based on EDL. If there is any extraneous material that needs to be captured for the VFX, you can add it to the EDL or create a seperate VFX EDL.

4. Capture the HD footage and get it on a hard drive; give the VFX shots to the artist [usually as an image sequence, this is the easiest and safest way]

5. The artist comps the shots and will [hopefully!] try to match them up to each other himself and maintain continuity between them

6. He gives you the completed shots as image sequences and you give them to the online editor. The online editor plugs them into the timeline.

Grading can happen either during Step 3 or Step 6, depending on your workflow and exactly what you're doing. If you're transferring to HDCam or D5, you'll get far more latitude when grading if you do it during the transfer. The downside of this is your VFX artist has very little room to tweak the colors; partially because the look is already set, and partially because there is less latitude remaining for him. For what you're doing in this particular film, however, this may work out for you just fine. The other option is to get a flat transfer, and grade during the online. This gives your VFX artist a bit more information to work with, but since you're now stuck in 8bit linear color space, there is less information for your online artist to work with when grading, so it limits how hard you can push the grade. Or you could kinda go halfway at step 3 and the other half at step 6. Talk with your colorist and your VFX artist to find out what their thoughts are regarding this project.

There is normally some color correction done during the online no matter what, and this is when you'll even out color differences between shots in a sequence. In general, unless you have a very specific reason to do otherwise, you should treat your VFX shots like any other shot, and just grade them to fit in. The grade during the online gets put on top of whatever color correction the VFX artist had applied to it. Depending on which system you're doing your finishing on, you might be able to import an image sequence of alpha channels and use that to selectively grade just the CG part of the frame. This can sometimes give you more control over it, but I'll caution you that you're probably more likely to screw it up this way than to get the result you're looking for. If your VFX artist knows what he's doing, he should be able to get the shots looking pretty consistent on his own.

Does that help?

Robert Houllahan · January 15, 2008

As of now, we are transferring to and finishing on either HDCAM or D5. I'm wondering at what stage the CG elements should be added---

I would try to get a pin registered scan (to DPX files) instead of a telecine transfer for this stuff, every bit of stability will help with compositing elements, there are allot of scanners in LA I would think you could find a good deal on a northlight or something similar...

-Rob-