Post-production workflow: VFX before color correction?

Question

I'm shooting my first internet video tomorrow. It has a spaceship chasing and shooting at a superhero in it. With knowledge accumulated between ~a year ago and now and more focused research over the past week, my idea of post-production workflow for a VFX video in the style of FreddieW or Corridor Digital YouTube channels is something like this (prerequisite: I am talking about color correction as getting all shots into a uniform look color-wise, and color grading as getting the feeling you want from the color):

1. Edit raw shots into a sequence, and lock them there.
2. Do music, sound and VFX. Go through each shot in the video and add VFX elements, matching them to the shot's lighting, exposure, color, shadows, ...
3. Go through each composited shot and color correct and grade it.
4. Replace shots in locked edit with composited and color corrected and graded shots.

I have now these questions:

1. Is this workflow "correct"? Are there many "correct"'s (is it mutable)?
2. You're supposed to shoot in as low contrast and saturation as you can (this is called flat?) to make color correction easier, like you have a blank, unbiased canvas to "pull" the colors out of, right?
3. This is the real core of my question: how can you match VFX elements to flat footage when there's no amplified color detail to work with? Can I color correct before VFX, and leave color grading where it is?

Answer 1

Any workflow that works for you is "correct". I don't see any obvious problems with your proposed workflow, though I would think you would want to do sound after VFX so that it can be properly timed to the VFX. Depending on the software you are using, you may not have to lock the shots but may actually be able to move directly in to working on the shots.

As far as low contrast and low saturation, I would personally disagree with shooting for low contrast/low saturation. This limits the depth of your available color space too much. Your camera has a limited number of colors it can describe, if you ignore a large portion of that color space, then you can't describe as many colors and the resolution of your color after boosting the saturation and contrast will be very greatly reduced. That said, if you have issues with lacking detail in your shadows, exposing it up in to mid-tones can help since your format doesn't actually support using that portion of the dynamic range anyway, so you wouldn't be losing anything meaningful.

The key is that you want to make strong use of the color space your camera can capture and encode that way you have the highest possible use of the color space as long as you are careful not to exceed the color space (color clipping needs to be avoided). It is always easy to make colors more subdued and to lower contrast as this is moving from a high color resolution to a low one, but you can't move the other way because you don't have the detail to fill in.

What flat means is not necessarily low contrast or low saturation, but rather that the colors are not heavily tinted and that details are well represented within the usable portion of the color space. They can be fairly vibrant, but they have a neutral response and gamma. In other words, you want it to evenly capture detail across the DR, but you also want the DR to be fully utilized. Think of flat as the way the curve looks in the Photoshop curves tool. Flat would be a straight line rather than a curved one. It can still have very strong contrast and saturation since the DR is being fully used.

For the third part, it really doesn't matter. Either you mix to subdued colors and boost everything or you mix to vibrant colors and bring it down. Either way, as long as it fits at some point, it will fit when you adjust the color grade of the entire frame if it fit well before.

Answer 2

I'll echo AJ on most of his response, but I would caution against doing color correction after compositing unless the elements are already fairly congruent. Ideally you would apply at least gross color correction to each layer or element.

The notion of shooting 'flat' or low contrast IMO is a holdover from film, where you could pull contrast from the negative in a subsequent transfer or interpositive step. It's really not wise in a pure video / digital workflow. You do want to avoid clipping on either end, or oversaturating, but otherwise, use the range you have available. You can always compress or clip later, but undoing premature work later can be tough or impossible.

Answer 3

One really effective way to match VFX elements to real ones is to manually correct each channel. What you want to do, is start with let's say the red channel, match the lows, then match the highs, and then the midtones. Once you repeat it for the rest, they all basically fall into place.

Just to clarify, your footage color correction should only be to fix the exposure and any color errors in your footage. It shouldn't include any grading.