Controlling The Eye By Relighting With Motivated Power Windows

Jim – I’ll have to hunt down Dado’s approach to using composite nodes with stacked Power Windows. I also agree, it’s a huge mistake if your power window is NOT invisible.

When doing relighting, I know I’m on the right track when the client (or DP) asks if we can add a Window – then I toggle what I’ve already done and their response is: “Oh! Holy cow. Nice job. No worries.”

Valid points andi – but I think that anything we do, can fall into that scenario. I have always felt that if I notice something obvious in a film that is distracting me from the story, that it is probably overdone. Anything we do is stepping on the photography to a point. The on set photography sometimes has mistakes, and when there is a long or mid shot where the subject has to be lit from a distance, can change drastically when the scene is relit for a close-up.
Our job in my opinion would then be to give that consistency between the cuts. On set they may not see it in sequence pf the edit and it can be subtle enough change that the DP seeing it in context might agree.
We should enhance not change – so I do agree. But you may have watched thousands of films that the colorist did it correctly and if they did, you might not have noticed it.
I think direction of the viewers eyes, can be done to enhance the story and not be obvious even to the point of subliminal suggestion.

Andi –

I can definitively state: I use far fewer windows today than I did 10 years ago – because I generally agree with you. Mostly, it’s too easy to destroy an image with a poorly implemented shape. Plus, as you move up in the size of budgets you’re working on, you’re more likely to get images that don’t really need them. This is especially true of single-camera narratives with good budgets.

Also, windows are almost always the last thing I do after balancing, shot matching, and laying in a look. Then, if I do use one, it needs to be motivated by *something*. The light I’m adding (or subtracting) has to be based on some sort of internal logic.

I have *definitely* seen too many bad power windows, even on tentpole series for the major streamers.

I think it’s important if you’re going to do relighting that you need to check your work on a large format display. I’ll frequently turn and re-watch a scene on the client 55″ OLED, because that’ll reveal artifacts that aren’t obvious on a smaller 24″ reference display or (even worse) on the UI display inside the software.

As I said: this is an age-old conversation with lots of differing opinions and approaches 🙂

This depends mostly on the budget etc. and what the DP was able to do with the time he had.

I do mostly TV these days and i would say most DP’s i work with absolutely count on me to use windows, especially on single camera comedies with large cast. The  just don’t have the time to do on set what they wanted/needed to do and that’s our job to make it right. Power windows are a very large part of what i do and i honestly cant think of very many shows that didn’t need them all over. I probably have at least 1  or 2 on every shot and you would never know. I like to think in terms of shaping every shot and i rarely get any push back on that but it does happen occasionally . I like to get a feel up front from the DP as far as that goes. It’s easy to have a quick conversation at the start to get an idea of his or hers taste and needs as far as whether or not they will be needed.

I don’t have a hard & fast rule for Power Windows since there are so many uses for them.

I usually do ‘controlling the eye’ Power Windows (as discussed in this Insight) after my main balancing and shot matching. If I’m using LUTs or a Color Space Transform in the node tree, I usually place those nodes before the ‘controlling the eye’ Power Windows.

If I’m using a Power Window as a corrective/fix, then I like to do those before the LUT/Color Space Transform. But I don’t have a militant opinion on this and will frequently change up the placement of my LUT/Color Space Transform node on a per-project basis – just to test and see if I like those results better.

Sometimes I’ll test two or three different node trees, with the placement of all these nodes in a different order, at the start of a project to see which results I like best and then commit to that node structure for that project.

Did I answer your question?