How do you plan to develop your production pipeline in the future?

Why grade dailies at all? Why not create LUTs right on locations and play them back with FrameCycler, or even SpeedGrade DI? For me, this is one link in the chain where things can go wrong: If I send dailies back to be graded, what I get back may not be what I was looking for. The problem is, with the demands of a production, there is just no time to send things back for changes, so we missed the chance to get it right the first time.

I think a much better way is to just create the dailies 'flat' in logarithmic space, without grading. Then we can apply the SpeedGrade .Look files right then and there. You can tweak the .Looks and once you are happy, you can save the results. Yes, it takes a bit more time out on the production, but this way you can be absolutely sure you have the look you were looking for.

Does your approach help the colorist?

With our workflow, we know we've got the looks right and the colorist knows exactly what he or she has to do. This means that when we're finishing the film, the colorist can work for three days or so before I even need to show up. Then I can spend a few days there tweaking things a bit. Realistically, this means that the whole film can be graded in a week. And when we get more facilities using SpeedGrade DI - so that the colorist can just load the .Looks and do the conform right in the system - it will go even faster than that.

Continue >