Is there a recommended practice as to where a lut should be applied (Display, input, output)?
I would assume that if i apply a LUT to the display on the deliverable then it would be applied to all clips of that deliverable and that i would not be able to remove it without modifying my deliverable or am i missing something?
When should the LUT be applied as input vs output.
Regarding Display LUTs, the assumption above is correct. When you apply a LUT as a Display LUT in the deliverable config, it is applied to all clips on that deliverable and cannot be removed.
If you require the ability to selectively apply a LUT to each clip, then doing so as an Output LUT instead would be the best approach.
The processing order is like this:
Source Image > Input LUT > CDL Values > Output LUT > Display LUT
Both Output LUTs and Display LUTs are applied downstream of CDL values.
So if you just have a single LUT and CDL values, using an Output LUT or a Display LUT will give the same results:
Source Image > -- n/a -- > CDL Values > -- n/a -- > Display LUT Source Image > -- n/a -- > CDL Values > Output LUT > -- n/a --
Input LUTs are applied before the CDL values, on the other hand, so you’ll get different results with the same LUT if you are also applying CDL values.
Source Image > Input LUT > CDL Values > -- n/a -- > -- n/a --
Which effect is preferred will depend on the workflow.
In general, for something like a LogC workflow using a REC709 LUT, I recommend using it as an Output LUT. Putting it downstream of the CDL values means your creative color is working in LogC space and that you get the benefit of any soft-clipping built into the LUT. This also gives you the flexibility to turn it off on footage from cameras that shouldn’t have it applied.
That said, its more important to be matching what will be done upstream or downstream in the workflow at times. So if they are using an Input LUT on set or in final color, then you’ll want to do the same, otherwise you’ll get different results.