Cary Knoop

Use Layers (or Parallel nodes) overlaying different intensity pastel colors.

Attached a simple example with parallel nodes. With layer mode you can experiment with different composite modes.

Pastel_1.1.1.drx

On 1/10/2018 at 2:38 AM, Thomas Singh said:

Does an log image have linear gamma? Is that the same as a linear line that can be seen on a ramp from black to white? What happens when you apply an s-curve to this linear line, is the gamma curve then re-defined? 

It all boils down to semantics as gamma is in the video world often used to describe any kind of curve.

Technically gamma refers to the exponent (or the reciprocal value) of a function and defines how much a linear function is "pulled up or down".

For instance below is a graph with various gammas, starting with the red line which is x to the power of 1 thus a gamma of 1, then going upwards with a gamma of 2, 2.2, 2.4 and 2.6 (or the reciprocal values).  So for instance x to the power of 0.45 is a gamma of 0.45 (or 1/0.45 = 2.22).

A log curve is something different and based on the log function, below is a graph with Rec709 (using a gamma of 2.22 or 0.45) and Panasonic's V-Log as an example.

By the way Rec709 is not a complete gamma curve as it is linear at the bottom.

On 7/31/2017 at 0:46 PM, Thomas Singh said:

When working with low light footage and sensitive camera chips I sometimes exerience some really bad macro blocking and other artifacts. What are you prefered methods to deal with that?

First I would check if the black point is not raised and that the shadows do not have too much contrast, I think both of these will "invite" the encoder to use a higher compression.  Sometimes it helps to add large grain noise in the blacks. 

24 minutes ago, Steve Shaw said:

Well, that's wrong...

And why is there an 'S' profile to the curves? sRGB and Rec709 are pure power law. (but as I don't know how the graphs were created I'm probably missing something).

Steve

Here is a straight line ramp (Gamma 2.4), the resulting difference is equivalent:

I would avoid sRGB and instead use Rec.709 Gamma 2.2

9 hours ago, Steve Shaw said:

Just so you know the ONLY difference between Rec709 and sRGB is Rec709 is (nominally) gamma 2.4, and sRGB Gammas 2.2.

There is no other difference.

Steve

In Resolve the transforms are different:

DNxHx should work fine but I would make sure you test if the levels and gamma are correctly interpreted at all stages in your workflow.

12 hours ago, Virgil Edward said:

what do you mean by burn in? sorry I'm very new to this aces workflow.

I just want to utilize the color space of ACES inside resolve then output it as non-ACES format like TIFF. Is that possible?

You might want to consider using EXR output.  But the cc version of ACES might not be your best option.

It would take only one constant (DISPGAMMA) to change to make a Rec.709 2.2. ODT, but as far as I know there is no way to add custom ctl files to Resolve.

How do you do the transform in ACES?  A transform would have to take an ACEScc(t) gamma and make a change that will in effect become equivalent to Rec.709 2.2 once transformed with Rec.709 2.4.