User Manual

Table Of Contents
Supported LUTs
DaVinci Resolve uses both 1D and 3D LUTs, and supports LUTs in a variety of formats.
.cube: DaVinci Resolve uses both 1D and 3D LUTs in the .cube format. 3D LUTs are
exported as 33x33x33 cubes with 32-bit floating point processing. 65x65x65 cubes
created outside of DaVinci Resolve are also compatible but may cause issues with
performance. DaVinci Resolve can also read and use Shaper LUTs in the .cube format,
but these must also be created outside of DaVinci Resolve. The .cube format used by
DaVinci Resolve is no relation to the Iridas/Adobe SpeedGrade .cube format.
Panasonic VLUT format: A LUT format associated with Panasonic VariCam cameras.
CLF (common LUT format): CLFs use an XML format that is capable of encompassing
a limited number of mathematical transforms in addition to traditional lookup-tables to
do image processing. Promoted by the academy as the ideal LUT format for use with
ACES, LMTs for ACES are recommended to be in the CLF format due to its increased
precision and flexibility.
DCTL: DCTL files are actually color transformation scripts that DaVinci Resolve sees
and applies just like any other LUT. Unlike other LUTs, which are 1D or 3D lookup
tables of values that approximate image transformations using interpolation, DCTL
files are actually comprised of computer code that directly transforms images using
combinations of math functions that you devise. Additionally, DCTL files run natively on
the GPU of your workstation, so they can be fast. For more information on DCTL, see
Chapter 174, “Creating DCTL LUTs.
What’s the Difference Between a LUT and a Shaper LUT?
DaVinci Resolve is capable of importing and using LUTs within its 32-bit floating point
image processing pipeline. The .cube format can be used as either a simple 33x33x33
3D LUT, or as a shaper LUT, which is actually a method of using two LUTs, a 1D LUT and
a 3D LUT together, that addresses signal processing issues that 3D LUTs alone
can’t handle.
For processor efficiency, 3D LUTs are designed with reasonable lower and upper limits
for the data they will handle. It’s well known that when a 3D LUT is fed values that are
outside of the range that LUT is designed to handle, the out-of-range data will be
clipped. Since many LUTs are designed with digital cinema workflows in mind, the
practical result is that feeding a video signal with super-white in it to a 3D LUT that’s
designed for full-range data (0–1) will clip the super-white part of the signal.
Shaper LUTs handle this issue by first using a 1D LUT to process video signals with
out-of-range data, fitting the signal into a range that the 3D LUT won’t clamp. The
output of the 3D LUT includes the reverse transformation, to effectively zero out the 1D
LUTs transform, while retaining whatever processing the 3D LUT was meant to apply.
Shaper LUTs are also useful for dealing with extremely large data sets, such as
OpenEXR files that can theoretically handle an image data range of –infinity to +infinity.
Using a Shaper LUT, you can remap the incoming data to fit more precision in the 0–1
range, leaving less important data outside the range.
Chapter – 130 Using LUTs 2924