Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Lin2Log: fix inconsistency between log2lin and lin2log #79

Merged
merged 2 commits into from
Feb 13, 2021
Merged

Lin2Log: fix inconsistency between log2lin and lin2log #79

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c89ac63
Log2Lin: fix inconsistency between log2lin and lin2log
devernay Sep 8, 2020
55d4487
Log2Lin: fix the formulas to be consistent with Sony's code
devernay Feb 10, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
51 changes: 37 additions & 14 deletions Log2Lin/Log2Lin.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,14 +82,48 @@ OFXS_NAMESPACE_ANONYMOUS_ENTER
#endif

/*
Following the formula:
# The Log2Lin/Lin3Log formulas: which one is right?

## From the Sony source code: good formula!

The reference Sony code [2] as well as the Colour Python package [3] and the ACES
code [4] use the following formula, which also seems to be used by Nuke's plugins:

cineonBlackOffset = 10.0 ** ((95.0 - 685.0)/300.0)

def fromCineon(x):
return (10.0**((1023.0 * x - 685.0) / 300.0) - cineonBlackOffset) / (1.0 - cineonBlackOffset)

equivalent to:
linear = gain * (pow(10,(1023*v - whitepoint)*0.002/gamma) - offset)
cineon = (log10(v/gain + offset)/ (0.002 / gamma) + whitepoint)/1023

## From the Cineon doc: most probably a wrong Formula

The original Cineon doc [1] gives the following formulas:

offset = pow(10,(whitepoint - blackpoint) * 0.002 / gamma)
gain = 1/(1-offset)
linear = gain * pow(10,(1023*v - whitepoint)*0.002/gamma)

Log2Lin [1, Sec. 5.3 and 6.3]:
linear = gain * pow(10,(1023*v - whitepoint)*0.002/gamma) - offset

Lin2Log [1, Sec. 7.0]:
cineon = (log10((v + offset) /gain)/ (0.002 / gamma) + whitepoint)/1023

Here we're using: blackpoint = 95.0
whitepoint = 685.0
gammasensito = 0.6

Ref:
[1] Glenn Kennel (1995-07-26). "Conversion of 10-bit Log Film data to 8-bit Linear or Video
Data for The Cineon Digital Film System Version 2.1" http://www.dotcsw.com/doc/cineon1.pdf
[2] Sony Imageworks. (2012). make.py. Retrieved Feb 9, 2021, from
https://github.com/imageworks/OpenColorIO-Configs/blob/master/nuke-default/make.py#L163
[3] NumFocus. Colour Python package. Retrieved Feb 9, 2021, from
https://github.com/colour-science/colour/blob/master/colour/models/rgb/transfer_functions/cineon.py
[4] Sony Imageworks. red.py. Retrieved Feb 9, 2021, from
https://github.com/imageworks/OpenColorIO-Configs/blob/master/aces_1.0.3/python/aces_ocio/colorspaces/red.py#L79
*/

#define kParamOperation "operation"
Expand Down Expand Up @@ -138,17 +172,6 @@ class Log2LinProcessorBase
double _whitepoint[3];
double _gamma[3];

/*
Following the formula:
offset = pow(10,(blackpoint - whitepoint) * 0.002 / gamma)
gain = 1/(1-offset)
linear = gain * (pow(10,(1023*v - whitepoint)*0.002/gamma) - offset)
cineon = (log10((v + offset) /gain)/ (0.002 / gamma) + whitepoint)/1023
Here we're using: blackpoint = 95.0
whitepoint = 685.0
gammasensito = 0.6
*/

double log2lin(double xLog,
int c)
{
Expand All @@ -160,7 +183,7 @@ class Log2LinProcessorBase
double lin2log(double xLin,
int c)
{
double retval = (std::log10( (xLin + _offset[c]) / _gain[c] ) / (0.002 / _gamma[c]) + _whitepoint[c]) / 1023.;
double retval = (std::log10(xLin / _gain[c] + _offset[c]) / (0.002 / _gamma[c]) + _whitepoint[c]) / 1023.;

return retval;
}
Expand Down