There are questions about whether the clamp nodes are required.
The white-balance computation has also been fixed, the main issue is that the multipliers weren't applied to the samples before computing the IDT, resulting in a correction "doubling".
I also changed the way the cache is working which should fix some issues if the archive passed has the same name.
This PR improves CLF generation consistency and addresses #75.
Output for the P2013-001 Generator:
Output for the Prosumer Camera generator, note how the
Range
node is removed here:There are questions about whether the clamp nodes are required.
The white-balance computation has also been fixed, the main issue is that the multipliers weren't applied to the samples before computing the IDT, resulting in a correction "doubling".
I also changed the way the cache is working which should fix some issues if the archive passed has the same name.