This stems from a post I had several months ago… Maybe others have already solved this, but I had not.
About a 6 months ago we started rendering to a gamma of 1.8 more often than not. We were previously working more or less in linear color space with a gamma of 2.2.
When we switched to 1.8, we were saving our images as 16bit TIFF files instead of 16bit EXR files. As you know EXR’s are vastly superior in terms of file size compared to TIFF’s coming out of 3dsMax. Often it can be 40% of the size of the TIF, which is huge in terms of data storage, and the amount of network traffic when retrieving those frames the server.
It is a given that EXR’s are floating point color that are typically written out at 1.0, and then are displayed in Photoshop correctly with a gamma of 2.2. When working with at a gamma of 1.8 in Max, and saving at a gamma of 1.0, the image becomes overcorrected and washed out when opened in Photoshop. This was very predictable.
Which is the reason we went back to TIF’s, work at 1.8, output set to 1.8, perfect color match between Photoshop and Max. So maybe, you know where I am going with this… It just dawned on me today that I could set a ratio up between the output gamma in Max, and the correct display of floating point color in Photoshop.
What this means that if I set the output gamma in Max’s preferences to be 0.818, and then save as an EXR, my image displays perfectly when opened in Photoshop. This works because the ratio of correction is the same.
So, …a couple of tests on which compression is best when working with EXR’s.
EXR w ZIP Compression stored as scanline is the clear winner on this test. I don’t visually notice any difference between the compression types, and I don’t know the ins and outs of each compression type, or why there are so many. Maybe someone else can chime in?
The EXR saved in this format for my test case is 2.3 times as efficient with disk storage as the TIF file is.
As far as I know only the last 3 in the list were lossy compressors. This leaves me to believe that the others are lossless compressors.
1) I don’t use tiled EXR’s but I do know they have a benefit in some instances, but I think it is geared more to pulling high res textures for efficient use.
2) I don’t store render channels in my EXR’s. If you are doing this, you may find different results.
3) I use the Max frame buffer, and not the Vray frame buffer because we write out several Render Elements when putting together an image.