Hmm, I just noticed there is a branch from GIMP for a film-dedicated image editor, 'Cinepaint'. I just checked the (painfully outdated) Mac version, and this one does indeed allow to input straight 16Bit color values and do basic image editing/painting photoshop-style.
I couldn't test much more, since the save dialog is broken (sigh...) in Mac version 0.24 under Snow Leopard, but the Linux version is a lot more recent it seems.
- Carsten
XYZ 12bit test chart
-
- Posts: 2804
- Joined: Tue Apr 15, 2014 9:11 pm
- Location: Germany
Re: XYZ 12bit test chart
A few days ago I dived into this again. For some reason, I downloaded GraphicConverter (legendary Mac OS/OS X app). Playing with it, I noticed that, besides 16Bit image support, it offers a float colorpicker, 0-1 with 3 decimal places (0,000 to 1,000), so, roughly 10Bit resolution. I contacted the german programmer wether he would see any difficulty in upping this to direct 16Bit input or more decimals in the float version. He wrote back that he would think about it. Today, he sent me a color picker that allows me to dial in color palettes with 0-65535 definition. So I can create precise 16Bit RGB TIFFs, that, when I choose 'none' in DOMs color conversion dialog, should be transformed 1:1 to XYZ values in the resulting J2C/DCP. I will create some test patches now and see what DOMs vectorscope says about them. GraphicConverter also offers a 16Bit color pipette which allows you to pick colors from certain parts of the image in 16Bit definition. Needless to say, GraphicConverter will also directly display J2C images. So, this comes out as a nice test chart and color conversion analysis tool.
Of course, these 16Bit values will be scaled down to 12Bit values in DOM, but that is easy to calculate. I wasn't keen enough to ask for direct 12Bit input, which would probably be easy enough to implement. Note that there is no formal 12Bit TIFF, so, all 12Bit XYZ TIFFs are in fact 16Bit.
Edit: Looks good. When I open the 16Bit TIFF I created in GC, DOM shows me the 16Bit color components in the vectorscope scaled down to 12Bits, but otherwise exactly how I dialed them in.
Now, which are the legal boundaries for XYZ-P3 conversion? I can dial in all possible XYZ values, many of them are probably illegal values converted to P3 in the projector? I will see what our projector thinks of it...
- Carsten
Of course, these 16Bit values will be scaled down to 12Bit values in DOM, but that is easy to calculate. I wasn't keen enough to ask for direct 12Bit input, which would probably be easy enough to implement. Note that there is no formal 12Bit TIFF, so, all 12Bit XYZ TIFFs are in fact 16Bit.
Edit: Looks good. When I open the 16Bit TIFF I created in GC, DOM shows me the 16Bit color components in the vectorscope scaled down to 12Bits, but otherwise exactly how I dialed them in.
Now, which are the legal boundaries for XYZ-P3 conversion? I can dial in all possible XYZ values, many of them are probably illegal values converted to P3 in the projector? I will see what our projector thinks of it...
- Carsten