Easy method to distinguish between 16bit and 32bit images?

Quick question: is there a quick method to tell if an imported image is truly 32bit, rather than converted to 16bit? Blender 2.6 no longer automatically converts a 32bit tiff to 16bit depth, does it?

16bit gradients are pretty obvious. You can right click the image and then properties. It should say. Blender posts and saves in the output you select.

That is not what I meant to ask: the tiff save option in Blender only offers 8 and 16bit per channel. I am working with a 32bit tiff in Photoline, import that in the compositor. I do the same for a 16bit tiff (change bit depth to 16 bit of the original file first). Both files look different in Blender: one is darker than the other. So, I gather that Blender does keep the bit depth of each file. (?) Or not? In the properties, both show up as float. So does Blender keep the 32bit information, or does it downgrade to 16bit automatically when a 32bit tiff is imported?

However, there is no 32bit output option to save changes to a full 32bit depth tiff version. I can use Radiance HDR or OpenEXR, but both are unsupported in Photoline. So, export becomes a problem as well.

AFAIK, Blender does respect the bit depth but only in the compositor. Unfortunately, this doesn’t work with the sequencer (yet).

Regarding which is the higher depth pic, I’ve got an easy solution but no reasoning behind it: it’s the darker pic that’s always gonna be the 16bit one. I’ve confirmed this through empirical tests using other apps and I know this for a fact to be true. The thing is I don’t really know why. It’s as if creating a higher depth pic is messing up the gamma which doesn’t really make any sense. Should anyone have a clue, please share! :slight_smile:

If it displays darker does the brighter part clip? If so perhaps you are only seeing the toe of the image?

I don’t really think that there’s any clipping involved. Unless clipping of the highlights starts at 60% IRE or something which I would find weird. It appears that highlights are (almost) gone, leaving mostly midtones and blacks. That’s why I said that it reminded me of a gamma transform of some sort.

You could try it yourself. Render something out from the viewport/compositor and save it as 8bit tiff and 16bit tiff. Then import each file in the compositor and have a look at it. Doesn’t the 16bit pic look darker? The difference should also be visible in most image viewers (Gwenview, Okular) and editors (Gimp 2.9, Krita).