A way to increase bit depth in Photoshop?

[Pierrard]

Hello,

I've been using image stacking for long exposures, where I average together several images to get a long exposure effect without high-stop ND filters (arriving soon). This gives a much stronger signal-to-noise ratio, and mitigates a few other technical challenges, but that's beside the point.

What I'm wondering is that since I'm using multiple images (and hence more information on each pixel), is there a way to convince Photoshop of this and treat the averaged stack (a smart object) as a 32-bit layer?

If my understanding is correct, this would give greater (colour/luminance) bit depth, and allow me do the adjustments I already do with more confidence of staying within the limits of the file (i.e. not over-processing).

Cheers!

 

[Wayne Fox]

I did a little video on this (skip to 2:22 to get past all the stuff meant for my store customers who need some background instruction on what HDR is).

Lots of videos out there like this, but this one from Julianne Kost from Adobe is certainly worth watching as well.

I’m pretty sure that once you create the 32 bit file with Photoshop and save it, you can then open it in Adobe Camera Raw to tone map similarly to doing it in Lightroom.

 

[Pierrard]

Sorry, I guess I didn't explain myself properly - I'm taking multiple exposures with the same settings and merging those together, not trying to make HDR images.

I expect that since I'm using multiple images just like in HDR, I could still combine them to make a 32-bit file and get smoother files with stronger processing, but I don't know how.

Any idea how?

 

[Wayne Fox]

What are you hoping to gain by combining them? Seems the only reason to use a 32bit file instead of a 16 bit file is to handle extreme dynamic range ...

Maybe you could use this method to “HDR” the files to create 32bit file and see what you get. Guessing you would get a 32 bit file with all the data in the middle ... maybe that would be useful in pulling the data apart to preserve some information.

 

[Pierrard]

Well, I'm using a Nikon D90 which does not have amazing bit depth compared to other more modern cameras (especially those with larger pixel pitch). What I'm trying to do is end up with cleaner (smoother colour/tonal gradations) output files rather than having artifacts such as banding. Averaging together multiple files does this in part, and quite well (especially averaging areas of subtle colour/tonal gradation), but I was wondering if it is technologically possible in photoshop to produce 32-bit files from those multiple images, since it can be done via HDR.

 

[DigitalSteve]

I think you go to: the Image Tab, choose mode, then 32 bit, for each image and then do the stacking?

 

[Pierrard]

Thanks, that makes sense.

Unfortunately it doesn't seem to smooth the gradients much/any more than normal averaging has. It does help with microcontrast though.

 

[Wayne Fox]

I don’t think you can “manufacture” bit depth by just increasing the file bit depth, and I’m not sure stacking in the manner you describe changes that either. If you average 10 files, you still end up with a single RGB value which is averaged of those 10.

If your concern is smooth gradients, I assume you are coming from raw to 16bit ProPhotoRGB ... adobeRGB 8bit certainly can have issues with smooth gradients. And sometimes gradients that appear not smooth are because of your display limitations, and do not show up in print.

 

[Pierrard]

Thank you for your explanation - that makes things a lot clearer.

From what I've seen so far in prints (of which I have done unfortunately few) I haven't seen the issues I have on screen, so it might be a display issue as you've said.

 

[Bob]

Information can be destroyed but not created.
increased bit depth implies that there is actual information in those low order bits.
Cameras are limited in providing that information by their signal to noise ratios. A camera with a 10 stop dr might be best thought of as a camera that has no meaningful information in any bits other than the top 10.
HDR techniques create a file that "fools the eye" in shifting tones from one region to another often with what some, including myself, consider awful halos.
What can be done is that an image can be shot at an exposure that protects the highlights followed by a shot that overexposes by three stops. Combining those shots via some sort of masking distorts the actual luminance of the scene for the purpose of mapping that into some lesser range that can be printed or displayed on a monitor.
Nonetheless, that mapping does not create more information and does not justify any more bit depth than what was contained in the original files.
-bob