RAW work flow

Hi All,

I’ve read a number of threads related to RAW file capture and processing but still not sure if my process flow is optimized or even close to it.

Here is the simple flow I follow:

  • Shoot in RAW (as per many forum recommendations). Nikon D810, 35mm prime lens
  • In CameraRaw, adjust WB (daylight), no other adjustments (sharpening, contrast, etc.)
  • Export a 16-bit TIFF (not sure if this is necessary over 8-bit)
  • Import to RC and go from there

Does this sound right? 

Anything missing or extra?

Thanks

Hi Tim,

I wouldn’t say there is a right or wrong, just a better or worse, depending on your needs and your source images.

Is 16 bit TIFF supported now? I don’t remember since I only use high quality JPGs developed from RAW. This is ok for many cases, I would thinkt that the difference is rather small. But if you want the last possible bit of information, then your approach might be worth the additional time and storage space. If you want to avoid the detour via TIFF, you might want to consider using RAWs directly, which is possible when you have a driver installed that enables windows to display your RAWs natively.

So far I’ve always used camera WB, since for whatever reason my developer does not seem to like my RAWs since they get all magenta with automatic WB…  :-)  My point is that ok is what meets your visual needs. I can’t imagine any problems arising from changing WB other than the colours of the output, because the “features” within the images will stay the same.

Here is a post (no. 3) https://support.capturingreality.com/hc/en-us/community/posts/115002690471-Starting-and-continuing-a-project- where Luxcubed says that he alters the images extensively (sharpening, de-noising, the whole bit) which is usually not recommended. But he says his results are better with the alterations as oppose to almost unchanged ones. So you really need to try and see for yourself…

Hi Gotz,

Thanks for the advice! (as usual :slight_smile: )

The post you linked was one of the many I’ve read on the topic and, I agree, the results Luxcubed is getting are surprising. I’ll try messing with the adjustments and see where that takes me.

The program took the 16-bit TIFFs without complaining, so import at least is supported, but I’m not sure if it actually uses full depth info. 

I’ve seen some references to delighting RAW images for better photogrammetry results. 

Have you tried any “delighting” techniques (removing shadows, etc.)? 

 

I think the only thing that’s currently supported for greater than 8-bit. is exporting undistorted photos.

I think everything else internal is done in 8-bit. and no support for 16-bit textures.

so not much point now.

but hopefully soon we will get 16-bit and float support through out.

 

 

 

 

Hi Tim,

no I haven’t messed with the shadows much since in my experience RC is already magnitudes better than it’s often mentioned competitor. But I should try it once. When I started out and again when I was beginning to use RAWs, I also looked for the ultimate workflow. But now I am more and more convinced that those few recommendations out there are only targeted at beginners and only describe a way that will work best in most cases. As usual though, reality is much more diverse and there are undeterminable many different objects and needs out there, so that in the end everybody has to work out their own best suitable workflow. I am certain that delighting or basically every alteration like smoothing or de-noising (which is similar anyway) will potentially lead to a less accurate model on the pixel level, but can have benefits that outweigh it, like giving a rough geometry in otherwise too dark areas.

Chris, thanks for the input. I wonder what the point is of exporting 16 bit with only 8 bit input? It would only make sense if RC operates with 16 bit internally, which I doubt since the input is limited. Hmm…

Thanks Chris and Gotz!

Looks like I’m wasting space with 16-bit TIFFs :slight_smile:

 

it was done as a quick update. so you can do texturing in other software like mari. where you can use higher bit-depth images. which opens up the ability to delight the models. but this also brings in all other issues of transferring distortion as well. so its not ideal. 

really much easier just to shoot on an overcast day that if you can.

looking forward to having it work all the way through… preferably with float images as well.

" really much easier just to shoot on an overcast day that if you can. "

 

Indeed. One order of overcast day for me please! :slight_smile:

Hahaha!
Can you order 10 for me please? No rain though, or snow, or or strong wind. Is there an option for no people as well?