Nan with Spot WB

photoflow-20170315 (Windows)

When attempting to Spot WB the rim of the bowl on the left side, I get a black preview and:

When changing from CAMERA to Spot mode, the area starts clipping in the shadows. That is probably why we get this problem. However, when I alter the multipliers to bring the area out of clipping, Spot WB works as expected.

Indeed, something is not normal… thanks for reporting, I’ll look into that tomorrow, and let you know.

@afre The SpotWB issue should be fixed in git stable branch. Do you need an updated package for testing, or are you using those from your Linux package manager?

Thanks!

1 Like

Thanks! :slight_smile: I have been using photoflow-20170315.zip (Windows), which is kind of old.

Also, I was wondering if you could add the ability for (Color) Spot WB to assess a region; e.g., by allowing the user to specify a radius or draw a rectangle.

Looks like it is time to make a new windows build!!!

Yes, this is in my plans, but I had no time yet to code such a feature…

1 Like

New Windows version ready for testing: photoflow-20170521.zip

I found some time to try it today. When I attempt to use Spot WB, PF becomes unresponsive and starts consuming increasing amounts of memory. Testing it on the CR2 file from “[PlayRAW] Autumn Island”.

Would you be able to start photoflow.exe from a terminal, and send me the terminal output while the program tries to compute the spot WB?

Thanks!

I forgot mention that I don’t know how to do that in Windows. cmd.exe just opens PF and that is it.

This should work:

  • open a terminal with cmd.exe
  • navigate to the directory where the photoflow.exe program is located
  • start photoflow.exe and redirect the output to a file:
    photoflow.exe>c:\log.txt 2>&1
  • try the spot WB, and kill the program after a bit of time (to collect some output…)

More about output redirection under windows here.

I didn’t consider redirection since cmd itself didn’t output anything; i.e.,

C:\Users\afre> photoflow.exe
C:\Users\afre>

Anyway, here is the full output → log.zip (22 Bytes). Upon inspection, this only happens when I set the working color space to LAB.

PS photoflow-20170521.zip compared to photoflow-20170315.zip is a bit sluggish.

Ah! That is definitely the reason! I admit that I have overlooked the Lab case, I will fix it as well…

Edit: see this issue for details.