`.postprocess()` always returns a black image about rawpy HOT 8 CLOSED

New version is out, let me know if you still have troubles

from rawpy.

How weird! Which OS and Python version do you run?

from rawpy.

macOS 10.12.3
Python 2.7.13
rawpy 0.8.0
(if it's important) numpy 1.12.0

from rawpy.

One more question: are you sure that the image is not just very dark? The maximum value you get via np.max(rgb) is 6381, so it can't be all zeroes. The combination of 1,1 gamma without auto brightening may lead to dark images if the RAW image doesn't have much signal. What happens if you just do a postprocess() without any arguments?

from rawpy.

Output is similar with no arguments passed to postprocess(). The max changes to 255. The image is still black. Opening the postprocessed image in, say, Photoshop, and cranking up the brightness / contrast shows no sign of any image or vignetting.

raw_image is [[243 260 241 ..., 253 243 253]
 [264 191 258 ..., 189 256 189]
 [243 260 245 ..., 251 241 251]
 ..., 
 [243 181 239 ..., 179 229 179]
 [223 241 225 ..., 231 225 231]
 [241 183 247 ..., 177 231 177]]
rgb max is 255
rgb is [[[0 0 0]
  [0 0 0]
  [0 0 0]
  ..., 
  [0 0 0]
  [0 0 0]
  [0 0 0]]

 [[0 0 0]
  [0 0 0]
  [0 0 0]
  ..., 
  [0 0 0]
  [0 0 0]
  [0 0 0]]

 [[0 0 0]
  [0 0 0]
  [0 0 0]
  ..., 
  [0 0 0]
  [0 0 0]
  [0 0 0]]

 ..., 
 [[0 0 0]
  [0 0 0]
  [0 0 0]
  ..., 
  [0 0 0]
  [0 0 0]
  [0 0 0]]

 [[0 0 0]
  [0 0 0]
  [0 0 0]
  ..., 
  [0 0 0]
  [0 0 0]
  [0 0 0]]

 [[0 0 0]
  [0 0 0]
  [0 0 0]
  ..., 
  [0 0 0]
  [0 0 0]
  [0 0 0]]]

Here's what Preview.app shows for the image:

https://drive.google.com/file/d/0B_-3PXrqwAXBc0NkekZEdk9MYk0/view?usp=sharing

(GitHub, Imgur, imgbb, countless other image hosting services are down due to S3's outage, sorry for the Google link)

from rawpy.

Would you be able to provide the .nef file? You can send an email to me (see profile) if you like.

from rawpy.

No problem: https://drive.google.com/open?id=0B_-3PXrqwAXBUGE4eTlCYkJRUEU

from rawpy.

I found the issue. Due to a build script mistake the macOS package was built using an old LibRaw version (0.16, see rawpy.libraw_version, whereas it should be 0.18), and that apparently can't read your NEF file. I tried it on Windows with libraw 0.18 and it works. I'll do some tests and release a new version later today. Then all should be fine. Thanks for reporting, nasty bug you uncovered!

EDIT: Travis CI is affected from the Amazon AWS outage so I can't do proper testing currently (don't own a Mac), will probably take another day for the release.

from rawpy.