Project

General

Profile

Bug #10512

nef image != exported image

Added by Pascal Obry over 2 years ago. Updated over 2 years ago.

Status:
Confirmed
Priority:
Medium
Assignee:
-
Category:
-
Target version:
-
Start date:
06/04/2015
Due date:
% Done:

10%

Affected Version:
git development version
System:
Debian
bitness:
64-bit
hardware architecture:
amd64/x86

Description

Something strange (to me) and I'm not sure this is correct or not. Take the attached NEF (old one one found on the net, just 3Mb, but I have the same result with a NEF from my D800) Convert it to TIF (using dt, 16bpp). Open both images and add +1.0EV to both. The exposure applied to the TIFF is far more pronounced than the one on the NEF.

Is that expected?

base_curve.PNG (9.09 KB) Stefan Lober, 10/05/2015 08:26 AM

History

#1 Updated by David Houlder over 2 years ago

Different basecurves? I suspect the exposure adjustment may be done before the basecurve is applied.

#2 Updated by Pascal Obry over 2 years ago

Maybe, but when I open both images they are with the same exposure and images looks really alike. So I suppose that changing exposure to both at the same level should give the same output. But I doubt that exposure is done before basecurve anyway.

#3 Updated by Roman Lebedev over 2 years ago

If you export to 32bit float tiff and try to change exposure on that image, does it look "like nef exposure" or "like 16bpp tiff exposure"?

#4 Updated by Pascal Obry over 2 years ago

Using a 32bit float tiff is even worst, that is the exposure is even far more pronounced than for a standard tiff. Really something wrong I think with exposure on tiff.

#5 Updated by Roman Lebedev over 2 years ago

  • Subject changed from nef exposure != tif exposure to nef image != exported image
  • % Done changed from 0 to 10
  • Priority changed from Low to Medium
  • Status changed from New to Confirmed

Not limited to TIFF, something is indeed wrong...

#6 Updated by Stefan Lober over 2 years ago

Hello,
I had a look at the issue.
I tested with this image: http://www.rawsamples.ch/raws/nikon/d1/RAW_NIKON_D1.NEF

The base curves are indeed different:

NEF base curve

vs standard linear for the exported image (also module is off per default)

When the NEF base curve is applied to the exported image, the effect is even more pronounced. Also, applying the NEF base curve to the exported image without exposure module already gives a different result.


The NEF has the following modules applied (with activated exposure):
module: raw black/white point
module: white balance
module: highlight reconstruction
module: exposure
exposure: 0,995000; black: 0,000000; white: 0,501736; scale: 1,993080
module: base curve
module: input color profile
module: sharpen
module: output color profile
module: gamma

The exported image has:
module: white balance
module: exposure
exposure: 0,995000; black: 0,000000; white: 0,501736; scale: 1,993080
module: input color profile
module: output color profile
module: gamma

If applying the exposure module results in different outputs whereas before they were the same, one of the modules (especially after exposure) has to mess with the image.

I did a test, setting all pixel values in exposure to RGB (0.5, 0.5, 0.5). This results in RGB byte values of the outputs (from color picker):
NEF (231, 231, 231) | exported (127, 128, 127)
without base curve: NEF (188, 188, 188)

So, the exported image is left as it is after exposure and the NEF is altered. Should be the source of the difference. Interesting that this has no effect without exposure module applied.

Also available in: Atom PDF