Color correction in exposure fusion

[hermer-blr]

Color correction in exposure fusion

by hermer-blr » Fri Aug 26, 2011 2:37 pm

I feel that the answer to my question below has already been given in a post, but I cannot find where. I also remember having already tried once exposure fusion in APG and having been disappointed by some color modifications compared to the original TIF images.

a. Inputs:
- 10 tif images x 3 brackets (-2, 0, +2) = 30 images

b. Purpose:
- Process those 30 images with APG 2.5, in "exposure fusion" mode

c. Experiment 1:
- The 30 images are grouped in 3 layers (-2, 0, +2)
- Exposure fusion activated
- Color correction to "none"
- "Exposure fusion" rendering
- Result perfect, except that the main background color (subject = Bryce Canyon, USA) has changed from yellowish to reddish (by comparison with the Nikon capture generated tif images)

d. Experiment 2:
- - The 30 images are grouped in 3 layers (-2, 0, +2)
- Exposure fusion activated
- Color correction to "LDR"
- 3 anchors set on the same location for (-2,0,+2) - adjust all gamma and color tone - optimize by layer
- Preview worse than ugly - I have however launched the "exposure fusion" rendering - Result perfect (however different from the one achieved in experiment 1: e.g. the cloudy sky is more beautiful now); the color change, unfortunately, stays !

e. Solution ???
- What is the right solution in term of color correction (as mentioned in the past, this is not documented as part of Workflow B)?

Note: I made a simple test with Photomatix on one set of images - By applying exposure fusion ("adjust" template with saturation to 0.5), the colors do not change compared to Nikon Capture...

Last edited by hermer-blr on Fri Aug 26, 2011 2:59 pm, edited 1 time in total.

[HansKeesom]

by HansKeesom » Fri Aug 26, 2011 2:49 pm

f. Experiment 3 :
- use enfuseGUI to fuse 10 * 3 tiff images into 10 tiff images.
- Load the 10 resulting images into Autopano and do detection.

[hermer-blr]

by hermer-blr » Fri Aug 26, 2011 5:20 pm

To HansKeesom - Thanks for your help

To any reader of my post: The issue is resolved !!!

The point is that I was comparing an input image viewed with Nikon Capture NX2 and the panorama viewed with my favorite (albeit not perfect) viewer called "FastStone Image Viewer":
- Both software take into account the image ICC profile
- But, in addition, Nikon Capture takes the monitor ICC profile into account !
- while FastStone does not.

When I viewed both the input image and the panorama with Nikon Capture, I could observe the same colors !!!

Conclusion: APG did a perfect job in exposure fusion. It is interesting to note that "none" and "LDR" do not exactly give the same panorama. I tend to prefer the "LDR" one...

[hermer-blr]

by hermer-blr » Fri Aug 26, 2011 7:10 pm

To be complete:

I have also rendered the panorama made of the 10 images obtained with Photomatix from the (-2,0,+2) set of exposures.

After long comparison between the Photomatix Pano and the LDR Pano, I decided that the LDR one was better. The difference (following some contrast and saturation adjustments to make both panos look exactly the same) is very subtle; but, at the end, at 100 % (and eventually 200 %) resolution viewing, I got the impression that the LRD one was a little, just a little, sharper, with perhaps a little, just a little, more details in the colors...

From an other perspective (without any importance, however, in term of image quality), the RMS with 30 images was a little better (1.4 vs. 1.6) and I had nothing to do to make the pano horizontally aligned and well shaped...

Conclusion: APG with exposure fusion is great !

[lumelix]

by lumelix » Fri Aug 26, 2011 7:32 pm

Do you use the auto alignment for the input images in Photomatix ?
It fails sometimes an give a not optimal result for some HDRs

[hermer-blr]

by hermer-blr » Fri Aug 26, 2011 7:39 pm

As my panorama had been shot handheld, I had to use it. The result was excellent.

As mentioned, the difference between the Photomatix Pano and the APG Exposure Fusion one is very thin and I really had to observe the two panos during quite a long time to make a decision...

[HansKeesom]

by HansKeesom » Fri Aug 26, 2011 7:42 pm

Good to hear Autopano is doing a great job.

As I use the free open source software of enfuseGui (http://software.bergmark.com/enfuseGUI/Main.html) I really would like to hear you opinion when you would use it instead of Photomatix.

[gkaefer]

by gkaefer » Fri Aug 26, 2011 11:52 pm

Good to hear Autopano is doing a great job.

As I use the free open source software of enfuseGui (http://software.bergmark.com/enfuseGUI/Main.html) I really would like to hear you opinion when you would use it instead of Photomatix.

well for me photomatix can make better results with danger of more options take more time to find best result.
if shot handheld... in my cases enfusegui mostly failed.
enfusegui has an faster workflow from testing options until batching and having filal images...
Georg

PS enfusegui does not HDR, it does exposue fusing...

Last edited by gkaefer on Fri Aug 26, 2011 11:55 pm, edited 1 time in total.

[HansKeesom]

by HansKeesom » Sat Aug 27, 2011 12:09 am

PS, I know exposure fusing and HDR are different, well explained it is at http://www.hdrsoft.com/support/faq_photomatix.html#differences

I like exposure fusion as it gives me quickly what I want to have, a good realistic picture.
For me the result counts and I wondered how hermer would judge it compared to the other ways