Deprecated: preg_replace(): The /e modifier is deprecated, use preg_replace_callback instead in /www/htdocs/w0068b6d/smf/Sources/Load.php(225) : runtime-created function on line 3
strange colors
forum@fdrtools.com
December 17, 2018, 03:08:20 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
News: SMF - Just Installed!
 
   Home   Help Search Login Register  
Pages: [1]
  Print  
Author Topic: strange colors  (Read 22010 times)
zero
New
*
Posts: 2


View Profile
« on: July 17, 2008, 12:50:41 AM »

Hi,
I downloaded a trial version of FDRTools, input three images from a sample sequence found here http://www.mpi-inf.mpg.de/resources/hdr/calibration/pfs.html, but I'm getting strange color artefacts? Is there something specific I should do to avoid this? Actually I used default settings. Sorry, but I'm new to this thing.


* Untitled.ldr.jpg (58.25 KB, 1024x768 - viewed 854 times.)
Logged
AndreasSchoemann
Administrator
Resident
*****
Posts: 551


View Profile Email
« Reply #1 on: July 18, 2008, 09:06:30 PM »

Hi,
I downloaded a trial version of FDRTools, input three images from a sample sequence found here http://www.mpi-inf.mpg.de/resources/hdr/calibration/pfs.html, but I'm getting strange color artefacts? Is there something specific I should do to avoid this? Actually I used default settings. Sorry, but I'm new to this thing.

The exposure series you mention has been constructed to show that pfscalibration works as expected. You can see this when you look at the exposure differences which vary wildly and arbitrarily from about 0.25 EV (img08, img07) to about 4 EV (img09, img08). 4 EV means a serious gap between two subsequent images, especially for nonlinear JPEG images. Blending this seamlessly without linearizing the images is not trivial.

By default FDRTools does JPEG linearization using the embedded ICC profile or - if this is not available - the profile configured at Preferences -> Devices -> Input Device : Color Space. This method works always and without calibrating the device. The drawback is that it is not very accurate. However, when you keep the exposure differences small the method is sufficient.

Hence the recommendation: when you shoot an exposure series with JPEG images, use a constant interval of 1 EV or 2 EV. Then you won't get into such trouble. Or even better: use RAW images. They are linear by nature and will always blend seamlessly, even with larger exposure differences, say 4 EV.

However, it is possible to merge the images so that the resulting HDR looks fine. You have to use the latest version v2.2beta2, HDR creation module Average with Preset weighting curve "Triangle 2". I have played a bit and found that the current weighting presets are not ideal (this is especially true for "Triangle 1"). With slightly changed values for "Triangle 2" the blending works even better, see the attached screenshot.

I have also attaached the resulting image, tone mapped with Compressor.

I have also found a few other problems which will be solved with the next version 2.2beta3. This will improve the blending result a bit.


* average.png (19.57 KB, 424x820 - viewed 831 times.)

* sample.jpg (46 KB, 600x450 - viewed 897 times.)
Logged
zero
New
*
Posts: 2


View Profile
« Reply #2 on: July 19, 2008, 01:22:17 AM »

ok. thanks.
are the triangles shown on the attached screenshot (over the histograms) what you are referring to?
Logged
AndreasSchoemann
Administrator
Resident
*****
Posts: 551


View Profile Email
« Reply #3 on: July 19, 2008, 08:36:53 AM »

ok. thanks.
are the triangles shown on the attached screenshot (over the histograms) what you are referring to?

Yes. You can preset weighting curves (here "Dreieck 2" in the screenshot, "Triangle 2" in english) and you can then adapt these curves to your taste by adding, modifying and deleting curve points. You can alleviate this process by selecting several layers and then modify the curve. The curves of the selected layers are then modified synchronously.
Logged
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!