On Sep 29, 10:34 am, kfj <[email protected]> wrote: > On 28 Sep., 19:21, JPW <[email protected]> wrote: > > > Easy, Kay. > > Please assume the best of me, as I do of you. And remember, as I have > > told you, I am a totally naive user of any but the most basic feature > > of Hugin, PanoTools and command line programs in general. > > Ooops... my fault. I did mean you should set exposure sigma to 50, not > exposure mu. I gave you a mixed message:
OK. Appreciate the correction. I had deduced this was your intent after your response when I reread your original post and the manual. > > On Sep 28, 6:59 am, kfj <[email protected]> wrote: > > > ... > > > pixels from all images and just use the fusing mechanism. But a large > > > value for exposure-sigma should do the trick as well. so for your > > > enfuse options, try something like > > --saturation-weight=0 --exposure-mu=50 <<<< typo. arrgh. sorry > > Let me explain: enfuse looks at all pixels in all images in turn. For > every pixel, it calculates it's 'quality' by a certain criterion. When > taking well-exposedness as criterion, how can well-exposedness be > defined? Enfuse's approach is to look how 'near' the pixel is to a > 'good' exposure. --exposure-mu defines 'good' and --exposure-sigma > defines 'near'. The default value of exposure-mu of 0.5 means that per > default pixels which are precisely between black (0) and white(1) are > considered 'good'. The 'near' parameter is more difficult to grasp - > it's the standard deviation from the mean defined with exposure mu, > hence the sigma. The default of 0.25 here has the effect that roughly > the middle half of brightness values will be considered better than > the rest, and the closer to mu the better - it's a gaussian bell > curve. If you want to fool enfuse into considering all pixels equally > good, you can take a large value for sigma and get a bell curve which > is so flat it's almost plane. That's where the sigma of 50 comes in - > a value you'd never use for 'real' exposure fusion but which should > work for your specific case. > > After all images in all pixels have been given a quality measure, > enfuse combines them, weighting each according to it's quality. If > they all have the same 'quality' enfuse will simply produce an average > of all inputs, sice it considers them all equally good. > > Kay Although complex, I actually do understand the concepts (the graphs in the manual are helpful.) My real problem is knowing exactly what to use as commands and how and where to enter them. This is where I have difficulty using CLIs — I understand their incredible power and simplicity makes them desirable, but I'm a terrible speller with dyslexia, and my brain just doesn't work well with text based abstraction. Anyway, I deduced from your posts that I should put --saturation-weight=0 --exposure-sigma=50 in the enfuse options box and check "Panorama Outputs" Exposure fused from stacks" Is that correct? If it is correct, I did run some experiments with various sigma at . 25, 5, 25, & 50 so I could observe the effect. Assuming they ran correctly— again, where can I review a log to see exactly how they ran? They all looked equally good to me In terms of noise reduction and far superior to any of the jpegs or tiff. Should there be a great deal of difference between them? I looked particularly in the blacks and in the glossy areas. The light glossy areas in the originals had the most obvious sensor noise. The jpeg crud was distributed everywhere pretty liberally. This was all disappeared with the default settings. Change in sigma seemed to have little effect. Again, perhaps this is because of the nature of my test images. Perhaps they were so noisy even a small sigma repaired the majority of the noise. Or perhaps a high sigma might produce a more obvious effect using only 2 or 3 input images (I used 5.) Thanks again. If you would like to see the test images, let me know. John -- You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/hugin-ptx
