Category Archives: Graphic Editing

Underwater Color Correction

A post by Brianne Bilyeu about her diving adventures at the coast of Belize reminded me that I was never really satisfied with the common underwater color correction methods. The method used by Photoshop is quite good but who can afford Photoshop? I, for one, can not.
The methods offered in the form of ready made scripts for Image Magick and The Gimp were not to my taste, so in the good old tradition “If you don’t like it, change it!” I will describe a process that gives better results (depending on the quality of the source image, YMMV, and so on), for me at least. If you have a different opinion: tell me.
The underlying method for all scripts is quite simple: the water reduces the red part of the spectrum. so just add some red and you are done. It is not that simple, of course, but comes quite close. This method is known as the Mandrake method because David Kusner, who uses the nym mandrake, described it first. It goes as follows:

  1. make a copy of the background layer
  2. desaturate the layer (make it black and white)/ make a new grey (0x808080) layer
  3. you may need to persuade the layer to still be in “color”
  4. make a new layer, fill it with red
  5. combine the red layer and the black and white/grey layer with the method “multiply”
  6. combine the layer resulting from step 5 to the background layer with the method “screen” and adjust opacity to taste; 50% seems to be the suggested value that fits on average.

A tutorial for Gimp can be found at Tanked Up Underwater Imaging. He also made a SCM-script for Gimp. But this process gives still unsatisfying results, at least fo for my taste. The color correction is good but the picture gets quite flat and dull. Yes, one can do something against that flat- and dullness but that involves some work, hence time and I’m a lazybusy guy without much time at hand so I looked for a fast and simple method. The one I liked most was changing the last step from combining with the “screen” method to combining with the “overlay” method and raise the opacity to more than 50%. A bit of sharpening (with FX-Foundry ➜ Convolution Matrix Presets ➜ Sharp. The defaults are ok in most cases ) and it’s done.
To show the effects I shamelessly stole (no, I asked first, of course!) Brianne Bilyeu’s picture of a large turtle. That means the pictures belong to her and are not included in my CC-BY licence! But she has put her work under the Creative Commons by-nc-sa license, so you may use them, too. (Please be aware that in some jurisdictions (e.g.: Germany) having advertisements on the page may already count as commercial use!)

Original picture

Original picture

At first with the unchanged Mandrake method:
Mandrake Method

Raw Mandrake Method

The colors are quite correct, but with an overall dullness. Now with just the change of the combining methods and opacity set to 100% (which is too much but shows the effect better):
Combining method "overlay" instead of mandrake's "screen"

Combining method “overlay” instead of mandrake’s “screen”

And finally sharpen it (click on the picture to see the effect):
Final Picture

Final Picture

The other underwater pictures at Brianne Bilyeu’s page profit from the algorithm,too but to a lesser extent. You just have to accept that you need artificial lights if you want to make color-correct pictures at a depth of more than 1-2 meters to get professional and therefore sellable results.
On the other side…

Counting Fish

Goldfish and other Cyprinidae

Goldfish and other Cyprinidae

I tried to count estimate the number of fish in our pond with some technical help. The cheap diggi-snappers today have no threads at the lenses anymore, so I had to try to adjust a polar filter with one hand, hold the camera with the other hand and throw the fish food with my third one. This seemed to work, call me Zaphod from now on.
But the camera didn’t want to play fair: I had either no correct focus (the polar filter seems to disturb the auto-focus and this thingies have no way to focus manually) or didn’t have the correct exposure—sometimes even both—so I just took some normal pictures and enhanced them digitally to get some rough information about the ichthyofauna of our little pond.
Most pictures were useful but boring with the exception of that one at the top.
Oh, and the fishes found and clearly identified were:
Carassius gibelio forma auratus (the goldfish)
Rhodeus amarus (European bitterling)
Cyprinus carpio (錦鯉)
A couple of small fish were not identifiable by me together with the poor method and the already starting growth of algae which makes the water quite opaque. The pond is small and muddy and a tough environment for most fish so I doubt I will find more species than those above.

That is sooo Last Year!

Forrest in the evening

Forest in the evening

The setting sun gave the forest behind my backyard a nice autumnal glow, so I tried it again with my cheap digi-snapper. This time was quite a disaster: too shaky, too much noise and a bit too wet but it is recognisable at least that the reddish thing in the background could be a forest. With a bit of good will. The whole process was automated with Hugin and I was a bit lazy when I checked the result, so the fence in the foreground does not fit very well. No further postprocessing because made no sense: you cannot change things that are not there in the first place. But I decided to play around with the different projections this time. More after the fold. Continue reading

Comet 168P-Hergenrother

The comet named 168P-Hergenrother broke up. With whom is yet unknown. This is old news, but I found the not-raw picture not very good to detect all of the four pieces they were talking about. If somebody knows about new and/or better but especially raw pictures (in the sense of having at least the full 16 bit resolution. I know I can’t get the real raw pictures without a lot of paperwork and a bigger mailbox), please tell me. Continue reading