Sunday, December 16, 2012

About picture correction and work optimization

As I have written in my post CyberView Settings, scanning is only the beginning. It is completed for me now (I have freed a lot of space in one of the closets by moving the boxes with slides to the cellar) and I am concentrating on next two phases: correction and tagging.

It is time to make a confession; this is the first post that is not “historical memory”. When I am writing this I have already scanned all my slides and negatives; all together almost 6000 pictures. All the posts so far (with exception of “View pictures on TV set”) were about events that happened several months ago, now I am setting myself to present time.

My first idea was, to do the correction and tagging at the same time with the same program, but soon I changed my mind and decided to do it separately. The reason is simple: I am lazy. Or, to say the same thing in better sounding words: it takes a lot of time to do the work on thousands of pictures so it should be optimized and automated as much as possible. But correction and tagging are very different tasks and each of them has different optimization possibilities.

Most programs have some kind of “automatic color correction tool”, but for my taste they are too aggressive and the resulting colors look unnatural, therefore I prefer to do it manually. But typically pictures come in sequences that were taken on the same film and in similar conditions (10 pictures of the indoors birthday party, then 6 outside skiing, etc.) and most of the time require the same color correction. The idea is to set the required correction values (color histogram is very helpful at this exercise, but more about it later) once on the first or on the most important picture in the sequence, save them and reuse them on all other pictures in the same sequence.

Tagging is a completely different task, it involves writing labels which tell the event or the place where the picture was taken (for example: “skiing”, “Italy”, “Corvara”), the names of people on the picture (for example: “Daša”, “Ženja”), etc. Tagging can be optimized by selecting pictures which need the same labels and applying the labels by selecting them from a pre-prepared list. But I will write about tagging in future posts. Now let’s continue with color correction.

So what is the best program to use for correction? I never especially liked Windows picture editor and I never even considered paying for one of commercial (Adobe, Corel, etc.) programs. My favorite and one of the first programs I always install on any system I use, was the freeware IrfanView. Although it started as a simple picture viewer, the functionality was later significantly extended and current version is quite a powerful picture editor which includes all basic tools (even more tools are available as plugins, including the IPTC tags editor). But I decided it is worth to invest some time to research first. As always I goggled around, downloaded several freeware picture editors and tested them.

The typical tasks to correct errors on scanned digital pictures include:

  1. Crop: Sometimes the scanner failed to find border and the picture contains black edge. Sometimes the photographer (for each such case I was surprised how I could have zoomed so stupidly) could have better selected the motive. In such cases the solution is to select a smaller rectangle and crop.
  2. Rotate: Sometimes the film was a bit sideways in the frame. Sometimes the photographer (for each such case I was surprised how I could have turned the camera so stupidly) forgot that vertical lines (like edges of buildings) must remain vertical on the pictures too. In such cases the solution is to slightly rotate the picture for a degree or two.
  3. Correction (in some programs it is divided to several tools like Brightness, Contrast, Gamma, Color balance, etc.): Older films are faded, each of the 3 basic colors differently. Sometimes the photographer (for each such case I was surprised how I could have …) or the automatic exposure of the camera under or over exposed the picture. In such cases the solution is to adjust colors.
  4. Mirror and rotate 90 degrees: It is easy to have a slide in the tray oriented incorrectly; in fact only one possibility out of eight is correct. If the people are standing on the head or the inscription is right to left, the solution is to mirror or rotate the picture.
For testing I selected several digitized pictures which contained at least one of above errors and tried to correct them with different downloaded programs. I took it as a sporting competition; each program was given 10 minutes to convince me it is the one I am looking for to do the tasks at hand. There were three programs that made it to the final: IrfanView, Gimp and DigiKam.

I have already mentioned IrfanView. I already knew it and I like it, but to use it on 1000’s of pictures I had doubts. It is a program developed for Windows and no other operating system. On the other side both Gimp and DigiKam are programs developed for Linux, but both are also ported to Windows. Gimp is known as the unofficial free replacement for Photoshop with countless features and possibilities to correct pictures, but (this is strictly my personal opinion, which might be completely wrong) it is an incredible tool, if the task is to correct several pictures to perfection, but not if the task is to correct 1000’s of pictures as quickly as possible.
In the header of web site DigiKam claims “Manage you photographs like a professional with the power of Open source” and it is not an exaggeration. Although the main focus of DigiKam is tagging of pictures (isn’t that nice, maybe I can use the same program for tagging and correction) it is also a very capable picture editor. As you probably already guessed, this program was my winner. I could easily use it for the 4 tasks above, that I have chosen for the test.

The question “is keyboard faster than mouse” is more than decade old and a lot of very smart people strongly disagree on the answer. I my humble opinion the correct answer is “it depends”. Personally I prefer working with keyboard in programs I know and for tasks I use a lot. I feel more comfortable to put fingers on the keyboard and keep them there then to haunt around for the mouse. In DigiKam most commands are available with keyboard key combinations and additionally it is possible to configure any command to any key combination. Next feature that convinced me is that last used settings of each correction command are automatically remembered. And another one I liked was the editing screen layout with size adjustable icons of pictures in current folder on the left, selected picture displayed in the middle and settings for currently applied tool on the right.

I am deeply thankful to DigiKam authors; obviously they value optimization highly; which is not a surprising fact in Linux world. It makes it possible for me to correct a sequence of pictures with several keystrokes and to typically process all 80 digitized pictures from 1 tray (now in 1 folder) in time under one hour. By process I mean using at least one, but possibly all of the 4 typical correcting tasks (see above). I will describe my preferred workflow and the commands I use in details in my next post.

No comments:

Post a Comment