This is perhaps not one of my usual posts on this site since I generally do not post stuff about my photography hobby side here. But then, it is about computer software after all so, here you have it anyway…
I use Adobe Lightroom for all processing of my photos, especially the ones I shoot in RAW-format. The latter which I, of course, use as much as I can. The only ones that I shoot on JPEG are the ones that I take using my Ixus which I only use when I, for some reason (usually space constraints), cannot or do not want to bring any of my “real” cameras. Oh, and of course the “emergency” shots that I shoot with my smartphone. I find it amazing that some people are content with the absolutely abysmal quality of photos taken with a smartphone (any smartphone).
Anyway, back to the actual subject. Lightroom v5 came out some time ago. This week I decided to upgrade my installation from v4 to the new version. The changes are not exactly blowing your socks of (it actually feels quite a bit faster but that might just be my imagination) but it has one feature that I really appreciate. It has an automatic perspective correction. When I am not using my EOD 5D with it’s proper viewfinder I often end up “shooting from the hip” in the sense that I need my reading glasses to see any details on the screen on the back of a digital camera. A lot of the time, I just do not have the time to put them on. So I pretty much make a best guess when taking the shot.
Since a lot of my shots are made with a wide angle lens setting I often get lines that are distorted all over the place. Lightroom v4 had tools to deal with that that worked quite well but they were all manual. Sometimes when you had to make multiple corrections it could become quite tedious to get it right. In Lightroom v5 there is a automatic function that analyses the image and computes the right corrections to make. Sure, it is in no way perfect and sometimes it runs completely amok with your poor photo but more often than not the result is quite good and if not perfect so at least a good starting point from which you just have to do some finishing touches.
Below are two examples of the results of the automatic perspective correction in action. They are both fully automatic examples in that I have not made any manual corrections afterwards. I am somewhat impressed that it managed to straighten up the old stronghold (Glimmingehus) in the second picture without, at the same time, screwing up the house to the left.
As I said, I also changed my workflow a bit at the same time. Specifically, before I did all the tagging, star rating etc. in Lightroom except the face tagging. This I did in Microsoft Live Essentials Photo Gallery by letting it first auto-tag the ones it recognized and then add the ones it didn’t find afterwards. Although I could use the face recognition of Photo Gallery this had several drawbacks.
The most annoying of them were that it operated on the published photos. If I made some changes to them in Lightroom and re-published them I had to redo the face tagging step. The automatic face recognition was also less than perfect. Actually, in my opinion, it became less and less reliable as time went on. It either missed the faces altogether or just marked them as not recognized. Of course it would tag photos which had people in them that did not look towards the camera.
What I really wanted was some solution where I had a single source, that is Lightroom, where I could re-publish the photos after some minor changes without having to post-post-process them again after each manipulation. A lot of people in various forums deplore that this new release of Lightroom do not have automatic face recognition. Personally, I start to think that this is not really a bad thing. My experience of automatic face recognition is that is really do not work very well and you have to do a lot of additional manual work anyway. So why bother actually? I read a post on the Lightroom forums that said that the face-tagging by tools like Live Photo Gallery or Picasa made you lazy and that you missed a lot of photos that should have been tagged. I tend to agree.
So I decided to see how it would be if I scrapped the automatic face recognition and just tagged everything manually in Lightroom. To my surprise this actually proved to be not only superior in that you got a more accurate tagging of who was in each shot, but it was also less work. Being the person that took the photo in the first place it is really dead simple to identify who is in the shot and with the keyword sets in Lightroom (as well as the other tools like multi-selection, copying of tags etc…) where you can set a keyword tag with a single click it is really fast to do the people tagging at the same time as you do the rest of the tagging. Also, when you do it in Lightroom you do it at the source and do not have to redo it if you make some changes or publish it to some other place than the original one.
Since Lightroom supports hierarchical tags it is easy enough to keep a nice structure of your tagging and even though a lot of tools do not pick up the structure itself you can use it fully as long as you are inside Lightroom.
Now I just have to decide if I should go back and re-tag my old photos. The gadget/computer/fiddling geek in me tempts me to do just that 🙂