Earlier this year, a photo buddy of mine (Jeff Kohn) introduced me to TuFuse. It’s a program that combines multiple exposures (from a high dynamic range, HDR, scene) into one image. TuFuse is shareware.
The basic interface of the program takes you back to the old days of DOS To run it in Windows, you have to create a batch file that establishes the path to your files, calls up the program (with optional switches that adjust the processing options), and tells the program which image files to combine. It’s fairly easy and straightforward once you’ve done a run.
TuFuse will accept 16-bit TIF files and produce a 16-bit TIF as output.
I finally got around to seriously using TuFuse and tackled some of my older, difficult images. First up was a 5 exposure scene shot last year.
click for larger pic
The above scene was shot with 5 exposures (all in RAW). The result, using nominal processing settings in the program, was fantastic.
I still ended up layering (with masks) some content from the original files over the TuFuse output in PS (Photoshop). Using layers and masks is my normal approach to combining multiple expsoures, and it’s tedious and labor intensive. I’ve tried manually combining the exposures for the above scene about 6 times in the last year and never got a result that completely satisfied me.
For comparison, here’s my original attempt at exposure blending – manual blending only (i.e. using only layers and masks in PS):
With a scene like this (one that has a very bright source of light in a localized region in the scene) simple layering with masks is quite difficult. Imagine concentric rings of differing exposure around the point light source (the sun). This is further complicated by the mountain in the upper right – this mountain is one of the darkest parts of the scene and renders nearly black in simple layering because of its close proximity to the sun in the scene.
The beauty of TuFuse is that it handles these areas of difficult exposure transition well. And then you can layer in bits of the original files into the “easy” areas of the scene (not that the TuFuse output is bad, but the original pixels just seem like a better choice to me where possible).
Since I was pleased with the TuFuse results, I also tried the sister scene to this shot, one that I’d not processed before:
click for larger pic
This was another 5 exposure blend done in TuFuse, and then supplimented by bits of the original files in PS (i.e. layering in, using masks, some of the original files over the TuFuse output in PS).
To reiterate what I’ve said before, shooting wide dynamic range scenes (i.e. scenes that contain a dynamic range greater than what the camera can capture in one shot) is quite difficult, and there are different approaches to doing it. Each approach has its pros and cons.
I prefer not to use graduated neutral density filters (the traditional method of shooting such scenes) because it’s a one-size-fits-all approach that isn’t always well suited. Take the above scenes for example. A graduated ND would have made the upper right-hand mountain very dark (much like my original efforts in blending the scene manually). There is no ND filter made that has a ND spot to handle scenes like this, and if such filter was made, the size and ND strength of the spot would have to be just right.
HDR programs (like Adobe HDR and PhotoMatix) have not worked well for me in the past. In fact, my first attempt at blending this scene was done in Adobe HDR. I spent long hours with it trying to fine-tune the processing and output with only poor results.
Manual blending (using layers and masks in PS), when done with care and skill, seems to produce superior results. The con is that it requires a lot of processing work. But I feel that the results are the best that can be achieved and also make the most and best use of digital post-processing (or, digital darkroom, if you like). Using TuFuse definitely helps to handle those tricky areas, like the upper right hand corner of the scenes above.