I'd like to shed a slightly different light on the topic. Why does
anyone want to process RAW in hugin? There are different motivations:
- to wrench the last bit of quality from the images that would be lost
in a TIFF: the green channel effectively yields 1.4 times the
resolution, for example. Potentially very useful.
- to save the RAW conversion - the RAWs are there, you want a pano,
you'd like to not have to launch the RAW converter before feeding the
data to hugin
I'd assume the first motivation will be relevant for special purposes,
e.g. calibration or focus stacks. But I don't think it's awfully
relevant in most paographers' every day work - converted RAWs or even
good JPGs will do the trick just fine. So my suspicion is that the
main motivation to use RAW is the second: save the conversion. Maybe
it's enough to just defer the conversion? I don't know about other RAW
formats, but Canon's CR2 contains en embedded 'thumbnail' - this sound
pokey, but in fact it's something like a good 1.5 megapixels in size,
and has well enough detail to run a CPG on it. In fact it's very
crisp, having been scaled down from the original 12 MP my camera
yields. So it's entirely feasible to extract this image, using
something like
ufraw-batch --embedded-image IMG_XXXX.CR2
and make a panorama from the preview images. Don't laugh. The embedded
preview has the same geometry as the RAW image. So you can - later,
and maybe in a batch run overnight, have the RAWs expanded and the pto
file upsacaled to work on the full-size images, and do a stitch from
them. I've written a pto upscaler precisely for the purpose - batching
the other stuff is (relatively) trivial, but a nice batch file would
of course help... The upscaler is here:
http://bazaar.launchpad.net/~kfj/+junk/script/view/head:/main/scale_pto.py
This one isn't hsi/hpi but plain Python. I admit I haven't looked at
the script in a while, but it should work nevertheless. The approach
has a few good points, like saving time and space - a bonus for laptop
users and travellers. Instead of the emberdded preview image, it's of
course just as well to use output from the RAW converter at smaller
scaleand in a compact format - in fact this is even better. Why?
In the proposed workflow I always felt there was one thing missing.
Hugin can very nicely take care of exposure, colour and lens geometry,
but there is one point which can be crucial which is left in the
domain of the RAW processors: tca. Correcting tca with a RAW procesor
changes image goemetry and stands in the way of my 'upscaled
thumbnail' approach. You can use an external tca corrector in your
toolchain on both the thumbnail and the processed RAW, but this
approach has always had less appeal to me since it involves yet
another warp with the accompanying degradation in quality - not to
mention the extra processing time.
Kay