Helpful Posts:
0
-
3rd November 2009, 08:29 PM
#1
HDR in general. First post here.
Hi to everyone using these forums. I have lurked for a while and must say that the "atmosphere" here, as far as I've seen, is distinctly amenable.
I've been using HDR in a modest way for quite a while with VR interior panoramas, for which it seems to be highly appropriate. I've stuck with Enfuse and the associated front end EnfuseGui. My approach has been pretty crude, in that I tend to use Enfuse on its default settings: every time I've changed anything I always seem to get worse results. However generally I find that Enfuse does a very good job. To control the end result I tend to select three shots from a seven-shot (@ 1 stop) bracket sequence. Which exposures I select tends to depend on how well I've estimated the "central" exposure value. Amongst other things I'm doing a project on notable churches and cathedrals, so the DRs are quite extreme. I use PTGuiPro and although this has HDR functionality, the couple of times I tried to use it it appeared to be slow and none too great: not a comprehensive test by any means. Workflow wise it sems a lot more sensible to pre-process the shots before doing the stitching. I'd be interested to hear what others think about these options.
I know there are a few commercial HDR/tone-mapping applications out there. Part of the attraction of Enfuse, of course, is the fact that it's free. It's pretty fast too, like most lightweight DOS-based applications. Right now I don't want to spend anything I don't have to. However I'd be interested to know how people who have tried the range of options intensively feel about the best approach to HDR in this context (ie HDR interior VRs).
I suppose what I'd like to be able to do a little better than I can with Enfuse (sometimes anyway) is to control what I can only describe as the EQ curve of the tone mapping. Sometimes the results tend to be biased toward one end of the dynamic range and need some PP on the final stiched TIF which doesn't help the final quality.
The other thing that occurs to me right now is that in the past I've done most of the RAW conversion (I shoot Nikon) in ACR, just correcting the ca and not tweaking anything else. Nowadays I tend to use NX2 for non-VR stuff as its gives much better results, however I haven't tried any DR tweaks on files to be HDR-d. Obviously I should, however applying a set of batch adjustments to three sets of ten files is an additional overhead. Anyone got a comment on this?
That's enough for a first post. Of course none of this stuff has any urgency.
R
-
3rd November 2009, 09:17 PM
#2
Re: HDR in general. First post here.
I only used Enfuse once, but the results were very similar to TuFuse, which is the tool I use now from time to time. It's the only HDR tone mapping (well, the purists will say it's not a tone mapping tool, just exposure blending, but it's understood) that provides results reallistic enough for me. Since it simply mixes the colour pixels you provide it in the source files, you will never occur to get strange colours.
I always use if with the default parameters, like you do, because I prefer a centred histogram with good local contrast in all areas of the image (which is the difficult part of any tone mapping process). In that way with a single additional curve I can adjust overall brightness and contrast to get the finished image.
I would recommend however half the shots you are currently doing. Shots 2EV apart are perfectly usable and will save you space, processing time and possibly improve sharpness for reducing the progresiveness of the blending.
Regards.
-
3rd November 2009, 10:44 PM
#3
Re: HDR in general. First post here.
Hi Guillermo
thanks for your comments. I have considered Tufuse but not yet tried it.
On bracketing. Nikon only allows 1-stop increments. I've found that three either side covers me even if I miscalculate the "centre" ev. Sometimes I get away with five shots but in the average unlit church, in summer, it's easier to over-shoot and discard. Just to be clear I only ever use a maximum of three exposures, chosen unscientifically, however they are usually two stops either side of centre, as you suggest. Except the "centre" exposure isn't always in the numerical centre.
R
Tags for this Thread
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules