Home Creators Posts Import Register Favorites Logout

Content

As promised, I've now published the script that I use to merge exposure brackets to HDR files.

Github repo: https://github.com/gregzaal/HDR-Merge 

It uses Blender's compositor to do the dirty work of choosing which parts of each image should be included in the final HDR file to get the maximum dynamic range, and then uses Luminance HDR to tonemap those HDR files into JPG files that are easier to stitch with.

No compiled binaries are available (yet?). It's a python script which uses PyQt and Exifread libraries for the GUI and exposure bracket detection respectively.

It can be used now, but could still use some polish to make it easier and fully functional for you (e.g. ND filter color cast compensation is hard-coded for my filters, some assumptions are made about the order and consistency with which the images were shot). But I figure it's better to publish it now and sort those out later. Any help with this is appreciated.

Next week I'll work on documentation on exactly how to use this script and some details about my workflow, including how I handle various cases like capturing bright sunlight or stars, avoiding lens flares, dealing with crowds, etc.

Comments

Ashfall Art

That's awesome thank you for sharing this

Nikola

Hi! I just get the 5$ tier, congrats for the crazy good hdris! Will you get into the details of how the blender compositor is used for obtaining the best DR? Can i use LumariverHDR to get the same effect? All this is so very interesting :)) I Get good results even with PSD HDR merge tool^^

Poly Haven

Thanks! I definitely want to get into those details. It's been a while since I've tested all the possibilities, but there are definitely multiple reasons why I decided to come up with my own solution. I plan to write some blog posts about my workflow this week, though I'm not sure what and how much I can cover.

Nikola

I can't encourage you enough to cover everything you know about the subject, i have so many questions about everything :)) Maybe do some parts on video, often techy stuffs are easier to explain/watch than to read on the blog. The Interior HDRI Stitching live stream is great with so much mysteries still in suspense^^

Tobias Jakobs

"Next week I'll work on documentation on exactly how to use this script and some details about my workflow, including how I handle various cases like capturing bright sunlight or stars, avoiding lens flares, dealing with crowds, etc."

Poly Haven

Sorry for the delays, it's turning into a bit of a monster. I don't want to spread bad techniques or false information, so I've been doing some research and refining my workflow, testing various things to do with color accuracy, etc. I also actually want to rewrite this script, as it seems there are some issues getting the UI library to install. Rest assured, I haven't forgotten, it's a high priority for me atm, it's just more work than I anticipated.

Nikola

"i don't want to spread bad techniques or false informations" But your current workflow works :) You are using it everyday and we are more than happy with the results! Maybe you can release your research by parts with a summary so we are aware of your process and where you are at the moment and share your experimentations. Your process is interesting to us, i think we are not looking for a perfect finished polished documentation. Knowing where and how things don't works is important too :) Do you use the magnetic filter holder 'Xume' to manage your filter use during the hdri capture process?

Poly Haven

I'm a bit of a perfectionist though :P once I've discovered that I've been doing something wrong, I'm not going to waste time documenting the wrong workflow. I have to research the better workflow enough to be confident that I can teach it. Those Xume filter adapters sound wonderful! I've been contemplating getting a square filter holder with new filters just to speed up filter swapping, but those magnetic adapters would mean I can just use my existing filters :D

Nikola

Haha yeah i know that perfectionist thingy too^^ but we know it's kind of a curse ;) By how much can your current workflow be so wrong while producing good hdris? What's off, the capture process or the merging? Anyway i am way too impatient! Yeah Xume filters looks good but i'm afraid that the magnetic force would be too strong that each swap would risk moving the camera alignement + the added thickness giving vignetting. I returned my 67mm filters and bought an up ring 67 to 72 and a 72mm nd400 haida slim proII, to see if the vignetting is ok with the Xume, on the 10-18mm+200D. But what happen if a second filter nd8/64 is added? idk yet.. Square filters aren't the best because of the filter holder, the light can seep between the lens and the filters, just in case we need more artefacts.. But i have seen square holder filters with special design that prevent the light from seeping. Price's not the same. I use an app on my phone qDslrDashboard that gives me the bracketing thing i don't have on the 200D or via Magiclantern. It's good but there is some time between each shot.. but totally usable, more if the lowest shutter speeds can be avoided. I will try a capture process like that (it’s just a base) : 1/ 4000 f11 100iso nd400+nd8/64 (for the strongest lights). 2/ 4000 f11 100iso nd400. 3/ 500 F11 100iso nd400. 4/ 60 f11 100iso nd400. 5/ 4000 f11 100iso no nd. 6/ 500 f11 100iso no nd. 7/ 60 f11 100iso no nd. So each shot is 3stops different and i avoid the 1/8s and 1s shutter speed. But then i don’t know how to use your script to do the merging in blender (I’m a basic c4d user on hackintosh with no programming skills, sorry ;) but i have a Win10 double boot, so Yay!) But i get good results from the HDR merge tool in photoshop. Does the HDR merge math inside ptgui is any good? idk how to merge hdris when filters aren’t used for the whole picture (but only for the brightest frames). But i know how to change the exif with exiftool thanks to one of your old post on photo.stackexchange.com ;) So, here we go!