HDR, Lightprobe, IBL Workflows


#1

The process of taking photos, prepping, stitching, fixing and massaging into a usable lightprobe or spherical panorama can be a little confusing and convoluted sometimes.

I have looked around for the ‘definitive’ way to do this, but like many aspects of CG there is no such thing as the ‘definitive’ method. It seems there are many many different ways to go about it, and of course pros and cons to each.

I want to share my particular workflow, and see how others do it.
My process works well for some lightprobes, but I have had trouble fixing things when they do not quite stitch perfectly.
First off, here’s my process:

1) Take photos:
-Set camera to manual everything and set a neutral white balance/exposure - checking this in bright areas and dark areas and finding a middle ground.
-Check angle (I start low first), check tripod is level (want perfectly horizontal panning, and lock-off tilt.
-Take 3 bracketed exposures (3-stops apart (the max on my canon 7d)). Pan a set amount (keep this amount consistent - depends on lens focal length).
-Continue taking photos, 3 exposures at each angle, tilt up, repeat, tilt up, repeat in necessary.

2) Combine each 3 photo stack to an HDR image:
- I use Photomatix to convert each 3 image stack to a 32 bit .exr - no tone mapping / adjustments. Photomatix batch converts from a folder of images, so I leave this to do its thing.

3) Stitch .exr’s into a panoramic image.
-I use Autodesk Stitcher to import and automatically stitch together the images. This may take a while, but the results are generally quite good.
-Render out a spherical panorama at 10% of original size to an .exr (I don’t need exceptionally hi-res for my IBL images).

4) Fix up:
- Clone stamp away any errors in photoshop, remember to stay in 32 bit mode.

  • Get rid of tripod and cameraman shadows etc., the sort of thing that would show up in reflections.
    -Save as 32 bit exr.

Done!

Now this works for me, but I feel it is a bit of a roundabout method, going from application to application, with a lot of in-between conversion and processing time.
Also, you are limited in the kind of editing you can do in photoshop in 32 bit mode - many of the adjustments, tools, filters etc. are 8 / 16 bit only, and don’t work with 32 bit imagery.

So, that is how I do it. How do you do it?

Some pointers on how to speed up and simplify my workflow would be great.
Thanks!


#2

Canon 1DS, Sigma 8mm fisheye lens, tripod with a nodal head.

  1. Do some test shots to work out what your ‘neutral’ exposure timing should be, and what your fastest should be (when the brightest light source in the scene is captured with no clipping).
  2. Then work out how to get from one to the other in 2-3stop increments.
  3. Shoot those exposures with the camera level at 90-degree increments.
  4. Load all the images into PTGui, set the crop and press go.
  5. Enjoy your hdri.

#3

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.