Pages: [1]   Go Down

Author Topic: Tone Mapping Operators for Photographers  (Read 631 times)

Jack Hogan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 695
    • Hikes -more than strolls- with my dog
Tone Mapping Operators for Photographers
« on: March 01, 2018, 04:37:44 PM »

I've been reading up on tone reproduction operators, it seems to be a messy field of research still in quite a bit of flux.  Papers from a decade ago appear stale and recent ones become more and more esoteric, trying to be all things to all people.  But it seems to me that as it pertains to photographers the problem is not that complicated if the intent is perceptual accuracy: given a properly exposed/brightened raw capture with mid-gray at roughly L*50 or 18% in the linear data, what operator would best fit roughly 14 stops of DR into a user defined 8-10 stops viewed in typical conditions?  For Landscapers like us, in practice I think it could be simplified down to a couple of variables, the rest could be assumed away.

So assuming a typical sunny scene input captured and viewed on typical current quality (FF) hardware, what Tone Reproduction Operator would work best today if the intent were 'perceptually accurate'?

Grateful for any pointers and suggestions.



  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 448
Re: Tone Mapping Operators for Photographers
« Reply #1 on: March 02, 2018, 08:55:26 AM »

Although this post is from 2010, John Hable wrote about a few filmic TMOs with code snippets included:


Back in the day, John wrote the utility called GingerHDR - unfortunately no longer supported.  It was a plug-in for After Effects that merged and tone mapped HDR video that was output by the Magic Lantern Canon firmware (alternating high and low iso exposures, merged to EXR and tone mapped).  The video aspect made for some interesting issues, and Hable included optical flow analysis of successive frames to try to stabilize movement between captures, etc. I used it for still HDR images and the filmic mapping of the dynamic range gave very pleasing, "perceptually accurate" results with a little tweaking here and there depending on the nature of the scene.

« Last Edit: March 02, 2018, 04:35:51 PM by kirkt »
Pages: [1]   Go Up