1
\$\begingroup\$

I want to return my original picture from an HDR-picture which I made with Adobe Photoshop CS6. Does an inverse function exist to do this?

The HDR technique takes two pictures where there is more light in one picture and then makes one picture from these two, I think.

Is there any way to return the exact original picture from the HDR picture?

If so, I am interested in the mathematical characteristics of this technique.

\$\endgroup\$
1
  • 3
    \$\begingroup\$ Without the original file(s), you would only be adjusting the results of the HDR processing to approximate the image before the HDR processing was applied. \$\endgroup\$
    – Michael C
    Nov 12, 2013 at 23:27

3 Answers 3

8
\$\begingroup\$

It is very important to understand the difference between a high dynamic range image, and a high dynamic range image that has been tonemapped back down to low dynamic range for display on a standard monitor.

This is a high dynamic range image that has been produced from multiple exposures:

Looks dull and lifeless because you're viewing it on a low dynamic range monitor. What you can do is locally adjust the contrast, so as to maximise the dynamic range of your monitor in every part of the image. This process is called tonemapping:

If you've done this tonemapping step (which is easy to detect as it makes your images look like the above), then you're out of luck, you'll never get back to the original(s), as it's a highly nonlinear filter that takes whole neighbouring image areas into account.

However if you've merely done a merge-to-HDR-image, then that process is entirely reversible, you can get back any of the original images by just truncating the dynamic range at the appropriate points:

\$\endgroup\$
4
\$\begingroup\$

The short answer is no. The long answer is, how would it be possible, if you think about the image you now have. The process effectively blends different exposures together to produce a final result and then discards the original information.

The essential idea of HDR is to capture and expose the information in the brightest and darkest areas such that you end up effectively evening out the exposure across the range, maximizing detail. At that point, there's no obvious way to determine what areas should then become darker and what areas should become lighter without having the original information at hand or having some means of evaluating the subjects to determine which would logically shift darker or brighter.

\$\endgroup\$
1
\$\begingroup\$

There are plenty of tonemapping operators (you can take a look to some here: http://osp.wikidot.com/parameters-for-photographers) and each one is different. But what all have in common is that they perform a local adjustment of contrast.

Basically: you split the image in regions with different lightness ranges and apply lightness and contrast adjustments individually to each region.

Most of them make a continuous gradation between regions to avoid abrupt changes, thus when the contrast change is very sudden in the original image you see typical halos around edges.

I don't think any of them is reversible, since there could be two different images that yielded the same output and it would be impossible to tell which one was the original.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.