Linear conversion, anyone?

I've been doing some reading about tone curves and how they are used for processing raw file. And the one thing I learned is that there are more image data captured in a raw file than is actually used in a straight processed image. In other words, that blown shot you took the other day may yet be salvageable, depending on how blown it is.

To put this to a test, I got one of my shots that I know is way overexposed. Converting it from raw "as shot" in DPP gives me this:

Now, using the PSCS raw converter and adjusting the exposure and shadow levels in the converter, I get this:



And finally, using Image Browser (see note below why DPP was not used) to convert the raw file to linear TIFF, I ended up with this by applying a suitable tone curve in PSCS.



There are the 100% crops of the back and the nape of the horse, for comparison purposes. These are from the image converted "as shot".


These from the one converted using PSCS.



And this one is of the processed linear conversion image.



Other than the tone curve to bring out the "normal" image of the linear conversion image, no other post processing was done to any of the images to simplify comparing them.

For those interested, this is what I did to process the linearly converted image.


I started off with the linearly converted image which has that characteristic "very dark"� look:



I created a Curves adjustment layer and applied the following curve:



And the resulting image was shown above (the 3rd image from the top).

If this were a keeper instead of just a demo, I would continue on with the rest of the post processing steps in my workflow.

Please note that linear conversion can only be done with RAW files.

NOTE: I just discovered that Image Browser does a better job of converting raw files to linear images than DPP does. This really surprised me as I would have expected the opposite to be true. As you can see from the image below, there would be no way to recover the lost highlights if the DPP linear image were used.