The Power of Raw

Check out this tutorial by Mahesh Thapa on getting the most dynamic range out of a photo by using a "double exposure" raw process:

http://forums.dpreview.com/forums/read.asp...message=3850801

Since you can effectively "bracket" your RAW conversion, you can overlay two photos and expose one for highlights and one for shadow detail. In this example, the process yields Mahesh an extra 1.5 stops of exposure lattitude.

Cheers

James

Thats a great technique James. I have used similar stuff before on land.

Now If I can just figure out how to get a tripod underwater so I can get two or three identical shots I will be all set!

Thanks for shaing.....I have copied that one.

You've just gotta love RAW, at least in situations where there's a chance you won't get the exposure perfect (ok, for me that's every picture )

Now If I can just figure out how to get a tripod underwater so I can get two or three identical shots I will be all set!

ULCS Sells a UW Tripod ;-)

Hi Bob,

You don't have to take two different shots. Just make two different exposures out of the same RAW file.

So you don't need a tripod and you don't actually bracket w/ the camera. You bracket w/ the RAW conversion software.

Cheers

James

James,

I think youwill find in practice that you almost have to have a tripod, if the two images dont match you will blur edges when you try to overlay them. At least that is my experience.

If I am wrong, please enlighten me on how you could match the subject edges on two exposures that aren't identical in composition.?

Bob

oopss I just read that closer........ That is a cool idea, and raw would let you get a stop or two if you nailed it on the front end.

My Bad

Yes, you got it Bob.

You take the shot - setting the camera exposure for the highlights (sunball, sunset, shiny barracuda, etc).

Then you go to the RAW converter and make TWO exposures from the RAW file - one "light" and one "dark." Then you overlay the two using layers. Finally, you erase part of the topmost layer to let the highlights come through. (for example, you lay the "dark" layer over the light layer, then you erase the "dark" layer from over the "open water" or the rocks, or whatever.

Cheers

James

I think it would be interesting for someone using this technique to produce a composite they feel is a compelling example of dynamic range enhancement. I'm completely convinced the same effect can be produced without using the double RAW conversion. All the data comes from one single exposure, so how can dynamic range end up improved? It can only end up worse that the original though manipulation. In this case, the technique simply produces a gamma "curve" made up of two line segments joined at a discontinuity. Provided you use 16-bit TIFF, you can do this with one conversion.

Fred Miranda has an action that is designed to take two bracketed shots and merge them to increase dynamic range.

Is that a challenge?

I'll provide a RAW file if you want to have a "shoot out" :-)

I'm sure yours will look better than mine - I suck at Photoshop.

Cheers

James

Well, doing it manually has advantages because you can selectively allow the brighter or darker image to show through....

But I'd try the action first to see what happens.

If you take two shots then you clearly have an opportunity for improved DR. That's what FM's action is for. Taking one shot and converting it twice isn't the same since all the data is coming from the original shot. Now, if people are converting from RAW to 8-bit, they could easily be convinced that it does work.

I think you may well like the results from the technique, but I'm confident it can be duplicated without the double conversion. I'm curious to know if others disagree.

Unfortunately, taking two identical shots underwater is virtually impossible.

Cheers

James

Unfortunately, taking two identical shots underwater is virtually impossible.

Hey! I said that first.

I've actually done this with a single JPEG as well; it's often amazing how much detail is 'lost' in the shadows. If you only want to bring it up by a stop or less you usually can. With raw of course there are more bits to play with.

Luminous Landscape has a tutorial on how to do this in PS, and he lists several methods:

http://www.luminous-landscape.com/tutorial...exposures.shtml

Nice link Sean.

I have been doing it with JPEGs too. You can get improved balance with a single exposure, though obviously not ideal.

Assume I have an image that has some dark areas which I want to bring into balance.

I don't have a raw convertor, so once in Photoshop, I duplicate the image onto another layer. Then, I choose the top layer and use the Adjust>Levels to bring up the shadows to the desired level. This will improve the dark areas, but make the highlights blown out. Then I apply a mask, and paint away the areas that did not require increased exposure. Once you flatten the image, you have brought the image into balance.

You can't recover missing data this way, but if, say, you have an obvious shadow due to strobe placement, you can improve the image quite a bit.

I have used a different, more automated technique, using a combination of free stuff and my own perl scripts:

1) capture images at different exposure time values and load them into the computer.

2) use the AIR automated image registration package to line up the images, with no manual aligning. http://www.loni.ucla.edu/NCRR/Software/AIR.htm. I had to write some converters and scripts to make this usable. Even with a tripod, imho,

this is necessary.

3) import the group of aligned images into HDRShop (http://www.debevec.org/HDRShop/). HDRShop is a basic image editor capable of manipulating images representing many more orders of magnitude of brightness levels then fillm or your monitor can capture. HDRShop takes the multiple exposures and combines them into one "high dynamic range" image whose "exposure" can be adjusted after the fact, etc.

4) use the "tonemapping" plugin from HDRShop to produce a displayable image. Think of this plugin as doing "automated dodging and burning",with a rough simulation of the eye's response to light.

Unfortunately, I only have one example image that is web accessible, and it is a BAD snapshot taken to email to someone during the big blizzard. It is a picture of the view outside of my front door at night. However it does show the technique off well:

http://www.lplizard.com/chrisg/tmp/snowntm.jpg

1) The difference between the brightest and darkest area that I was able to capture was over 500000:1.

2) this image is IMHO pretty much impossible to have produced from this scene using conventional photographic techniques. Look carefully - the snow in the foreground is infinitely brighter than the houses in the background, yet both show fine detail, while still conveying the impression of a being a night scene. Flash was not used.

3) the output image matches pretty well what the scene looked like to my naked eye. No single exposure value on the camera looked much like what my eyes perceived.

4) it wasn't much work to make (for me..the computer had to work hard).

No comments on the quality of the actual photography! I'm hoping to get some time to apply this technique to actual interesting well-shot photos soon.

How you make two different exposures out of the same RAW file?

I did not sow that option in the RAW converter that came with my Canon and not in the BrrezeSys converter that I have.

Just make the first one and save it, then up the EV by 1 stop (or whatever's necessary) and make another one.

Cheers

James

Just make the first one and save it, then up the EV by 1 stop (or whatever's necessary) and make another one.

Thanks James,

Maybe is my poor English that I miss something here, what I understood is:

You take one shot in RAW mode, and then in the RAW converter you save two images from the same RAW file each in different EV stop.

The problem is that I don’t see any option to change the EV setting in the BreeseSys converter that I have, maybe it just don't have it and I need another converter?....

Arnon