In earlier discussions we’ve looked at video of how to read a histogram and read why RAW files hold more information than JPEG. I’d suggest checking those entries out, if you haven’t already.

To date I’ve stressed proper camera exposure so file manipulation will be reduced in post-production (especially important when shooting JPEG). Speaking in general terms, the less an already processed file (JPG, TIFF, etc.) is manipulated, the more integrity there will be to the information contained within. We’ve seen that RAW files have greater flexibility for manipulation, allowing us to do much more with an image with fewer penalties (and sometimes gains) on output.

Now that we know how to obtain proper exposure, today I’m going to suggest adding a little extra exposure to your RAW files for greater detail. Yep, let’s push those RAW files toward overexposure…but without actually overexposing the image. This is also called giving the histogram a “push to the right”. We may even unlock some extra detail and more subtle tonalities.

Why would we want to do this? Because there is a lower signal to noise (s/n) ratio in the dark (shadow) areas of an image. In other words, shadows are noisier (we really saw that in the examples in RAW vs. JPEG, when we severely underexposed). But if we can add additional exposure to our image then we will lift the dark areas and shadows, record more detail and get less noise. The trick when adding exposure, though, is to not overexpose the highlights, blowing them out. We want to push the histogram to the right without pushing it off of the right. Then, when processing the RAW file, we will adjust the tones so they are where they look the best visually. By lowering the slightly overexposed shadows back down to where they belong we should (theoretically) see less noise and better detail in, especially, the darker areas of the photograph.

So let’s put it to the test. Here are two photographs taken with a Canon 20D (ISO 100, daylight WB). The image on the left was shot on aperture-priority auto exposure and on the other I pushed the histogram further to the right, using exposure compensation to +2/3 stop. You can click on the overall scenes to see them full-sized for comparison.

*Note: your camera’s histogram display doesn’t accurately represent the image’s 12- or 14-bit RAW capture, it represents the 8-bit detail in the image thumbnail. So there may be detail in the RAW capture that is not accurately represented by the camera histogram. In other words, the camera histogram is a conservative measure.

Canon 20D on auto exposure, resulting in a pretty accurate, middle of the road, real world exposure. This is maybe a touch dark but pretty close to how the subject looked.	On this shot I overexposed 2/3 stop, pushing the histogram to the right a bit. This image is lighter than the actual subject appeared.

The normal RAW exposure corrected in CS2. It looks good.	The overexposed RAW exposure corrected in CS2. It looks similar to the shot at left, just a hair different.
Here’s where we see a difference, in the details. This is the bottom of the door, a dark area. These are 200% enlargements so we’re really blowing things up. More noise is apparent here in the corrected “normal” exposure.	The originally overexposed image is much smoother in the gradations. At this 200% enlargement the difference that 2/3 stop makes is obvious.

This was a pretty simple visual explanation. For further details and the mathematics behind this process there’s a great entry on Luminous Landscape I highly recommend. As always, I hope this has been a practical help.

My camera is set to shoot RAW about 95% of the time. The reasons are many, too many to point out in one article so let’s hit the big ones, as far as I’m concerned. Using a pretty standard DSLR like a Canon 20D/30D/40D:

1. Think of your camera as having two components, a picture capture device and a picture processing device. When you shoot a RAW file the camera takes the capture and then stops there, it doesn’t process the image out. It’s up to you to process the image out later in software. You may then process that image out to an original JPEG, TIFF, psd, whatever you like, choosing your settings at that point. Let’s say you mistakenly had your camera set to daylight (outdoor) lighting when you were shooting under tungsten (indoor) lighting. No problem, since the file wasn’t processed out in the camera you can simply choose tungsten white balance in your software because you are just now doing the second part of the process. You then output an original image with the correct settings.

When you shoot directly to JPEG the camera both captures and processes out an original image to a set of specifications you’ve set in the camera (color balance, color saturation, contrast levels, etc.). You may have set the camera to normal levels or “punchy” color levels. If something was set wrong, like the color balance, you can’t simply reprocess the pixels in development, as you can with RAW. Yes, you can correct JPEG images in Photoshop. But there are problems with this…

Okay, it’s extreme time. I simultaneously shot a RAW (Canon CR2) file (left) and a large JPEG (right) of the same scene with a Canon 20D. These were shot with correct exposure but under tungsten light with the camera set to daylight color balance.
Here’s the RAW file, straight from the camera: ugly.	The JPEG file looks much the same straight out of the camera.
See the QPCard () in the middle of the scene? I put this RAW image into Photoshop CS2 (ACR), clicked the eyedropper on the gray square in the middle of the QPCard and voilá, it looks pretty darn good. Done with one correction to the RAW file.	I opened the JPEG in CS2, opened Curves and clicked the gray square in the QPCard with the eye dropper to neutralize white balance. Ugh. Trying this simple correction has give some whacked out results. Way worse is that the highlights are magenta and the shadows are green, a terrible cross-curve that’s hard to correct.
Detail of the RAW file shows that it’s sharper (we would have expected that) but look at the smooth tonal gradations in the metal. All with one click to an initially poorly white balanced image.	Noisy, off-color and not as sharp, the poorly color-balanced JPEG needs lots of love to make it acceptable. Way more than I’m willing to give.

But what if you are going to process that RAW file out to a JPEG anyway? Then both files end up as 8-bit in color anyway. Won’t they be the same? Here’s where RAW shines. If you make color, contrast and exposure corrections to a 12-bit RAW file in your software, then when you output it to JPEG the pixels are all in the right place, like a perfectly laid stone walkway.

Now let’s look at the camera-original JPEG. It’s like a stone walkway that’s already been laid out with all the pixels in place. Yes, you can rearrange the stones / pixels but compromises will be made by the software. It will have to interpret / interpolate colors and gradations in difficult areas. It is only pulling from an 8-bit palette so the color options are more limited than with RAW. You will see this in action in these underexposed photo examples.

On these shots the white balance is set correctly to tungsten. But now let’s mess up our exposure by underexposing by two stops. This is an extreme example but any underexposure does degrade the image.
RAW at -2 stops.	JPEG at -2 stops. They look much the same
Okay, in CS2 (ACR) I did a really quick correction to the file and processed it out. This has pretty darn good shadow detail for being so poorly exposed in the first place. We can thank all that information in the 12bit RAW file!	In CS2 I did a quick curve correction to the JPEG and got the highlights and midtones looking acceptable. But where’s the shadow detail? There’s not as much info to pull from when starting with an 8bit original.
Here’s the real proof. Way more detail, sharpness and less noise in the shadows.	The JPEG can’t compete with the RAW. With more work it could look a little better but it’s clearly outclassed.

With RAW files you can take the RAW image and make all of your choices about size, color, contrast, etc., and then output it to a new original each time (which might be a JPEG). This is greatly aided by the fact that there’s 12-bit color. The software can make easier choices and less compromises. Nice.

And if you don’t like the image, just go back to the RAW file and reprocess it out how you want it. Look at that, it can be a very different but totally original image with uncompromised information. You can create endless originals from the RAW file, all with solid, full information, regardless of your processing choices. Amazon has great resources for exploring various digital RAW workflows.

3. I really think we’re in the stone ages of digital photography. That’s another reason I shoot RAW. Since I’m processing the files out in software (and not the camera) I’m also thinking there could be a great advance in RAW processing software someday. Wouldn’t it be great to have this RAW, uncompromised digital information and be able to output it to even higher quality because of an advance in software?

Check out the examples and see if it makes sense. I hope this has been a help. If it has, please let a friend know. If it hasn’t, please let me know!

Thanks,
Jim T.

(subscribers, video may be viewed on the website):

	Reference photographs used in the video
normal histogram	overexposed histogram	underexposed histogram
	Canon 20D on aperture priority auto / no exposure compensation
black card	gray card	white card
	black/gray/white cards