Digital Photography Concepts: Camera RAW

by Paul Goodchild on March 4, 2009

This topic is discussed and questioned all over the web, so why do I feel the need to talk about it here? There’s just something not intuitive about some attempts to articulate it… we all have different way to describe things, so the more explanations out there, the better. Hopefully this will help clear up any/some understanding gaps for you…

Before the days of digital sensors, the main medium used for photography was ‘film’. This film is a thin strip of plastic upon which are laid silver-halide crystals that undergo a photochemical reaction upon exposure to light (electromagnetic radiation). This reaction is what forms the basis of film-based photo capture. The film is then ‘developed’ using various chemicals in a controlled environment to produce ‘negatives’, which are used as the template for the prints that are typically produced in a lab. As long you retain these rolls of negatives you have the ability to re-print as many copies as you desire. The negative roll doesn’t change, it cannot be edited, and is a representation of the light that hit the roll of photosensitive film right back when the shutter was clicked.

In the same vein, Camera RAW files are the digital equivalent of these film negatives. Basically light hits the photosites on the camera sensor, each site produces a signal and the camera processor interprets and stores the firing of these sites. With the appropriate software, the stored data that is the collection of these fired signals (RAW files) can be interpreted and output as an image either on-camera or on your computer.

As with negatives, these files don’t change and cannot be edited – at least the image data can’t be changed. So how does that differ from the JPEG files you get on your camera when you normally take photos?

In a big way. Or rather a lot of big ways. It may help to describe briefly what happens when you go from snapping a picture to receiving a JPEG image in most all digital cameras…

You click, the image is captured by gathering up the sensor data at the photosites, the colours are determined and a RAW image is created. The in-camera processor performs some post-processing such whilte-balance adjustments (based on what it believed the surrounding light to be at the time of capture), image sharpening, contrast and brightness adjustments and of-course, noise reduction. As it does this, it scales down from the RAW data range to fit the JPEG data range.

This JPEG image file is what the digital camera processor believes, given the settings you specified (if any) and the surrounding environment, what your final image should look like. If you want to edit it afterward so that it reflects what you saw, you need to consider the limitations you’re faced with. So what are the primary differences between camera RAW and JPEG?

Dynamic Range – RAW files store the image data in (currently) either 12 or 14-bits per pixel. This compared with JPEGs that store in 8-bits. This means, that for each pixel in a JPEG file, the pixel value can take 1 of 256 values. In the RAW file, each pixel value can take 1 of 4096 (12-bit) or 16384 (14-bit) values. That’s massive! If you make edits (which you will) you would much rather have a finer range of 4000+ options than a mere 256.

RAW file data are completely ‘pre-processing’. This means typical image manipulations are not yet performed such as white balance, sharpening, brightness and contrast, and noise reduction. This is good and bad. Bad, because there is time required by someone (you?) to do this necessary post-production, and good, because you can use much more sophisticated tools to make the changes than would be available in-camera. Not only that, it is a subjective change based on how you want the image to look, not what the camera believes is expected.

Following on from the point above, creating images from the RAW data means that you can create multiple images based on different styles from the original data while losing nothing as you move from one style to the next. For example, given a particular photo you may decide you’d like this in grayscale and you make the change and generate image (let’s say JPEG) from this. But, later, you decide perhaps you’d like a colour version of the image for a project you’re working on… had you snapped in grayscale, there would be no way back. If you have a RAW file, you can edit a colour image from that data. And you can do this all the while you have the original RAW file.

Extending the very first point, there is a cost with having more data per pixel: file size. A high quality 15Mega-pixel JPEG might cost you approximately 5MB. The RAW data file (in the case of my Canon 50D) costs on-average just over 20MB. When that is multiplied over hundreds of pictures, it’s a lot of space. This is also a consideration for the memory cards you will use within the camera itself.

I could keep going with further explanations and more pros and cons, but I think I’ve covered the main points. One more thing to note is that before you can have the option of camera RAW, you usually require a more advanced camera system, namely a Digital SLR. This however is not necessarily the case, as there are now high-end “compacts” that allow for RAW capture.

I hope this helps to clarify the whole RAW questioning you may be having. Should you take your photos RAW? My answer is: Yes. But you need the necessary storage space, the software to process, and the time to run the post-processing. The results are always better and once you’ve worked with RAW, you wont go back.

Digital Photography Concepts: Camera RAW

Further Reading