This thing we call RAW is not an image format. “RAW” is just a description of a digital image file’s state, much like calling a vegetable “raw” is a description of its state. And those two states are similar in other ways: both the vegetable and the image file are unprocessed, but might be in any number of forms.
If you don’t know whether you have a raw tomato or a raw squash, then you can’t know how you should prepare and cook it. This also is true for differing RAW image files. Each camera manufacturer defines its own RAW formats, and those formats tend to vary between camera models, as well as manufacturers. Keeping up with all these formats is one of the challenges of developing RAW conversion software. But that same challenge can also be a revenue stream, as your users need to purchase updated versions of software for each new generation of camera they buy. We’ll look at this last issue in a later article.
Camera manufacturers provide software of their own for converting their proprietary RAW image formats into standard formats. But few of these programs see wide use, and it is not unusual for camera buyers to never load any of the software from the disc accompanying their new camera, and to work entirely with third party image processing applications that support many brands of cameras, and offer a broader range of features, instead of manufacturer specific applications.
RAW formats came about because camera manufacturers, to be competitive, needed to provide images in a new type of format that would allow the user more control in the rendering of images into the usual formats. Lacking a standard for such a format, each manufacturer took their own route in taking what was previously a proprietary internal format, and turning it into a RAW format for use outside the camera.
The key advantage of RAW files is that they are high bit. This means that each of the three color channels in the file contains more than the 256 levels of color data that is available in an eight bit per channel file. This extra data is pretty much invisible when viewing the file through the video path from the application to your screen. But it’s still there, lurking between the visible bits, waiting to be brought to light by editing adjustments such as changing the gamma (midtones densities) of the image.
So it’s edits that make the high bit nature of RAW files valuable. Even with a perfectly exposed image, you might wish to open shadow detail further that reality, or adjust the white point to a state you find more pleasing than the actual white point. In other words: if you edit your images, there is always justification for the extra data in the RAW file, no matter how accurate your camera’s settings.
There are a couple of other characteristics of a RAW file that are worth noting: gamma and gamut.
Devices measure color in a linear manner, twice as many photons means twice as light. But the human eye sees color is a less linear way, perhaps because it provides a more useful view of the world. The tonal response of the eye is defined by a complex function referred to as L-star. However the much simpler mathematical formula called gamma 2.2 comes quite close to matching the eye, and is much simpler to process.
RAW files are in device gamma, and require conversion to a human viewable gamma to be viewed. When a RAW, gamma 1.0, image is viewed in a RAW conversion applications, then it is converted on the fly to the display’s gamma, as defined by the display’s ICC profile. And when an image is rendered out to JPG or TIF, it is converted from gamma 1.0 to something in the range of gamma 2.2 as well.
A somewhat similar situation occurs with gamut. Gamut simply means the area that something covers. With color gamuts, it means the range of colors that a device or color space can represent. A printer has a color gamut, though changing the paper you print on can change that gamut. A display also has a color gamut, which in recent times has often been used in the marketing information, where it states what percentage of AdobeRGB the display can reproduce.
Source devices, such as scanners and cameras, do not have a gamut, as such. They have a response function, which needs to be mapped to a gamut (as well as a gamma) in order to be translated into meaningful colors. JPG files are converted in-camera to a practical color space. With high end DSLRs the user gets a chocolate or vanilla choice of mapping the images to sRGB, or the somewhat larger AdobeRGB.
RAW conversion software applications get the image files before the application of gamma and gamut conversions. They then use definitions for the particular camera involved to define what the original data means, and convert from their internal wide gamut, high bit, linear gamma format, to the display profile for viewing, or to the user’s choice of output spaces when rendering images to TIF or JPG formats.
So we have now described bit depth, gamma, and gamut, as they apply to RAW images, Image conversion software, and rendering to screen or output file formats. This information will be the basis for further articles in this series about Adobe’s DNG file format, and how end users can decide on what software upgrades to purchase, to retain the flexibility they need in converting files from their newer cameras, without breaking the bank on software upgrades.
(Thanks Ellis and Jim for the proof reading assistance; writing on the iPad has its limitations!)