Bryson Leidich

Photography and Photoshop

Digital Basics II - Camera Capture

Camera Capture Bit Depth

Most current cameras capture images using a 14 bit system, which provides original capture information of 16,384 values per color channel. This allows the camera to provide a considerable variation in the interpretation of color and brightness values in an image. This additional information is used to improve the accuracy of the interpretation of color values by the processor in the camera before providing you with a usable 8 bit image, usually a jpg. This additional information can also be used to produce raw capture files with a higher image quality. The digital conversion to 16 bit in the raw processing software expands this quality even further.

A camera processed image is converted and downsized to an 8 bit file which is then compressed using the jpg algorithm, which is a lossy compression that permanently discards a substantial portion of the original capture information. Compared to a 256 level per channel jpg a 4096 level per channel raw file has 16 times as much usable information. Skin tones in portraits, smooth tonal transitions in clear skies and transitions from specular to diffused tonal ranges are part of the reason photographers choose to use the raw file capability of the camera over a camera processed jpg. Of course, sometimes the convenience of the camera jpg for some types of images is fine where modifications to the image are either unnecessary or the jpg image quality is sufficient. The more manipulation of the image that is needed, the more the higher bit depth of the raw file becomes valuable.

Camera Resolution

Resolution easily confuses people as a pixel must be seen in the context of the device that displays it, or prints it. Since this can change the effective size of the pixel, we need to be sure we speak the same language when describing resolution. Understanding that pixels are not an absolute, but a relative measurement will help you understand resolution better. See Basics I for more information.

Cameras are usually described in terms of megapixels; the millions of pixels that they capture. A theoretical camera that captures an array of 5400x3600 pixels would be a 19.4 Megapixel camera as the total number of sensors equal 19,440,000. Sensors vary in size, which has nothing to do with the number of pixels on the sensor. A compact or point-and-shoot digital camera and a more expensive DSLR may capture the same number of pixels, but the quality of the capture will be very different. The captured noise, color fidelity, and dynamic range are functions of variables like the individual pixel size, the manufacturing tolerances of the sensor, the sophistication of the software in the camera and other factors.

Your decisions on what camera to buy need to be based on a number of factors with the pixel count only being one consideration. A high quality DSLR with a lower pixel count will produce a better quality image than an inexpensive digital camera with a higher pixel count. The actual size of the sensor must be taken into account as the size of the pixels is important. Smaller pixels invite problems with noise, so a lot of pixels on a small sensor could be problem compared to a smaller number of pixels on a larger sensor.

Camera sensors come in sizes. Full frame sensors are approximately the same size as a 24x36mm 35mm film camera. APS-H sensors are smaller, and yield an image that is the same as zooming into the image by 1.3 times compared to full frame. There are two APS-C sensor sizes which are equivalent to 1.5 and 1.6 times magnification. Smaller sensors do not really magnify the image - they crop into the image by those amounts so the effective image size given the lens focal length appears larger. The magnifications are relative to a full size sensor as we see the camera as the digital equivalent of a 35mm film camera.

There are also larger sensors for medium format cameras and capture backs and they are quite expensive. There are also smaller sensors, such as those in 4/3 cameras and point-and-shoot cameras. The smaller the sensor, the lower the quality will be, all things being equal, as larger sensors capture more light with less noise and more accuracy of color information. A cell phone camera captures on a chip approximately 4x6 mm compared to the full frame 24x36mm chip - a considerable difference. The phone sensor image may appear to be excellent when displayed on the small screen of the phone itself, but there are significant compromises. We are dealing here only with an array of pixels with a width and height, but no real resolution. Our image is simply x pixels by y pixels. While we sometimes need to know how this "resolves" to make an image of a certain size, this array size is actually the most important part of the description of an image, especially in the absence of other defining information.

Camera Settings

While less expensive cameras may restrict you to a jpg for capture, they usually allow you to set some parameters for the capture, such as the capture size, and quality. The best images will come from setting the camera at the highest quality settings available. Using the highest resolution settings will guarantee you a file that will produce larger images, but you may decide to back down a notch if your only purpose is to produce 4x6 inch prints for your photo albums. Recent cameras capture incredibly high resolution images, often more than you need for many purposes. Use the resolution information from earlier to determine the best setting for your purposes. Learn what the pixel count means and determine what you need for your photography. Shooting a photo intended for email or the web as a poster sized image simply requires more work after the shoot. That said, if retouching is going to be important, more pixels means more detail and easier fine detail work.

As you move into better cameras, you may also have options such as color space, color saturation, in camera sharpening, and contrast. If you take your image card directly to a lab for printing, rather than processing images in Photoshop, Elements, etc., then you should choose settings that provide you with the output you need. Notable would be the sRGB color space as choosing Adobe RGB will result in desaturated images from many printers such as kiosk printers.

If you intend to manipulate the images yourself, you will find that setting all of the parameters to the least of their values will provide you with a file that is less limiting. Increasing the contrast, saturation and sharpening is easily done, but your camera preview will be misleading, and many of these decisions are better made in post processing. Make your decisions based on what the files look like in your computer, not what they look like on the back of the camera.

One aspect of many cameras, and all higher quality cameras, is the ability to set the color balance. Images made in daylight have a different color than images made under indoor lighting, or fluorescent light. This is the setting most often referred to as the "white balance" and options include such things as daylight, cloudy, or tungsten, and custom. Auto is always a choice, and for the casual shooter often makes an acceptable decision in normal circumstances. Since the resulting image is "baked in" by the in-camera jpg process, paying attention to this option can help you improve your images and avoid color problems. These settings do not have an effect on raw files except to provide you with a preview.

Even further up the digital food chain is shooting raw files, or "unprocessed" files from the camera. This is a more complex decision, but provides you with files that are completely under your control. Processing the file for color, contrast, and even size, falls into your hands as opposed to the processor in the camera, and the results are what you decide they will be. The usual argument against shooting raw is that you have to process each image before you can use it, and that can be time consuming. Current software for processing raw images is capable of processing multiple images in batches, and once you are familiar with the process, making a set of final images is not all that complicated. While you cannot use the images directly out of the camera like a jpg, fixing a bad jpg image can take more time, effort and expertise than working with raw images. The best distinction between the processes, while not entirely accurate, is that shooting jpgs is like shooting slides, and raw is like shooting negatives.

File Formats for Capture

These are camera formats, or files that your camera can produce when you make an image. All cameras will shoot jpg files. A few compact cameras and most all DSLR cameras will shoot raw files.

JPG - Your camera captures a jpg image by default. This is a compressed 8-bit file (256 levels of information) produced by the camera's software from the original capture information on the chip. The file compression allows for a greater number of images on the storage device, but seriously limits your control over the results. Camera Style settings can modify the processing parameters. See Camera Settings for more information.

Raw - The original and highest quality capture information provided by your camera. Most current cameras capture a 12 bit (4096 levels of information) or 14 bit (16,384 levels of information) raw file and provide a processed jpg image to the camera for viewing purposes. However, the original file is not managed by the camera, and you will have to process the image yourself with special software. The format does not lock in decisions such as color temperature, for example, and offers the photographer the opportunity to produce very high quality final images with a lot of control. Except for noting the ISO, shutter speed and aperture settings, a raw file essentially ignores camera settings, but puts those settings into the file's metadata for use by some processing programs, which is usually optional. It is the preferred capture format for most professional and a growing number of amateur photographers. The gains in control over the image in all aspects is worth the effort.