The Shoot

Resolution

A digital camera has an imaging sensor consisting of millions of . To record an image, each pixel builds up a tiny charge of electricity in response to the light projected onto it by the lens. Each megapixel represents a million pixels. The more megapixels an imaging sensor has, the higher the camera's potential resolution. The higher the camera's resolution, the greater the detail and the larger the size of print that can be made with acceptable detail. The need for detail as well as final print or display size can be a deciding factor in choosing a camera for any given photography project.

The two charts below provide a rough guide for matching megapixels to print output.

	Figure 1 We use the term “guidelines” in regards to matching sensor size to document size. The interplay between megapixel rating, sensor size, ISO setting, lens quality, camera technique (focus, lighting, camera stability) and optimization methodology makes a straight megapixel size to document size chart only a guide, not a standard.

In our tests, a highly detailed subject is captured under comparable conditions, ISO setting, and lens quality, and aperture that shows that the additional detail provided by a larger sensor (both physically larger and containing more photosites/megapixels) will be evident even in a small print. However if two cameras have sensors of the same physical dimensions, the camera with more megapixels may not display better image quality due to a whole host of factors already mentioned.

In any case, since resolution is a square function, the megapixel rating of a camera has to quadruple before the print size can theoretically double (given the same pixels per inch). Our charts above assume a print resolution of 300 ppi. However there is actually quite a bit of flexibility with regard to print resolution. Very acceptable prints and offset reproduction can be made at somewhat less than 300 ppi.

Although 300 ppi has become a standard default resolution for reproduction of digital image files, the full story on ideal resolution for print reproduction is somewhat more complex. For instance, Inkjet printers normally work best with resolutions of 180 ppi to 360 (some say up to 480) ppi, such as 8x10 inches at 180 ppi. Continuous-tone printing (e.g., on film recorders or dye-sub printers) requires resolutions of 240 ppi to 400 ppi. One of the more persistent myths is the most commonly quoted offset-printing standard of 300 ppi. However, resolutions of 1.3 — 2 times the halftone screen are considered safe. For example, if the images will be printed as 150-lpi halftones (common for magazine printing), the appropriate image file resolution range would be 195 ppi to 300 ppi. Newspapers print with coarse screens. A common newsprint resolution is 85 lpi, which works best with an image resolution of 170 ppi.

With regard to inkjet prints, since our eyes are limited in their ability to discern between a print made at 180ppi and a print made at 300ppi, it can be hard to discern the difference between a 6MP and a 16MP camera in an 11x14 print especially if you factor in image viewing distance.

Another feature of digital capture files is that they can be interpolated up (and down) more successfully than scanned film. Although up rezzing doesn't add detail, it does generate extra pixels preventing pixelization.

Form factor

Figure 2 Digital cameras come in all sizes and shapes. You can even make phone calls with some of them...

"The best camera is the one that's with you" – Chase Jarvis

What Chase says is undoubtedly true. While Chase backs this up with a charming online portfolio of iPhone images, we can be sure that he has professional cameras with him when he is shooting corporate and advertising projects for paying clients. Small cameras that fit in a pocket may make interesting photos, maybe even great ones, but they cannot produce images that compare – in terms of resolution, sharpness, color depth and accuracy – to high-end DSLR or medium format digital back cameras. Single shot cameras, in turn, cannot compete with high-end scanning backs for ultimate resolution, but who other than Stephen Johnson is prepared to haul one of these monsters into the wilderness to shoot landscapes? 

For most of us, the trade-offs occur in a narrower range of options. A sports shooter is likely better off with a 12 or 16 megapixel camera that has a large buffer and accurate focus tracking than with a 21 or 24 megapixel camera with a smaller buffer and less robust focusing abilities. Photographers that shoot landscapes, architecture, advertising, fashion and beauty are often more than willing to deal with large, expensive camera equipment in order to give their images a technical quality that cannot always be obtained with smaller, lighter and less expensive equipment. If there is one generalization that can be made, it is that on the high end, especially with regard to medium format digital backs, image quality has plateaued to a great extent. The difference between incremental jumps in megapixel counts has become increasingly hard to discern while much greater improvements are being seen in the image quality of compact cameras and sensors.

Crop factor

Another important factor for photographers to consider is whether the sensor is “full frame” or will crop the angle of view of the lens. When the digital sensor is smaller than the original format covered by the lens, the angle of view of the lens is cropped. Some people create confusion by referring to this as "magnifying" the image. The image is not magnified, it is cropped. 

For instance, standard 35mm format is 36mm x 24 mm in size. Many DSLR cameras have sensors that are 23.6mm x 15.7 mm, which is approximately 2/3 size. This results in a crop factor of 1.5. A 300mm lens on a 1.5 crop factor camera will have the same magnification of a 300mm lens on a full frame camera, but the image will be cropped to the same angle of view that you would get with a 450mm lens on the full frame camera. This can be an advantage for certain types of photography where a tighter crop is desirable, such as wildlife and sports photography. This is not so advantageous for those needing wide-angle views, such as architectural photographers.

Figure 3 The red crop line shows the angle of view of a 1.5 crop sensor as shown within the angle of view of a full frame sensor.

Volume

Shoot volume can be another important consideration in camera choice. Although Moore's law has allowed us to keep up with larger camera files in terms of processing and storage, there are still choices to be made with regard to balancing workflow speed and shoot volume. Matching camera resolution to the output resolution needed for your project will ensure an efficient workflow.

Speed

Speed of capture involves capture format choice (raw, JPEG, or raw + Jpeg) frame rates, camera buffer size, and data write speed. Vast improvements have been made particularly with regard to camera buffer size and data write speed. Shooting raw files is no longer the painful experience it was just a few years ago, allowing even sports shooters to shoot as fast or faster than they could in the days of film and motordrives. Capture bottlenecks still occur with tethered shooting however. We expect that this will improve over time with the introduction of faster connection methods such as USB 3 and high-speed Ethernet/wireless and faster computer processing speeds.

Tethered

An important parameter is whether the shoot will be done with the camera tethered to a computer (or external monitor). While one advantage of digital is that images are instantly available as thumbnails on the back of the camera, these instant “digital Polaroids” can also be transferred to external monitors or as processed images by using software that imports the images from the camera to the computer. Many types of photo shoot are collaborative. A team approach may include art directors, graphic designers and clients. Many times it is an advantage for a stylist or make-up artist to collaborate during the shoot as well. Even the talent (or portrait subjects) may be easier to direct if they can see the images as they are captured.

READ MORE IN TETHERED CAPTURE

FIGURE 4 When the model can see images as they are shot, capturing specific actions such as this hair-swirl is considerably easier.

Shooting to card

Shooting to memory card is still the leading capture method for most photographers and most shoot situations. The key elements are keeping track of the state of your cards – in terms of shot, unshot and formatted – plus having a system for downloading, verifying and backing up the shoot.

Card handling

Thought they might seem obvious, these are some good habits to get into in preparation for any shoot: 

• Have a plan for keeping shot cards separate from un-shot cards. 
• Make sure that the un-shot cards are formatted and ready to go.
• Establish a numbering or coding system for your cards. If you find corrupted files, this will make it easier to track down which card may be creating the problem. It also just helps you keep track of your cards.

FIGURE 5 A useful way to keep track of shot and unshot cards is to use a ThinkTank card wallet. We turn the shot cards over, leaving the unshot cards with the manufacturer’s label facing up. This method coupled with a card numbering system helps us stay organized in the field. A cardholder also helps minimize any dust and dirt from touching the cards.

Camera set-up

Tone curves and picture “looks”

Once again, these settings are mostly important to JPEG shooters. Since JPEGs are raw files processed in the camera with the raw data then discarded, you are essentially applying a “preset” to your pictures. Camera makers have names such as: “standard”, “portrait”, “landscape”, “neutral”, and “faithful” for these presets. These are essentially a combination of tone curves and saturation settings applied (baked-in) to the JPEGs. To alter these “looks” later is problematic, just as is adjusting white balance. Raw shooters may apply such “looks” (and many others) in post-production. Furthermore, looks are non-destructive for raw files and may be changed or kept as “versions.”

Color space

Color space is another setting that applies only to in-camera JPEGs and does not affect raw files. Color space choices are usually sRGB, a narrow gamut space for use on the Web, and Adobe RGB (1998), a larger gamut space often used for prints or printing. However, not even the wider gamut Adobe RGB (1998) color space comes close to the potential color gamut of the digital camera’s sensor. Once again, raw capture allows for selecting a color space in post-processing. By selecting a very large color space such as ProPhoto RGB, for instance, much more of the camera’s potential color gamut can be realized (depending of course on the actual color gamut of the scene).

In-camera sharpening

As discussed, Bayer pattern sensors, especially those equipped with low-pass filters, give a slightly blurry or soft image unless an algorithm is applied to define and sharpen the edges of the pixels that make up the image. Most DSLR cameras give us control over a sharpening setting that is applied to JPEG files. When shooting raw, this setting is applied only to the JPEG preview (the image that you see on the back of the camera), letting you control the sharpening of the final image in post-production.

Figure 6 JPEG sharpened in-camera

Figure 7 JPEG with no in-camera sharpening

Most people prefer to apply at least a small amount of sharpening, if only to see a sharper picture on the camera’s LCD. When shooting JPEGs, this sharpening will be baked in, and since sharpening is inherently destructive, over-sharpening in the camera can degrade image quality, especially the ability to enlarge (up-res) the files through interpolation. Our recommendation is to leave set in-camera sharpening to 0 unless there is no opportunity for post-production sharpening.

When shooting raw files, a high amount of sharpening can be applied so that the images on the camera’s LCD, and the images downloaded and viewed in browsers that use the preview JPEG, appear sharper. Since the raw file sharpening will be handled in post-production, there is no danger of unnecessary file degradation due to sharpening settings.

Exposure and metering

Correct exposure for digital capture is complicated by several factors, including whether you are shooting raw or JPEG, (or raw + JPEG), and whether your camera has an exposure bias. Determining exposure bias is a required first step on which all other exposure decisions are based.

Exposure bias occurs two ways: the in-camera meter can have a bias, and the sensor response, relative to the ISO setting, can have a bias. To tell which might be occurring, you’ll need to photograph an 18% gray card with a continuous light source using the camera’s meter. Set the camera to raw + JPEG and shoot 1/3 stop brackets from -1 to +1. Next, use a trusted incident light meter to determine the correct exposure and shoot the same brackets. If the in-camera JPEG looks correct for the incident-metered “0” exposure, but the camera-metered “0” exposure looks lighter or darker, then there is an in-camera meter bias. If the histogram of the incident meter “0” setting is not exactly in the middle, then the sensor is not calibrated to the correct ISO.

If the camera metering results in JPEG files that are overexposed or underexposed, as judged by the histogram, then the camera meter should be adjusted until the histogram of the gray card is exactly centered. If the camera sensor is biased in relation to the incident meter, then the incident meter will need to be adjusted to compensate for the sensor bias in relation to the ISO setting. In our camera testing, we noticed that different manufacturers have different philosophies toward matching the metering and sensors response to ISO values. Figuring out what your camera actually does in this regard is crucial to achieving optimal exposure. Achieving optimal exposure in digital capture is crucial in achieving the best possible image quality.

Once you have determined whether you need to apply exposure compensation, the next step is to determine if the raw processor(s) used have a bias with regard to your camera’s files. Different manufacturers define raw saturation (basically when the sensor reads “full”) somewhat differently. So the middle-gray exposure point can vary by as much as 1 stop. 

One instance of this effect can be seen in files from Phase One cameras when opened in Adobe Camera Raw where they appear to be one stop underexposed as compared to the same files opened in Phase One’s own software, Capture One. Despite how they appear, the Phase One files are not underexposed. The fix is to make a preset in Adobe Camera Raw that increases the exposure value for Phase files by one stop. To see if you need to create an exposure offset in your processor of choice, test your camera’s files by comparing the in-camera JPEGs to the default exposure settings in the raw processor.

Exposing to the right

Figure 8 Normal vs. ETTR histograms

Exposing to the right is not a foreign concept if you have had any experience shooting negative film. Negative film gives better results when it is exposed for the shadows instead of the highlights, which is the opposite of what works best for color transparency film. With regards to exposure, raw files behave more like negative film. Exposing to the right, just below raw saturation (highlight clipping point), will reduce noise and posterization — especially in the shadow areas. Once you have determined the camera’s real ISO values, additional testing will determine what the optimal amount of overexposure compensation might be.

 One caveat with this technique is that the JPEG previews on the back of the camera will look too bright, and raw files loaded into a browser or cataloging program will also look too bright. If the files are loaded into a browser/processor, i.e. (Adobe Bridge, Adobe Lightroom, Aperture or Capture One) that allows you to apply a preset that adjusts the exposure, then the files will look correct and you can achieve the highest image quality. For more information about built-in camera meters and digital exposure for maximum quality of raw files, see Luminous Landscape [http://www.luminous-landscape.com/tutorials/optimizing_exposure.shtml]

Exposing to the right is not good for JPEG shooters

Early digital cameras had a tendency to lose highlight detail or experience blooming when overexposed so the prevailing recommendation was to underexpose slightly. Newer cameras have mechanisms to drain off light overflow, preventing blooming; however, overexposing when shooting JPEG files still results in loss of highlight detail due to the film-like tone curves that the in-camera processing applies before it discards the raw data. 

Matching metering to the subject

Although we don’t want to delve too deeply into what is basically camera technique, it is probably good to mention — since digital exposure is critical to image quality — certain things should be kept in mind when determining exposure. Often times, matching the metering area to the point of focus (especially in spot metering mode) is a good idea. It is especially important for photographing anything on stage, as the subject is often in a bright spot of light while the background is dark. 

Keep the exposure constant for panoramas and stitching: If your goal is a panoramic view involving stitching multiple exposures, you will greatly simplify your work by picking an exposure that is a compromise between the dark and light side of the panorama, making exposures the same for all the frames to be stitched. This is especially important if there is much sky involved since matching the density of a stitched skyline can be challenging if you let the camera autoexpose.

Metering for HDR

High Dynamic Range stitching requires an opposite technique from panorama stitching. For the best HDR results, it is necessary to meter the very darkest area of the scene (either with a spot or an incident meter), and then meter the brightest area of the scene. Sometimes the difference will be nine or ten stops (or more). To get the best possible HDR rendering, you will need to bracket that entire range. It is important to remember that the bracketing needs to be done by varying the shutter speed and not the f/stop so as not to introduce depth of field or focus changes throughout the bracket.

Trip wire: If using raw files for HDR blending, do NOT have the raw processor set to automatically adjust the exposure settings!

White balance settings in the camera

As we’ve pointed out, in-camera white-balance settings are crucial with JPEG capture. White balance is also crucial for raw capture, but best achieved through the inclusion of a spectrally neutral gray patch in a reference frame. Raw software lets the user click on the gray reference patch to achieve accurate white balance. If this procedure is followed, the white-balance setting is less important when shooting raw files.

 Digital cameras have auto white-balance settings as well as a short list of likely white-balance conditions, such as daylight, cloudy, shade, fluorescent, and tungsten. Many cameras have a defined color temperature in a range from around 2,500K to 10,000K. More sophisticated cameras can even bracket color temperature by going more blue/amber or more green/magenta on either side of a chosen color temperature. 

Photographers who want to get an accurate color temperature in the camera, however, use the custom color temperature option. This requires measuring the color temperature of the scene lighting, using either a white card or a gray card, shooting a test shot, and then letting the camera determine the white balance that will neutralize the white or gray card. The color temperature can be fine-tuned by shooting a warmer or cooler card then using that setting as the basis for the custom color calibration. 

Other methods involve using a diffusion filter over the lens instead of shooting a card. Although the auto white-balance settings have gotten better, it is not technically possible for cameras to always give you an optimal white balance. If you shoot JPEGs, you must rely on auto white-balance, a color temperature meter, or constant creation and use of custom white balance settings. Failure to get good white balance in-camera when shooting JPEGs will slow down your workflow and will inevitably result in lower image quality. The only instance where white balance does not require extra work when shooting JPEGs is with controlled, unchanging lighting, such as in a studio set up.

White balance

Although a prime advantage of raw capture is the ability to fine-tune white balance in PIEware, there are times when it makes smart workflow sense to create and use a custom in-camera white balance. One example is a portrait series where a person in a red shirt is followed by one in a green shirt; the white balance will change and will affect skin tone. The goal is to keep the tone neutral and not have to adjust each set of images. 

The solution is to generate a custom in-camera white balance for the lighting set-up using a white card as the target. Now, we can do 20 portraits and get the same consistent skin tone or background color. If a color temperature tweak in PIEware is needed, one global shift is all that is necessary. Using this technique will usually position the frames within +/- 50K of perfect white balance and tint.

We always use this technique for portrait shoots – whether lit or available light – since it simplifies optimization, even for raw files. It is also great for shooting under fluorescent light since auto white balance can shift a lot and, again, it's easier to make one global correction later during optimization.

Include a gray card or color checker when shooting raw

Figure 9 Even when a custom white balance is set in the camera, it is a good plan to include a gray card or a ColorChecker chart in at least one frame of a scene. When the scene or the lighting changes, a new reference shot should be taken. The gray card or the second gray patch on the ColorChecker will provide a spectrally neutral white balance tool that will be used in the raw processor to achieve an accurate white balance starting point. Without this tool, achieving white balance in the raw processor will involve guesswork.