Important Concepts for Beginning Digital SLR Photographers… by Charlie Davis Pixels and Photosites Q: What is a "pixel"? A: "Pixel" is a short word that means "picture element", thus it's the small colored pieces that form digital images. Q: What is a "photosite"? A: A "photosite" is a part of a sensor that is sensitive to a single "color" of light. That color is usually red, green, or blue, but other colors can and have been used. Q: Is a "pixel" the same as a "photosite"? A: No. Sensors don't have "pixels"…they have photosites. A computer algorithm is required to turn photosite data from a sensor into pixel data (so that it can be viewed and printed). Q: Doesn't a Foveon sensor have pixels? A: No. It also has photosites. The difference between a Foveon sensor and a Bayer sensor is that the Foveon sensor has the photosites stacked vertically (Z-axis), whereas the Bayer sensor has the photosites in a 2-D array (X-axis and Y-axis). If you don't know what a Foveon sensor is, just ignore this question and answer…it's not important. Q: How many pixels do I need? A: No way to answer that question w/o more information. In general, most camera manufacturers push the MP boundary too far…ie, most cameras have more pixels than are needed and the IQ suffers a bit because of that. The problem is that there are more ignorant camera buyers than educated buyers, so manufacturers will use any technique that results in higher sales/profits. How'd we get in this pickle? Well, we started building digital cameras when high pixel counts were not possible, thus the first digital cameras didn't have enough. For a few years all new cameras had more pixels. It became an easy selling point. Manufacturers printed their camera boxes with big red letters that said things like, "Now with 1.3 Mega-pixels!" Soon they shortened this to things like, "1.8 MP!" so the letters could be BIGGER! After a few years, consumers became so used to seeing these headlines, they assumed that bigger was always better. However, the manufacturers never said that more pixels equaled better pictures and indeed it's not true. The truth is that we've gone beyond what's reasonable. There are upsides and downsides to having more pixels. At the moment (2009), the downsides are dominating. As a general rule today, get the fewest pixels you can! Understand that there are LOTS of folks who can't or won't understand the truth to the above statement. Just ignore them. The truth hurts their ego (they took strong stands the other way and can't change positions w/o losing face). Resolution Q: What is the relationship between resolution and Image Quality (IQ)? A: It takes a BIG increase in resolution (linear) to make a small increase in IQ. After some number of pixels, more pixels don't make much difference. Ignore resolution if the camera has at least 6 MP or at least understand that resolution is a small factor above 6 MP. Q: How many Mega-pixels are required to make a _____ sized print? A: Not as many as you'd think. For example most computer screens have less than 1 MP…the highest resolution ones only around 2 MP. Modern LCD monitors have between 60 and 100 DPI. Ink jet printers are limited to about 600 DPI (regardless of what they may say), thus a 6.7" x 10" print would have (at most) 24 MP. But, the difference in IQ between 150 DPI and 600 DPI is tiny, so a 1.5 MP image (at 150 DPI) will make a really nice 6.7" x 10" image at 150 DPI. BTW, I selected 6.7" x 10" because it fits nicely on common 8.5" x 11" paper and has a 3:2 aspect ratio. Noise Q: What is noise? A: Noise is random data. It can be seen in analog TV pictures as "snow" when the receiver is far away from the transmitter (or the antenna is small and ineffective). It can be heard in audio recordings as "hiss". Noise in digital pix is similar to "grain" in film images. Q: What causes noise? A: There are several causes. Operating any electrical device causes noise. There are several different types (read, thermal, Schott, etc). These random signals add to the signal from the photosites and can be seen as tiny splotches on pictures. Q: How can noise be reduced? A: There are several ways to reduce noise. First, don't amplify the signal from the photosites. Gain increases the signal and the noise…plus, the amplifier adds some noise of its own. Second, cool the sensor and the electronics (or use semiconductor designs that stay cool). Hot sensors and electronics generate more thermal noise. Third, slow down the read rate. Slower clock frequencies don't radiate as much or as far. Use an analog-to-digital converter (ADC) with more resolution. Finally, noise can be reduced by using computer algorithms that recognize noise patterns and blur them, while leaving image details relatively intact. Q: Don't noise reduction algorithms always blur the image and reduce the IQ? A: Yes. But they also hide noise which increases the IQ. The trick is to balance the NR technique and amount so that the reduced noise improves the IQ more than the blurring reduces it. One technique is to restrict NR to areas of relatively low detail and areas of low brightness (shadows). This leaves the brighter, more detailed parts non-blurred. As cameras get more processing power, they can also use more complex NR algorithms. Alternately, by using NR programs during post processing, powerful techniques can be used that result in excellent NR while retaining detail. Exposure Q: What controls the exposure on a camera? A: There are four things that affect the exposure. Classically, people say that only the aperture area and the exposure time set the "exposure", but this is a limited, legacy view. In addition to these two, the illumination of the subject and the sensitivity of the detector (film or sensor) also affect the "exposure". This formula shows this simple relationship: Exposure = Light * Time * Sensitivity Note that "Light" in the above simple equation is a function of both the illumination and the aperture area. Expanding the formula: Exposure = Illumination Intensity * Aperture Area * Exposure Time * Detector Sensitivity If any one of these is increased, the product of the others must be decreased to maintain the same "exposure". Q: What is the relationship between Aperture and F-stop? A: First, "F-stop" implies one of the numbers in the geometric series, 1, 1.4, 2, 2.8, 4, 5.6...(2^.5)^(N-1), where N = term sequence. The more general "F-number" can also be values in-between the "stops"…apertures like F/1.9 and F/3.3. Note that every other number is actually more complex than shown…they have been rounded off in non-traditional manners. For example, F/22 is actually F/22.627, so F/23 would be a correctrounded result (somebody decided to make it F/22 so that it would be twice F/11). The formula: F-number = Focal Length / Aperture defines the relationship, where Aperture is the Effective Diameter of the lens. Q: What does Shutter Speed mean? A: It's really Exposure Time (not Shutter Speed) and is the interval (time) during which the sensor sees the image. Q: What does ISO mean? A: It's really Sensitivity (but commonly called "ISO"). It's sorta the gain setting. All cameras have a "native sensitivity" where zero gain is required. This is usually some ISO Sensitivity value between 50 and 200. At this "native setting", the sensor gives the highest possible IQ, because no amplifier gain is required. As you increase the gain, additional noise starts appearing. Some cameras compensate for this higher noise by turning on NR as you select higher ISO settings. This reduces noise but can smear the image. You have to decide whether the noise or the "plastic" look is worse. Some cameras allow you to turn off NR. JPEG vs TIFF vs RAW Q: What is JPEG? A: JPEG are a series of file formats developed by the Joint Photographic Experts Group. Often, the file extension of JPEG files is shortened to ".JPG". More information is here: http://en.wikipedia.org/wiki/JPEG. Q: What is TIFF? A: TIFF stands for "Tagged Image File Format". Often, the file extension of TIFF files is shortened to ".TIF". More information is here: http://en.wikipedia.org/wiki/Tagged_Image_File_Format. Q: What is RAW? A: A RAW image file is one that has not been converted to pixels…ie, it is still photosite data. Each camera maker has a different way of doing this and a proprietary RAW format. Some of the file extensions for these RAW formats are: Nikon = NEF, NRW Canon = CRW, CR2 Olympus = ORF Sony = ARW, SRF, SR2 Pentax = PTX, PEF Kodak = DCR, DRF Sigma = X3F Fuji = RAF Panasonic = RAW Q: I hear many people say that RAW is much better than JPEG. Is that true? A: No. They are mostly just quite different. One may be better for some photographers and worse for others! Q: How is RAW better? A: It retains the full range of color brightness values as detected by the camera electronics. Some cameras record 16-bit values for every color. Others record 14-bit, 12-bit, or 10-bit data. A few cheap cameras only record 8-bit data, but most of these don't support any RAW format (because there is little benefit). The normal JPEG format only supports compressed, 8-bit color data, so any RAW format gives you more latitude to correct images w/o creating visible artifacts…the cameras with 16-bit RAW capability are the ones that give the biggest benefit to using a RAW format (but there are not many cameras with 16-bit capability). Q: How is RAW worse? A: It has LOTS of data, thus the file is much larger. Thus, it takes more time to write a RAW file, so the camera burst rate is normally lower when in RAW mode. Finally, YOU have to edit EVERY RAW picture. With JPEG, the camera has some "rules" and normally can get the picture "OK". To go beyond "OK", you have to spend time learning how to make better decisions with a RAW editor on your PC. Q: How is JPEG better? A: Being a compressed format, it's MUCH smaller. That means you can save more images to a card and it takes the camera less time to write that data (ie, faster camera operation). It's also widely supported. All web browsers can display a JPEG picture…none can display any RAW format. In order to "use" a picture, at some point you will have to render it as a JPEG. Q: How is JPEG worse? A: JPEG throws away data to get the file smaller. It does this very intelligently, so that it can be undetected. But, worse case, a JPEG picture can have small "artifacts"…these are most commonly seen in the sky and the horizon (where the sky meets the ground). Q: Can I see the difference between various JPEG formats? A: Only if their file size is drastically different. A small JPEG size implies that the algorithm that encoded it was compressing the data heavily and some JPEG artifacts will be visible. As long as you use the highest JPEG format the camera supports, it will be "OK". Q: How do I know which format to use? A: If you are a beginner, use JPEG. Each format has its uses and there are MANY variations. For example, both TIFF and some RAW formats can be compressed...and you have a choice of how it's compressed. There are lossy methods and non- lossy methods. For long-term storage, many experts start with a RAW file, then use TIFF, compressed with the LZW algorithm (non-lossy). Others, say start with RAW and use Adobe's DNG format. But for a beginner, use JPEG, Fine. Aspect Ratio Q: What is meant by aspect ratio? A: The aspect ratio is the width divided by the height. For example, a 4" x 6" print has an aspect ratio of 1.500:1. Often the ":1" is omitted, thus written as just "1.500". It's also common to state the aspect ratio as the smallest whole-number ratio… thus 3:2 (most people wouldn't say the aspect ratio is 6:4). And convention has us state these ratios so that they are greater than unity (1). Q: How many different aspect ratios are there? A: There are LOTS of different aspect ratios! For digital cameras, there are two common ones…3:2 and 4:3. A relatively new one is 16:9. Each sensor has a "native" aspect ratio. Many cameras have options to select the aspect ratio…they do this by not using some of the photosites along two edges. In other words, if a camera with a native 4:3 aspect ratio switches to a 3:2 aspect ratio, small strips of pixels along both sides are not used. This reduces the resolution by about 11%. For paper, the range of aspect ratios is much wider and the ratios are often different than ANY digital camera produces (yes, that causes problems). For example, the common 8" x 10" and 4" x 5" print sizes have a 1.25 aspect ratio that is not supported by any digital camera…and the common 8.5" x 11" paper size has a 1.293 aspect ratio that is very strange. Finally, it's often difficult to find a picture frame that matches digital picture aspect ratios and common paper sizes. It's like the printer manufacturers, camera manufacturers, paper manufacturers, and frame manufacturers don't talk to each other! Dust Q: What is "dust"? A: Dust is a collective term for any small airborne particles. Dust comes from many sources: Animal skin cells, plant pollen, human and animal hairs, cloth fibers, paper fibers, minerals from outdoor soil, and other materials. Dust particles are generally thought to be less than 500 microns in diameter. Most dust that finds its way into cameras is much smaller than 500 microns. 2 to 50 microns seems the most likely to be seen. Q: Why is "dust" a problem for photographers? A: Dust is a problem for several reasons. If dust is allowed in and on cameras and lenses and is wiped off, it can scratch precision surfaces (much of it is hard and abrasive). If dust accumulates on bearings of moving parts (such as mirrors) it can cause wear. If dust particles settle on surfaces close to the sensor, they can cast a shadow that is visible on pictures. Q: If I see a dark spot on my pictures, do I have a dust problem? A: Probably. Dust particles cast a shadow that is most visible when small apertures (large F/numbers) are used. You can test for dust particles in the sensor light path by taking a picture of something uniform in color and brightness (a blue sky w/o clouds is a good subject) with a VERY small aperture (F/22 or F/32)…be sure to expose it correctly (it must look blue, not black or white). If you see dark spots on the picture, these are shadows of dust particles. These dust particles are located on the anti-aliasing filter, which is in front of the sensor…thus they are blurred. The distance between the sensor and the AA filter varies. The larger the spacing, the less well defined the shadow will be. Note that dust on the mirror is NOT visible, either in the viewfinder or in pictures. Note that dust on lens surfaces is also not visible to the sensor or through the VF. Q: I see spots in my viewfinder. What causes them? A: If you have a dSLR, the most likely cause is dust on the bottom surface of the viewing screen. The viewing screen is a delicate plastic part, located above the mirror. DO NOT TOUCH THE VIEWING SCREEN WITH YOUR FINGERS! Most often the dust can be blown off or dislodged with a very fine/soft brush. The plastic screen is easily damaged. It is easily replaced and is usually not expensive, but if you are careful, you won't have to do this. Q: How can I clean the dust off my sensor? A: First, understand that the dust is NOT on your sensor, but rather on the AA filter located a short distance in front of it. This is an important distinction because the sensor is quite delicate, but the AA filter is not. Most AA filters are reasonably rugged. Some AA filters have a conductive coating on the front surface to dissipate static electric charges (which can attract dust particles). This coating, composed of 90% Indium Oxide and 10% Tin Oxide, in addition to being conductive, is quite hard. In cameras that have a "sensor" cleaning system, the truth is that piezo elements shake the AA filter to dislodge dust particles. The sensor is not shaken (much). When dust can't be removed by the ultrasonic shakers, you will have to clean the AA filter (or have it cleaned). In spite of the Indium-Tin Oxide film on the AA filter being difficult to damage, you should clean it carefully. I recommend that you follow a 4-step cleaning process. Start with the least invasive method and in steps move to more invasive methods until all dust spots are cleaned off. As described above, take a picture of the blue, cloudless sky with your smallest aperture and look for dust shadows. Do this after every cleaning attempt (below). Step 1: Using a Rocket Blower, blow out the mirror box. Point the air in several directions. Be careful to NOT     insert the tip inside the mirror box where it could hit the mirror or viewing screen! Then, activate the     "sensor cleaning mode"…most all dSLR cameras will have this…and blow air bursts toward the sensor.     Again, be careful to not bump the Rocket Blower tip into anything…you can insert it part way into the     mirror box while the mirror is up. Step 2: Activate the "sensor cleaning mode" in the camera and using a VERY fine brush, wipe the face of the AA     filter. Be sure to wipe ALL of it. Then blow it off again with the Rocket Blower. Step 3: Use a "low-tack" cleaning pad system like "Dust-Aid" (www.dust-aid.com). Step 4: Use a pad wet with "Eclipse" fluid (www.photosol.com). If your AA filter has an Indium-Tin Oxide coating,     use "E2" fluid. There are many good sites that describe in detail exactly HOW to perform wet cleaning. Google for it. Note that if Step 1 removes all your dust shadows, then STOP…it's not necessary to go farther…ditto with all the other steps, stop when the dust is gone. Q: What is "ISO"? A: As I said in the Exposure section, "ISO" when used in a digital photography context, refers to the "gain" setting on a camera. Back in the film era, we had different constructions of the light sensitive layer in the film that varied the sensitivity to light...there were also different development chemistries that "pushed" the sensitivities higher...the relevant 3-letter acronym was also different; "ASA". Both "ISO" and "ASA" are industry standards organizations who published procedures for measuring the sensitivity of image sensors (and film). In general, ASA and ISO sensitivities are equivalent…ie, an ASA 200 film requires the same exposure settings as an ISO 200 digital sensor. The effect of raising the "ISO" setting is to make the camera more sensitive to light. Cameras do this by amplifying the signals from the sensor. As the "gain" is increased, both the signal (the picture) and the noise are increased in amplitude. Cameras with "good" sensors don't have as much noise as those with "bad" sensors, thus they can be amplified more before the noise gets too bad. Put another way, "good" sensors have a high signal-to-noise ratio. Also, camera manufacturers increasingly use noise reduction algorithms (NR) to mask the noise. These NR algorithms work by blurring areas judged to be mostly noise and not signal (subject). Each camera brand has different NR algorithms. These are constantly improving, but often cameras have too much NR and the pictures look like pictures of "plastic". Whether you need good ISO 800 or 1600 or 3200 depends on what type photography you are doing. If you don't have difficulty taking good pix at the lowest ISO setting, then you don't need a higher setting! If, in contrast, base ISO results in too slow exposure times (blurry pix due to motion) and/or too large apertures (limited DOF and marginal sharpness), you need to raise the sensitivity/ISO setting. How much you need to raise it depends. Typically, a venue that needs a fast shutter and high ISO is indoor sports...something like a basketball game. For this application, you can't have too much sensitivity! Auto Focus Q: How do I properly take a shot? A: ALL digital cameras have a special shutter button that is really two switches stacked, one on top of the other. The top switch, when pressed, tells the camera to auto focus, set the exposure, and set the white balance (if these three functions are selected to happen automatically). The bottom switch, when pressed, tells the camera to actually take the picture. You can't see these two switches…it looks like ONE button…but trust me, there are two switches in there! Usually there is tactile feel to indicate when the top button has been pressed, but on some cameras this is VERY indistinct and causes them to be difficult to use. The process of taking most pictures involves three steps: 1) Point the camera at the subject (what you want to be focused and exposed properly). Depress the shutter button half-way and hold it there. This will focus, expose, and set the white balance. Most cameras have a visible "focus area" indicator in the viewfinder or LCD panel on the back…some allow you to move the "focus area" around. Some cameras have several "focus areas" and select the one to use with complex algorithms. Learn what your camera does regarding "focus areas"! It's IMPORTANT that you communicate accurately what you want the camera to consider the subject! There will be a delay for the camera to focus/expose/set WB… better cameras do this more quickly. 2) Move the camera to frame the scene as you desire. 3) Press the shutter button fully. This will take the picture. There will be a delay between pressing the shutter button fully and the picture being captured…this delay varies between cameras. Unlike the delay following the half-press of the shutter button, the full-press delay is often greater on higher priced (better) cameras, because they have a mirror that must be flipped out of the way before the picture can be captured. Some high-end cameras allow the photographer to use another button to perform AF, as an option. This button is normally located under the right thumb. Jargon Q: I read that LOTS of camera thingies are described as "fast". What does that mean? A: Yes. There are "fast" lenses, "fast" shutter speeds, and "fast" films. These terms are very old and probably outdated in the 21st century. I think it started with photographers describing a short exposure time as a "fast shutter speed". Then, a lens with a large aperture (and thus better at collecting light) was termed "fast" because it allowed a "fast shutter speed". Then a more sensitive film that allowed a "fast shutter speed" was called a "fast film". Note that these things are really NOT fast, as in having or allowing a high velocity, but rather fast, as in having or allowing a short exposure time. Q: What does "clipping" mean? A: In the digital realm everything is represented by numbers and the number range is fixed. As an example, most digital pictures are represented by an 8-bit x 3 color scheme. This means that each color can have values between 0 and 28 (255). By convention, small numbers represent dark and big numbers represent bright. These "hard" limits to brightness values cause "clipping". When something bright (like the sky) is exposed such that some (or all) of it is equal to 255 in all 3 color channels, then it looks like absolute white. When this happens, people call that condition a "blown-out" sky...ie, you can't see any detail in it. Even reducing the brightness in post processing just converts all the 255 values to lower values (like 250 or 200)...thus, instead of the sky looking like a uniform white color, it looks like a uniform light-grey color…and in doing this, the dynamic range (DR) is reduced. The same effect happens at the other end of the brightness range. When the exposure is adjusted such that much of the shadows = 0 (zero), there is no detail available in those shadows. The subjective judgment of most people is that it's worse to clip the highlights than clip the shadows, thus they recommend that photographers try to get the highlights exposed correctly and let the shadows fall where they will. This is verbalized by the oft-heard mantra, "expose to the right"...referring to the histogram where by convention, the bright parts are on the right side of the graph. A scene with a limited DR can be exposed so that there are no brightness values at either end of the scale. These type scenes are easy to photograph...just adjust the histogram so that everything is centered. However, many scenes have a DR that exceeds the default ability of the camera to capture it. In these cases, the histogram extends across the scale of 0 to 255 and significant parts "bump up" against both ends. These type pix require the photographer to make a decision. There are several things that can be done: 1) Adjust the exposure parameters (aperture, exposure time, sensitivity, light intensity) to "expose to the right" and let the shadows "clip". 2) Adjust the exposure settings to "expose to the left" and let the highlights "clip". 3) Add light to the shadows (fill light) and "expose to the right". 4) Reduce the camera's DR by lowering the contrast setting or using an advanced feature to compress the DR (called something like "D-lighting")…not all cameras will have these features. 5) Take multiple shots with different exposure parameters (usually exposure time) and process them with HDR software to create a high dynamic range picture...THEN, map the expanded range into an 8-bit version that shows BOTH the shadows and the highlights properly...this compresses the brightness range. 6) Use a graduated ND filter to reduce the exposure for the highlights (normally most useful for dimming a bright sky). 7) Use RAW mode, which by convention has more "headroom" and "footroom" thus allows post processing to recover some highlights and shadows...this only works for scenes that have moderate DR excesses. Q: What is a "stop"? A: On old cameras, the dials that set the exposure variables (exposure time and aperture) were designed to detent when the exposure variable was double or one-half the original value. These picked up the name "stop", because the dials would stop there. More modern cameras are not set via "dials" as much anymore…even if they do use dials, the dials are electronic… and they allow you to specify if one "click" equals one "stop", ½ "stop" or 1/3" stop. Some old people still use expressions like "stop down", which means to reduce the diameter of the aperture in the lens (but for some reason they never say "stop up" to mean the opposite, especially to old people). ;-)