So, you just bought that thousand-dollar digital camera with full manual control to unlock the artist within. You play with the new features, and take some long exposure night shots and are horrified to see a bright spot in the same locaiton on every picture. No, you haven’t discovered a new star… it’s the infamous hot pixel, and unfortunately, it’s a fact of life.
Don’t panic, and don’t return your new camera to the store. Your CCD is not defective, and understanding what they are will help you cope. Nearly every camera I own has a couple, and the ones that don’t eventually will at some point. They really aren’t a problem unless it’s smack dab in the middle of the photo, and they appear no matter what your ISO rating is.
Your CCD is a grid of elements, each of which is sensitive to brightness (light). A hot pixel is created when one of the elements has a higher rate of charge leakage than its neighbors. During a long exposure, this leakage may cross a threshold of an exposed value. Cheaper digital cameras don’t permit an exposure longer than 1/4 second (or they lSO settings), which effectively eliminates the chance of revealing the hot pixels that probably exist.
All CCD elements leak current, and given a long enough exposure will show hot pixels, even in complete darkness. Temperature is also a factor in creating hot pixels. Generally the higher the temperature, the higher the charge leakage. So that hot pixel you see during those summer landscapes may disappear in the winter.
Another factor is the ISO rating of your CCD. When the ISO is increased, you are essentially turning up the gain, similar to turning up the gain on a microphone. You’ll notice a lot more hot pixels at ISO 1600 than you will at ISO 100, simply because the signal is amplified.
Normally, hot pixles are practically invisible unless you go looking for them every time. You really won’t see them unless you enlarge your photo to 200% or more, and with today’s high mega-pixel cameras you are more likely to shrink the photo down than enlarge it.
What do they look like?
Hot pixels (or sparkles) appear as a single (or few) unnaturally bright individual pixels. They will be exactly at the same pixel-location in every frame. They are locked in on the sensor, and they will be much sharper than anything else on the image. Some cameras show these pixels as tiny little crosses, which is due to Bayer interpolation. Bayer interpolation is also what makes the pixels a high contrast pixel red, green, or blue, and is the direct result of the color-striped mask over the sensor.
The dirty little trick about our “digital” cameras is that the sensor is analog! Sensors collect photon in microscopic wells (called pixels). Sensors assign electric charges to these photons, which are read as analog voltages. These voltages are sampled and quantized to make them into digital values, which then go through more digital processing before the picture is written.
Leakage currents are electric charges which leak into sensor wells, and this excess electric charge increases the voltage in the well and make it look brighter than it should. Manufacturing variations will cause some pixels to have more leakage than others, and there really is no way to make a “perfect” CCD. If there were, then the yield of the chips would be so low that digital cameras would be exhorbadantly expensive.
A pixel is a single point of color data in an image (short for “picture element”). Some pixels may be colored red, and others green or blue, but there are no such things as red or green pixels on the sensors. A pixel is a complete pixel only when the red, green and blue values for that unique location are known.
Digital camera makers all lie about megapixels. This is fine, because all legitimate camera makers lie in exactly the same way, which makes it easy to compare cameras from different makers.
All digital cameras (except for the $30,000 ones that have three seperate CCDs) have only a third of their claimed pixels. Instead of having seperate R, G, and B sensors for each pixel location, they have a single monochrome CCD with each pixel location covered by a color filter array (CFA). The alternative RGB matrix follows the Bayer patterns (50% green, 25% red, and 25% blue, hence called RGBG or GRBG). Bayer interpolatoin in named for Kodak scientist Bryce Bayer who invented this in 1976.
Bryce Bayer’s patent called the green photosensors luminance-sensitive elements and the red and blue ones chrominance-sensitive elements. He used twice as many green elements as red or blue to mimic the human eye’s greater resolving power with green light. These elements are referred to as sensor elements, pixel sensors, or simply pixels; sample values sensed by them, after interpolation, become image pixels.
The raw output of Bayer-filter cameras is referred to as a Bayer pattern image. Since each pixel is filtered to record only one of three colors, two-thirds of the color data is missing from each. To obtain a full-color image, various demosaicing algorithms can be used to interpolate a set of complete red, green, and blue values for each point.
Different algorithms requiring various amounts of computing power result in varying-quality final images. This can be done in-camera, producing a JPEG or TIFF image, or outside the camera using the raw data directly from the sensor.
Remember that before this interpolation that there is physically only a single monochrome pixel on the sensor. The algorithm creates values for each of the three colors at every location by averaging (interpolating) each set of partial R, G, and B vlues to create values at each location.
Different camera manufacturers vary the algorithms, which is why some makers have higher percieved sharpness or color saturation.
Most hot pixels are only visible at high ISOs. If you enough of an enthusiast that you bought an expensive camera, then your “good” stuff will not be shot at ISO 1600 or higher, anyway. They are very tiny, and are really only visible at high magnificaiton. You probably only saw it because the small LCD on the camera pointed it out to you. A typical LCD on a digital camera is probably only 300×200 pixels, so it has to scale down and average the high resolution photo you just took to display it. Since the hot pixel is the brightest thing in that region, it gets displayed on the LCD, so it looks worse than it actually is. Print that picture out and you probably won’t even notice it.
How do I get rid of them?
You can easily clean them up in Photoshop, because they stand out so sharply on a pixel-by-pixel basis. Use a small Spot Healing Brush.
If you absolutely cannot live with the hot pixel, you can send it in to your camera manufacturer for “repair”. They only replace the CCD if there are a certain number of “dead” pixels, not hot pixels. All they do is identify the hot pixel and tell the camera to not use it. The other pixels on the sensor will average and compensate for it. Some consumers might be distressed to know this, because they paid for a 10-megapixel camera and they want all 10 megapixels working. Well, did you know that the manufacturer mapped out defective pixels before it even left the factory? Because of Bayer interpolation, there is not a 1:1 direct correlation of the pixel on the sensor to the pixel in the photo. Digital processing interprets what the sensor sees and puts it together. There is so much information obtained by the sensor that a photo can be generated even if a few pixels are defective.
Recommendations
Every camera has at least one hot pixel, even if you can’t see it right now. Even if your camera is fine today, it might start doing it tomorrow. Don’t worry about them, because if you look hard enough you will always find them. If you return your camera to the store, then the replacement will have this problem, too. Even if you hand-pick a camera, the “clean” one you seelct will eventually have some sparkles. Since sensors are analog, the hot pixels will come and go. They may not move, but their intensity will vary. Stop worrying about it and concentrate on making great pictures.