Night Photography Noise
Night photography involves low light and long exposures which present special problems for digital photography.
Digital cameras suffer from random noise resembling snow on a television. A high ISO setting results in a high noise image.
Digital cameras may have some defective pixels. Normally the camera software works around these issues and you don't even notice. At night with low light and long exposures, the camera software seems to be unable to differentiate between noise and image data does not remove this.
There are a number of techniques you can use to reduce this in the final image. First shoot with the lowest ISO permitted by the subject. The random noise that does remain can be greatly reduced by software. Photoshop seems to be the gold standard in photo processing software but I'm broke so instead use Paintshop Pro which is less expensive. Paintshop Pro has a smart noise reducer does an excellent job of removing this type of noise usually without sacrificing image detail.
To get rid of the stuck pixel dots there are a couple of techniques you can use. They rely on taking a flaw picture to capture this fixed noise without the image. Just put the lens cap back on and shot for the same duration.
One technique detailed in the Digital Photography Review, involves using the reference frame to select the points on the real frame (using the magic wand selection tool) and then use the scratch and ding filter to remove them (uses surrounding pixels to replace them).
Another method is to invert the reference frame, add to the original, then tweak the resulting histogram back to where you want it. I've played with this a bit but not really got it to where I am happy. I think it's because the stuck pixels tend to be "ON" not at some intermediate value so underlying image data is lost.
With my camera there usually aren't more than two or three noticeable goofed pixels so I usually just use the clone brush to clone in undamaged data over them.
An annoying problem I haven't found a good solution to is bleed, on films a brightly exposed area may bleed into the area around it. This varies greatly from film to film. I've also experienced bleed with digital, and because of high contrast at night it is hard to avoid.
There are some trade-offs, for example, I've found I can reduce bleed by underexposing slightly and then stretching the resulting data. This reduces bleed but at the expense of noise.
One more problem I've run into, "stars" resulting from point sources, Point sources of light get radial streaks leading away from them. The iris on my camera has six leaves, point light sources get six radial streaks.
Normally night photography requires a tripod. The Canon Powershot S2 IS includes image stabilization and I have found this actually makes hand-held night photography doable in situations where an exposure of one second or less will suffice and there isn't wind or other factors causing involuntary movement.
Digital cameras suffer from random noise resembling snow on a television. A high ISO setting results in a high noise image.
Digital cameras may have some defective pixels. Normally the camera software works around these issues and you don't even notice. At night with low light and long exposures, the camera software seems to be unable to differentiate between noise and image data does not remove this.
There are a number of techniques you can use to reduce this in the final image. First shoot with the lowest ISO permitted by the subject. The random noise that does remain can be greatly reduced by software. Photoshop seems to be the gold standard in photo processing software but I'm broke so instead use Paintshop Pro which is less expensive. Paintshop Pro has a smart noise reducer does an excellent job of removing this type of noise usually without sacrificing image detail.
To get rid of the stuck pixel dots there are a couple of techniques you can use. They rely on taking a flaw picture to capture this fixed noise without the image. Just put the lens cap back on and shot for the same duration.
One technique detailed in the Digital Photography Review, involves using the reference frame to select the points on the real frame (using the magic wand selection tool) and then use the scratch and ding filter to remove them (uses surrounding pixels to replace them).
Another method is to invert the reference frame, add to the original, then tweak the resulting histogram back to where you want it. I've played with this a bit but not really got it to where I am happy. I think it's because the stuck pixels tend to be "ON" not at some intermediate value so underlying image data is lost.
With my camera there usually aren't more than two or three noticeable goofed pixels so I usually just use the clone brush to clone in undamaged data over them.
An annoying problem I haven't found a good solution to is bleed, on films a brightly exposed area may bleed into the area around it. This varies greatly from film to film. I've also experienced bleed with digital, and because of high contrast at night it is hard to avoid.
There are some trade-offs, for example, I've found I can reduce bleed by underexposing slightly and then stretching the resulting data. This reduces bleed but at the expense of noise.
One more problem I've run into, "stars" resulting from point sources, Point sources of light get radial streaks leading away from them. The iris on my camera has six leaves, point light sources get six radial streaks.
Normally night photography requires a tripod. The Canon Powershot S2 IS includes image stabilization and I have found this actually makes hand-held night photography doable in situations where an exposure of one second or less will suffice and there isn't wind or other factors causing involuntary movement.
posted by Nanook at 5:42 PM
0 Comments:
Post a Comment
Links to this post:
Create a Link
<< Home