P&S hot pixels

[lex_jenkins]

This is fairly typical performance for many P&S digicams.

 

From the time it was new my Olympus C-3040Z demonstrated anywhere from a sprinkling to a blizzard of green pixels at slower shutter speeds. Starting at 1/15 second a few show up - nothing that can't be fixed easily in editing. At 1 second and longer it's a green snowstorm.

 

At the time this was considered acceptable performance. Then Canon introduced a couple of models with better noise reduction and that set the standard for P&S performance.

 

Oddly, to this day Olympus digicams, even the E1, tend to have the noisiest sensors beyond ISO 100 or 200. They don't seem terribly concerned about the issue.

 

I dunno. Some "artistes" are winning critical acclaim using cellphone-cams, which produce half noise and half actual photo.

[Spearhead]

Lex, I've seen some great stuff taking with phonecams, check out <a href="http://zonezero.com/moblog/index.html">this site.</a>

[john_harris11]

Peter,

Not sure why folks have a hard time actually answering your question. I'll do my best.

Yes, P&S cameras are rampant with hot pixels. Quite frankly, they come cheaply for a reason. Some of that reason is cheaper chips, some is cheaper firmware, some is cheaper optics. The list goes on. On a daily basis I see image files from many many differing cameras. The P&S boxes are quite simply not going to yeild a file as clean as a quality SLR. You should expect more hot pixels, but shoud not be expected to accept stuck ones. Top notch firmware can catch these anomalies and dither them out.

 

You can obtain software apps such as Noise Ninja that can help with the hot pixel issues, depending on the severity in each file. You could also set up a few actions in Photoshop that would first take the image into LAB space, then apply an amount of either median or dust and scratches to the A and B channels only. You will have to be the judge as to how much. You may need to set up several different actions for differing severities of your issue. Just enough to mask out the color issues may be enough, and you may not need to make corrections to the luminosity (L) channel. The action should then convert the file back to RGB. The flavor of RGB should be determined by your output requirements. Check with your print provider to be sure.

 

You mentioned color build of the one stationary stuck pixel in your orignal post, but I did not say if hot pixles are identical in color build in every image. If so, your issue may be firmware related and not actually a sensor issue at all, as hot pixles are usually a result of image sensor sensitiviy (RMS Dynamic Range), and exposure, combined with the firmware's ability to deal with determining what is actual detail and what is noise, then properly dithering out the noise. I do not get that deep into DC programming, so you may wish to defer that question to a better authority.

 

I hope the above information is helpful.

 

John