I'm curious to this as well. I've read in the past that lowering the resolution in your camera will not help in noise reduction as the actual number of pixels used doesn't change..

But then again, I've heard it does. Hopefully someone who is certain on the answer will chime in.amateurapertures.blogspot.comMy fully documented blunders in Photography..

Seeing as how noise increases as pixel count increases and sensorsize decreases (as a general rule, I know this is not always thecase), if you picked a lower resolution (say, 5mp) on a 10mp compactsensor, would that reduce the visible noise in the picture? Or is thenoise level on a particular sensor always going to be the same, nomatter what resolution setting you have?.

You'll get less noise per pixel but the pixels will be larger - an exact trade-off in principle, as you've realised. It is not possible to create additional information simply by repackaging what the sensor saw in a different way - the best you can hope for is to break even. A fundamental principle of thermodynamics which applies here too .

If you take the picture at 10MP and then downsample it to 5MP in photoshop later by combing several small pixels into a fewer large ones you'll get the same result..

Best wishesMike..

So what you're saying is that taking a picture with a 10mp sensor set to 5mp will still use all 10mp (and get all the noise), and then just downsample it to 5mp?..

The "noise" issue is straightforward: the high ISO means high amplification of the signal. So, when it is amplified, any imperfections in the signal are amplified too. Making the pixel smaller means making the signal weaker, too, which is what lies behind the links..

As has been mentioned, the way that a camera will reduce it's resolution is similar to that photoshop might do. However image quality depends how good the camera program and the processor is. Often, if you can be bothered, software like Photoshop may do a better job..

Ham===http://londondailyphoto.blogspot.com..

So what you're saying is that taking a picture with a 10mp sensor setto 5mp will still use all 10mp (and get all the noise), and then justdownsample it to 5mp?.

That is my understanding. It would be hard to do it any other way... if you want a 5MP image from a 10MP sensor, which half of the pixels would be ignored? The camera can't just use the left hand half of the sensor, or switch every alternate pixel off... as far as I know . So it takes the normal output from a 10MP sensor and then downsamples it along with all the other messing about that goes into creating a JPEG - doing it in software is easy..

Don't forget that downsampling 10MP to 5MP will tend to average out some of the noise..

I am perfectly willing to be contradicted by someone who knows more than I do!.

Mike..

No, not really. What you're referring to is pixel binning. Pixel binning has not really shown much improvement. In pixel binning, you combine (for example) four pixels (well, 16 if you wanted to break it down to 4 each of the RGB components and combine all of those) but you don't get much performance increase. The theory is if you combine the pixels you end up with one super pixel that averages out the noise, but in reality you just end up with one noisy pixel. That's not to say there's NO benefit to binning, it's just not like you would hope..

What you have to realize is that noise is relative to size also. If you're willing to reduce your image from 10mp to 5mp, I'll assume that you're not going to print this as 48" x 72". Noise reduction software will reduce the sharpness, similar to reducing the resolution..

Ok, say you take a 10mp shot, and use some heavy noise reduction. I'm trying to say that image is similar to a 5mp shot with no noise reduction. The ratio of noise reduction to apparant resolution depends on a lot, these are just arbitrary numbers here..

End result here is shoot with full resolution, and don't feel bad about using heavy noise reduction software. You'll end up with a better image..

Output size means a lot. If you're just going to print to web or 4"x6", crank up the settings. For 16"x20" images, you have to be more careful and use masking, blurring, and stuff to get careful noise reduction..

-Porter..