24
\$\begingroup\$

I know that high ISOs tend to produce more noise, and some cameras' software can handle that noise better than others, but are there any other settings or conditions tha affect visible noise?

I'm using a micro-four-thirds camera (E-PL1) if it matters.

\$\endgroup\$

8 Answers 8

23
\$\begingroup\$

Keep the camera as cool as possible! High temperature increases the thermal noise in your images. That's why certain astrophotographers actively cool their camera!

\$\endgroup\$
5
  • \$\begingroup\$ That is interesting. I didn't know that. Does it actually make a noticeable difference (for non-astrophotographers)? \$\endgroup\$ Commented Jul 15, 2010 at 21:01
  • 1
    \$\begingroup\$ Exactly why I have a white K-x! And thankfully it looks better in person. Yes, it is significantly cooler than the black DSLR's I've used out in the sun doing stuff like birding or sports. Lens still is hot. \$\endgroup\$
    – eruditass
    Commented Jul 24, 2010 at 18:01
  • \$\begingroup\$ @Justin, yes the noise level is directly proportional to the temperature of the sensor. \$\endgroup\$ Commented Jul 26, 2010 at 18:24
  • \$\begingroup\$ @Justin & Danny: The electrical leakage in sensors (that cause hot pixels) doubles for every 8 deg C increase in temperature (source: books.google.com/…). However you will only notice this when you use very high ISO settings or very long exposure times. \$\endgroup\$
    – Marc
    Commented Jul 28, 2010 at 9:13
  • \$\begingroup\$ Some newer high end cameras are actually using actively cooled sensors now to control noise. Its probably cost prohibitive at the moment, but something to keep an eye out for, as it will probably become more common as it gets cheaper to do. \$\endgroup\$
    – jrista
    Commented Aug 2, 2010 at 23:42
21
\$\begingroup\$

You can reduce noise without lowering ISO by slightly overexposing your picture, especially if you shoot RAW.

From the Expose (to the) Right article at Luminous Landscape:

A 12 bit image is capable of recording 4,096 (2^12) discrete tonal values. One would think that therefore each F/Stop of the 5 stop range would be able to record some 850 (4096 / 5) of these steps. But, alas, this is not the case. The way that it really works is that the first (brightest) stop's worth of data contains 2048 of these steps — fully half of those available.

Why? Because CCD and CMOS chips are linear devices. And, of course, each F/Stop records half of the light of the previous one, and therefore half the remaining data space available. This little table tells the tale.

Tone Level       |  Levels dedicated | Stops of difference
==========================================================
Brightest Tones  |  2048 levels      |  Within first stop
Bright Tones     |  1024 levels      |  Within second stop
Mid Tones        |  512 levels       |  Within third stop
Dark Tones       |  256 levels       |  Within fourth stop
Darkest Tones    |  128 levels       |  Within fifth stop

The simple lesson to be learned from this is to bias your exposures so that the histogram is snugged up to the right, but not to the point that the highlights are blown. This can usually be seen by the flashing alert on most camera review screens. Just back off so that the flashing stops.

Now of course when you look at the RAW file in your favourite RAW processing software, like Camera RAW, the image will likely appear to be too light. That's OK. Just use the available sliders to change the brightness level and contrast so that the data is spread out appropriately and the image looks "right".

Read more.

\$\endgroup\$
7
  • \$\begingroup\$ I think it's the best answer here, but it could certainly explain the ETTR concept a bit better. \$\endgroup\$
    – Karel
    Commented Jul 25, 2010 at 6:52
  • \$\begingroup\$ Overexposing the picture will reduce the noise, but will also increase the motion blur (motion of camera and/or subject). Taking a burst of photos then detecting the motion and averaging them (minus the motion's effects) will give you the same affect, but hopefully minus the blur (if motion detection and reduction is good). \$\endgroup\$ Commented Jul 26, 2010 at 18:27
  • 1
    \$\begingroup\$ Taking a burst of photos and averaging them is one technique to reduce noise. Exposing to the right is a different technique, and each has obviously its advantages and disadvantages. Lowering your ISO also yields a slower shutter speed in the same condiitons. Everything's a balance, and the photographer needs to be aware of all the consequences. \$\endgroup\$ Commented Jul 26, 2010 at 19:29
  • 2
    \$\begingroup\$ I've added quotations of the parts of the article I think are most useful in describing Dave's point. Hopefully that will keep this answer useful even if the linked site ever goes offline. \$\endgroup\$
    – jrista
    Commented Aug 2, 2010 at 23:50
  • 3
    \$\begingroup\$ Overexposing your picture effectively lowers your ISO. It also has all the same effects as lowing the ISO: you also need to use a slower shutter speed (which may increase motion blur) or a wider aperture (which may not be available) for the same scene. Indeed, this is exactly how some cameras are able to offer an ISO lower than their sensor's base ISO: they set it to base ISO but overexpose. If your camera can lower its ISO, you're better off just doing that instead of twiddling with exposure. \$\endgroup\$ Commented Jan 24, 2011 at 5:40
14
\$\begingroup\$

A few options for reducing the noise other than lowering the iso or increasing the light:

  • Keep the camera's sensor cool.
  • Take a burst of photos, then average them.
  • Lower the resolution.
\$\endgroup\$
6
  • \$\begingroup\$ Personally, I think averaging multiple images is an awful technique. \$\endgroup\$ Commented Jul 26, 2010 at 19:29
  • 1
    \$\begingroup\$ If the burst of images is fast enough and you weigh a motion detection algorithm in the averaging the results should be good. If you increase the exposure time then you basically doing the same thing, only the burst it continuous and automatically averaged and you have no way to tell if something moved and compensate. \$\endgroup\$ Commented Jul 27, 2010 at 21:09
  • 5
    \$\begingroup\$ The averaging technique is widely used in the astrophotography field. Photographers take dozens (some even hunderds) of pictures and average them using varying techniques. They often shoot dark frames and flat frames to minimize various sources of noise: ccdastrophotography.com/article.php?id=2. \$\endgroup\$
    – Marc
    Commented Jul 28, 2010 at 9:20
  • \$\begingroup\$ So it's a technique primarily used in a specialized field of photography that requires specialized software. \$\endgroup\$ Commented Aug 2, 2010 at 16:11
  • \$\begingroup\$ @Dave Yes that is correct. Although the technique could be used for any low-light environment with non-moving objects. \$\endgroup\$
    – Marc
    Commented Aug 3, 2010 at 10:26
4
\$\begingroup\$

If you have to work with available light and you are already at a high ISO then the only way is either to expose for a longer period at a lower ISO or use noise reduction software in post-production.

Bottom line, higher ISO will produce more noise and the only way to reduce the ISO is to have more light.

\$\endgroup\$
3
\$\begingroup\$

An expensive solution: buy a cam with better ISO performance (usually containing a bigger sensor).

When you are a lot in situations with little light and moving picture content, the investment might be worth it. I bough my D700 mainly for stage fotografie, since the light onstage often is quite dark.

\$\endgroup\$
5
  • \$\begingroup\$ +1 for mentioning sensor size, as it's hugely important. Unfortunately it's probably not practical for the original poster. :-P \$\endgroup\$ Commented Aug 2, 2010 at 21:55
  • \$\begingroup\$ I think it should be noted that it is less so the sensor size, and more so the size and density of the photosites on the sensor that matter. You could have an 8mp sensor on both a full frame and an APS-C, and the full frame will have better noise performance because of the larger and less dense photosites. \$\endgroup\$
    – jrista
    Commented Aug 2, 2010 at 23:54
  • \$\begingroup\$ @jrista Indeed, and that bigger sensor gives you a wider field of view, so you have to increase your focal length to get the same field of view. But this results in a smaller depth of field, so you have to close down your aperture, resulting in longer exposures, so you have to up your ISO again to compensate for that. If you want keep the same look, that is. \$\endgroup\$ Commented Aug 3, 2010 at 6:51
  • \$\begingroup\$ @Dave: I think that is a highly subjective concept there. I seriously doubt the idea that one would always try to make a full-frame sensor "behave" like an APS-C sensor. Sure, if you want identical field of view and depth of field and overall exposure, sure, you mitigate some (some, probably not all) of the gains a better sensor has to offer. I seriously doubt such a scenario is very realistic, however. I currently use an APS-C size sensor, and while it is great for wildlife shots, when I move to a full-frame, I will do so because of the wider field of view, not to keep the same FOV/DOF. \$\endgroup\$
    – jrista
    Commented Aug 3, 2010 at 16:38
  • \$\begingroup\$ I'm only trying to make the point that when you change something on one aspect, you also change something in another aspect. Everything's related, and the subjective thing is what you want to get out of it. \$\endgroup\$ Commented Aug 3, 2010 at 18:10
2
\$\begingroup\$

The best way for most photographs is to simply get rid of the noise in post-production using software. In the end, it will be easier & cheaper than most of the other options described here.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ In terms of software, I've been pretty impressed with the noise removal in Lightroom 3. so I'd recommend that, for lots of reason as well - not just the noise removal. \$\endgroup\$
    – Wilka
    Commented Aug 2, 2010 at 22:21
  • 3
    \$\begingroup\$ It should be noted that this removal of noise with post processing has downsides of its own, as fine detail is often identified as noise by software, and may be eliminated. Significant noise reduction in any image generally has the effect of reducing image sharpness, and many tools often add artifacts of their own when doing extensive noise reduction. Unlike adjustments to white balance, exposure, etc. in a raw file, removal of noise is not a lossless process. I recommend it when you have a minor to moderate need, but I would do everything you can to eliminate noise when taking a shot. \$\endgroup\$
    – jrista
    Commented Aug 3, 2010 at 0:00
  • 1
    \$\begingroup\$ Wavelet denoising techniques are surprisingly effective. The main cost to this is the extra time taken to adjust the wavelet denoising parameters. Gimp has a good plugin to do this registry.gimp.org/node/4235 \$\endgroup\$
    – labnut
    Commented Nov 8, 2010 at 10:05
1
\$\begingroup\$

Pretty much the only way to alter the noise is to lower the ISO. If you are afraid of having darker photos, you can look into getting a faster lens (lower f-stop), or take longer exposures (at the risk of having blurry photos).

\$\endgroup\$
1
  • \$\begingroup\$ +1 Increasing the exposure time is NOT the only other option, but it will work in some cases (see my answers/comments above) and you were the first to suggest it here. \$\endgroup\$ Commented Aug 1, 2010 at 19:24
0
\$\begingroup\$

Image stabilization (or vibration reduction), or a tripod/monopod, along with longer exposures will lower noise. Assuming your subject is sitting still, that is.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Noise almost always refers to electrical or optical interference on the sensor itself. Stabilizing the camera won't help this. In fact, a longer exposure time will actually increase noise. Perhaps you're referring to blur? \$\endgroup\$ Commented Aug 2, 2010 at 22:00
  • \$\begingroup\$ @Craig Walker: A longer exposure means you can change the ISO to a less noisy setting.. \$\endgroup\$
    – ck01
    Commented Aug 4, 2010 at 7:25
  • \$\begingroup\$ Yes, but the question title specifically mentions not changing ISO. \$\endgroup\$ Commented Aug 4, 2010 at 22:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.