Which ISO to use

[Speck Tater]

I've just made a habit of shooting at ISO 100 for best quality, but that causes me to miss some shots because there isn't enough light, or the subject is moving too fast for the max shutter speed I can get with the light.

So I decided to do some experiments with ISO settings. Here are two samples, the first at ISO 100 and the last at ISO 25600. These are 1:1 crops. Surprisingly, when the full image is scaled down to 400 or 600 pixels for web page use there's no noticable difference in quality between the two.

For the complete series of tests see: This Link




Here are the two full frames scaled to 400 pixels: (top is ISO 100, bottom is ISO 25600)

[Celebration]

1. interesting, but I think you would get a better test if the subject was closer

Try zooming on something about fifteen to forty feet away, that has something different in the background, and repeat the test. I would love to see the differences then. On a distance photo like this, when the background can't be seen, I don't think the differences are not as apparent. I realize you were attempting to do that with the scaling above, but I would love to see the differences using the entire camera frame, but under the different circumstances. Imagine you are about twenty feet away from a bird on the dock of a lake, taking a closeup shot with a zoom lens, with the trees on the other side of the lake in the background........That is just an example, though.

[Speck Tater]

2. The conditions weren't particularly good yesterday

It was partly cloudy with the sun peeking out and then hiding behind a cloud, and a nearby forest fire filled the air with haze.

Yes, a closeup target and constant lighting would be better. As for using the entire camera frame, that's a HUGE picture, 4912 x 3264 on my camera. It's hard to look at a picture that big on the monitor screen. And that doesn't duplicate normal viewing conditions. A picture is normally either cropped, or reduced, or more commonly, both. Reduced to 50% size (2456 x 1632) and given a noise reduction pass in Paintshop Pro there's really no difference that I can see between an ISO 100 and an ISO 800 picture.

Reduced to 10% of out-of-camera size (491 x 326) there's hardly any noticeable difference between the 100 ISO and the 25600 ISO! So for small web pictures high ISO is perfectly OK. The noise will disappear when the picture is scaled down.

What I need to do next is test different size prints and find out what ISO settings are suitable for what print sizes. I know ISO 100 printed at 16 x 20 shows absolutely no sign of noise even under a strong magnifying glass.

[ManiacJoe]

3. Your experiment is leading you in the correct direction.

The guideline you are looking for is this: Use the lowest numbered ISO that still allows you to get the shot.

If you are shooting a static landscape using a tripod, then by all means use the camera's lowest ISO 100/200, whatever that is.

If you are shooting moving subjects in low light, then you need to push up the ISO to whatever numbers are necessary to get the shot. Better a noisy pic than a blurry pic since this noise can be worked with during post processing.

As you have started to notice, the use of the picture will help dictate how much noise is acceptable. Pics that are scaled down to web viewing or 4x6 prints hide any noise very well, 8x10 prints will need some work on them.

[Major Nikon]

4. Pixel peeping doesn't always provide all the information you need

As you noticed, differences that are readily apparent while viewing at 100% resolution don't necessarily translate to the finished product. So a lot depends on how you're actually going to use the images when it comes to what is acceptable vs what is not.

Generally speaking with a relatively constant light source there are only 3 things which determine exposure which are shutter speed, aperture, and ISO setting. Typically I have a pretty good idea of the range of aperture and shutter speed settings I want to use. From there I select the ISO setting that puts me in that range. If I'm using artificial light, I may also up the ISO if my strobe(s) are putting out full whack. I use an app called Light Meter for my iPhone which gets me in the ballpark pretty quickly.

If I'm in a situation with rapidly changing light conditions, I may be tempted to use auto-ISO, but this is generally a last resort for me because I want to know when ISO levels are ranging into unacceptable territory. My D7000 can change ISO pretty quickly manually, so this is almost always my preferred method.

[Blue_In_AK]

5. I use a higher ISO when I'm doing action shots,

like the Iditarod, because I like the freeze-frame look, not the blur (personal preference). If I have a static shot, though, if I have a tripod, I'll go 100 or 200, even when it's dark. I think I used ISO 400 for those aurora shots last winter, but only because my wide-angle lens is f/4.0. If I had been using a faster lens, I probably would have gone with the 200 ISO. I very seldom shoot with ISO over 800.