Introduction and ConclusionThe purpose of this test was to see what degredation of the image would be caused by using high ISO settings on the camera. My conclusion is that even the highest ISO settings are useable, and certainly 800ISO can be used as a matter of course, without worrying about degradation of the image.
My long-time experience is that plenty of light thrown on the subject also helps – so very low light shots may well still show chroma and luminance noise in deep shadows.
Which begs the question of why use high ISO if there is plenty of light about? The answer is that higher ISO enables higher speeds to be used, which is great for moving subjects; and a high ISO sharp shot is always better than a low ISO clean file, but blurred, shot.
A note on ISO
For those not familiar with the term ‘ISO’, it is a measurement of the camera’s sensor sensitivity to light. The higher the number, the greater the sensitivity. The term is a holdover from film cameras. Each film is made with a specific sensitivity.
As there is a relationship between the sensitivity of the film, the shutter speed and the aperture that can be used in a given situation, photographers are always looking for the optimum combination that will give them the best image quality. With film, the greater the sensitivity, the higher the shutter speed that can be used, as less light is needed to burn the image into the film.
But higher sensitivity comes at a cost and the cost is the increasing appearance of grain in the image. The particles or (or dye clouds in the case of colour film) become more visible, the greater the sensitivity of the film.
But what about digital cameras, which have no film? Surely they always have the same sensitivity, after all there is no film and it is always the same camera? Well, increasing the ISO in a digital camera is apparently a matter of increasing the sensitivity of the the chip/circuitry. But as with film there is a downside, which is that in the ratio of the signal to noise, there is more noise (the digital equivalent to grain) compared to the signal.
Or rather, in this fast-moving digital world, the downside is becoming less of an issue as time goes on. For the detailed results – read on.
The Results
I shot these images in RAW and opened them in Photoshop CS2 with my default RAW conversion set-up, which has no sharpening, luminance smoothing, or colour noise reduction. Each image was then Saved-For-Web with no sharpening or colour correction.
The photographs were shot with the 18-55mm kit lens with the focal length set, according to the EXIF data, at 48mm.
The shots shown here are crops of a few hundred pixels from near the centre of the image. I set up a standard crop size for the crop tool in Photoshop and cropped each of the shots to the same number of pixels.
For comparison purposes, the first shot is taken with a Nikon D200 with a 50mm lens and set at 100ISO.
CLICK on any of the photos for a larger view.
D40 at 200ISO (the lowest ISO setting for this camera)
D40 at Hi-1 (3200ISO equivalent)
There is more detail in the shot from the D200, but that is to be expected given that it has a larger sensor and that the 50mm lens is probably sharper than the 18-55mm kit lens on the D40. Nonetheless I think it is clear from the shots that even the highest ISO is useable.
Of course, I want to know whether whatever differences there are between the shots, would be visible in a print. I read somewhere that looking at an on-screen shot at 25% gives a good approximation of what can – and cannot – be seen in a print. If that is so, then these crops are certainly a stringent enough test.
I have not printed these images. Some time ago though, I printed a portrait I took at 800ISO shot on the D200 with 105mm f2 DC lens (a very sharp lens) to 20×30 inches (45x150cm) and it looks as smooth as one would want. Or rather, it has an ever so slightly sketch-like look if one looks very close; which is how I have found that the D200 tends to render images at high ISO. For portraits it is very appealing.
.