PDA

View Full Version : ISO impact on SHARPNESS ? Why



David
22-10-2009, 11:16pm
I get it that when you slow down the SS and expose the image for longer the NOISE risk/factor increases because of the longer exposure time. I get it that a faster Apeture (2.8 cf 5.6 for example) reduces the risk of NOISE for the same reason: less exposure time consequential to a wider apeture and therefore more light getting into the shot.

But what I dont get is when people say higher ISO means less sharpness overall in said image- Why dont I get it ? Because an increase in ISO means you end up with a faster SS (less exposure time) and the ISO based increase of the amount of light hitting the sensor compensates for the lack of light coming from the Apeture (smaller hole = less light) - no ?

Questions Im asking myself include

A. If that is factual, why ?

B. Does it apply to all focal lengths from say 10mm-400mm

C. Does it apply to all lenses (including L series Canons and other higher quality lenses)

D. Does it apply to all light conditions and

E. Is it the NOISE factor alone that messes with sharpness with higher ISO settings

I dont shoot above 200 ISO even with my 50D which copes with ISO alot better than my 400D did but I leave the AUTOISO alone (sometimes why I dont use it) so ISO/NOISE issues wont be a problem for me unless/until I am doing something like star trails with long long exposure times which I can overcome to a degree with the in camera NR function etc etc.

Thoughts, comments and rueful shaking of the head at my probably misinformed ignorance most welcome :rolleyes::)

kiwi
22-10-2009, 11:23pm
actually apperture has nothing to do with exposure time, only shutter speed affects that. Apperture does affect exposure though as it controls the amount of light that hits the sensor

High ISO introduces noise, like grain but a bit different. Noise starts to remove definition at a pixel level progessively as it increases. Noise reduction when applies effectively smoothes (ie blurs) the image, so either way you are up creel without paddle.

That's my basic understanding anyhow

ricktas
23-10-2009, 5:11am
Agree with Darren. It isn't so much a lack of sharpness as you see when a lens is not focused correctly, but the noise leaves less definition at the pixel level, so end up using some noise reduction software, that smooths it all out, then sharpen the bits that need it.

I @ M
23-10-2009, 5:51am
David, the electronics of the digital camera can be directly compared to an amplifier and speakers in a stereo system. A simple analogy is that the higher you crank up the volume ( high ISO amplification ) the more chance you have of getting fuzzy sounds from your speakers ( loss of edge sharpness in pixels ). That is a very basic way of looking at it and there are quite a few other influences as well.

David
25-10-2009, 2:34pm
Thanks for the responses guys.

Noise is inevitable for long exposure shots I am thinking of (Star Trails) and there is software specific to Trails that you can download or as Rick says, Ill have to go and learn how to sharpen up the softened pixels where it matters- best leave the ISO low by the look of it, enough noise from the long exposure as it is will be enough to contend with.

Prang
25-10-2009, 5:15pm
I think sharpness comes into it when using noise reduction converting to Jpegs.

All cameras seem to do this differently .

swifty
25-10-2009, 7:29pm
Thanks for the responses guys.

Noise is inevitable for long exposure shots I am thinking of (Star Trails) and there is software specific to Trails that you can download or as Rick says, Ill have to go and learn how to sharpen up the softened pixels where it matters- best leave the ISO low by the look of it, enough noise from the long exposure as it is will be enough to contend with.

I think you need to differentiate between ISO noise and noise you get from long exposures, called dark current noise I think.
As others have pointed out, softening of images from iso images are often the result of noise reduction that effective smooth out some detail.
Low iso, long exposure noise results from things like the sensor heating up during exposure. Things like dark frame subtraction noise reduction can be effective at removing these artifacts without detriment to detail.
Theoretically I think you should achieve better results with lower iso and longer exposures because of the overall higher signal to noise ratio (more photons collected) even if there might be more overall noise. But I don't really do star trails so I don't know which is best on a practical level.