[Gandhi]

higher iso. Longer exposure=higher photosite temp=more noise.

but that may depend on the original size of the photosites, quality and size of sensor, the way the camera amplifies the raw signal from the sensor etc.

Suck it and see I say.

My fz30's noisy anyway. alledgedly. Will get round to posting some shots when I actually get outside and it's not raining.

[SammyC]

What you are trying to do is not expose left or right but just make use of the full range of the RAW colour space. So, you use the ISO gain factor to try to widen your histogram to fill the full range, and expose correctly. In some circumstances you have a very high contrast image that has a wide range automatically, so no scaling of the range is possible so we use a low ISO. If you have a low contrast image then a high ISO can scale the values to fill the full range.

You should use the highest ISO you can without blowing any highlights. On a bright sunny day that might mean that ISO100 is your only option.

So, as I work it out Steve, it's preferable to get as wide a range of values as possible. In order of effectiveness (i.e. less noise):

1) More light, either by it being there or a long exposure
2) High ISO, so doing the light value amplification at the analogue stage in the camera
3) Under exposing and boosting at the digital stage in software.

This is because as you move down the path from light source to RAW file you are losing information, digitisation is inherently lossy.

But this is all theory!

SammyC I understood that explanation, thanks!.

[Steve]

Don't you just love all this theory and practice stuff, you learn as much as you can, read the manuals, scan the web, talk to people then go out and practice were you find that all the theroy makes no sense

Thanks Sammy for wading through it all and trying to help out us dumb people.

[SammyC]

That my thinking too Steve! I can see how this ISO thing works, I can see the implicatinos of it, but I have no ****ing idea how to apply this etc. lol

If I get the urge I might try to write this all up for putting somewhere, seems a shame to lose this now that we have the beginnings of an understanding.

[SammyC]

Originally Posted by Marcel

SammyC I understood that explanation, thanks!.

My pleasure! It's nice to be able to 'give something back' after it's just been take take take so far!

[SammyC]

Originally Posted by gandhi

higher iso. Longer exposure=higher photosite temp=more noise.

but that may depend on the original size of the photosites, quality and size of sensor, the way the camera amplifies the raw signal from the sensor etc.

That's a different type of noise than I was talking about. In fact what you are talking about is more true noise rather than the posterisation noise that I've been talking about which is more deresolution issue that just looks like noise.

Do you think this has much effect on the image then?

[Gandhi]

I think the test would be to take a short expsure, a very long exposure, and then another short exposure and compare the noise in the two short ones.

It may have an affect on smaller sensors, but when you get to aps-c sized and full frame sensors, i think the difference would be negligible.

My FZ30 is noisy relative a dslr. but that's cos it has 8 Mp on a 1/1.8" sensor. back to what I was saying about the size of the sites and noise, as the pixel density is lower on larger sensors, there is also less leakage of current from one site to another. again, less noise.

if you try and pull shadow detail you will always end up with more noise as less light hitting the sensor means a lower voltage and therefore a lower s/n ratio.

I think a lot of the stuff abut heat etc is theoretical and has little real world influence.

[Steve]

I understand what you are saying gandhi but this must also have some effect on the results gained from DSLR's with or without full frame sensors. Higher ISO has traditionally produced higher noise, what is being said (I think) is that if you get the exposure correct even at higher ISO's there is no greater noise. Its just proving it and wading through all the theories to get the facts with real world examples and then to be able to use that in our daily photography.

[SammyC]

Depends which 'noise' you are talking about there Steve, Gandhi's heat noise or the rounding error noise that you originally introduced and that I have continued with.

For the second case, the rule of thumb is that there is less 'noise' in higher ISOs (up until clipping occurs) for a given exposure (constant shutter and aperture). I think you will always get a better result if you have control over the exposure and use a lower ISO though.

Just as a little note. While I've been (trying to) calibrate my monitor today, I came across an article which explains the method of 'exposing to the right', which is basically what this method is.....

http://www.luminous-landscape.com/tu...se-right.shtml

[SammyC]

Good spot there Marcel! I personally didn't realise that it wasn't linear.

[Gandhi]

Cheers MArcel, I Knew I'd seen that somewhere, It's the reverse ansel adams theory, he always said 'expose for shadows and process/print for the highlights'