ISO - a con ?

[Morgan_Moore]

There has been various debate about expectations of high ISO performance of DBs

My poorly received argumant is that the ISO setting doesnt really affect the raws file at all

ie metering at  higer than base ISO will just lead to a raw that is under exposed

So I say just put the back on the base ISO plus one stop (to create highlight headroom ) and leave it there.

That is the ISO of your back - end of story

using it above this ISO is no different from under exposing

So I tried a shot today of my screen, exposed 'correctly' at both 25 ISO and 200 ISO , the 25 ISO frame oviously used a longer shutter speed

I made my exposures so that the histos (on the camera display) were both centred at the middle

Two correct shots.

Ive done screen scrabs of the process from camera to dng

Basically the 200ISO shot looks 'correctly' exposed (exept in brum emotion reader) until the histograms are examined

(straight curve nothing applied)

AT THIS POINT it becomes clear that the 200ISO shot has way less data than the 25 ISO shot

To eqalise the histograms I have to use the 'exposure compensation' slider to add 2.5 stops to equalize the histos

I dont think too much should be read into the 2.5 value (which should according to my hypothesis be 3) due to my very rough testing

Hopefully this may contribute to peoples expecations of Dbacsk at high ISOs

Dont beleive the (ISO) hype I would say.

Enjoy the attached image...

SMM

[Morgan_Moore]

A couple of further thoughts.


-----
I dont know jack about how backs actually work - they probably have some sort of amplifier after the recording

This would lead me to think that the pure argument above is slightly flawed, because that amplifier probably gets turned up a bit, but the argument is still not far off the mark and a good way to tailor ones performance expectations
-----

Metering at high iso handheld is useful becuase you are better of with noise than camera shake

-----

With long exposures (>2s) a more complex equation of ISO noise and exposeure time noise probably has an affect in favour of using a higher ISO

-----

I am not 'knocking' my sinar back - it is great at 25-100 and okish at 200 with a chunky exposure


-----

At least I hope this encourages people to run ther own tests

[John Sheehy]

There has been various debate about expectations of high ISO performance of DBs

My poorly received argumant is that the ISO setting doesnt really affect the raws file at all[a href=\"index.php?act=findpost&pid=156597\"][{POST_SNAPBACK}][/a]

That's quite possible.  Most sensors are read out with only one gain, and various ISOs are obtained simply by using another amp before the ADC.  If that second amp is clean, then there will only be a small benefit of using higher ISOs; the fact that the ADC-introduced noise will be lower, relative to the other noise and signal.  If the amp is bad enough, then it might introduce extra noise, relative to the same absolute signal, at higher ISOs, in which case under-exposure from the base ISO is better than using high ISOs, as far as the RAW data is concerned.

It wouldn't surprise me if some of the backs that give 16-bit RAW output do so in such a way that ISO settings only change the metering, and the brightness of thumnails, while using the same gain for all ISOs.

The only way to tell is to take the proper test shots and measure the RAW sensitivity and noise levels at various ISOs.  I don't own any MF backs, but I do this for all my compacts and DSLRs, and I can make more intelligent ISO/exposure decisions, based on the camera.

[KAP]

There must be a lot of electronic magic going on after the exposure, or we could just expose at any aperture shutter speed combination then choose the iso after the event. I'm sure there is a base sensitivity of the sensor, no doubt silver halide had a base sensitivity as well but treated differently we could use another set of exposure combination. It must be the same with the digital film.

Kevin.

[Morgan_Moore]

There must be a lot of electronic magic going on after the exposure, or we could just expose at any aperture shutter speed combination then choose the iso after the event. I'm sure there is a base sensitivity of the sensor, no doubt silver halide had a base sensitivity as well but treated differently we could use another set of exposure combination. It must be the same with the digital film.

Kevin.
[a href=\"index.php?act=findpost&pid=156669\"][{POST_SNAPBACK}][/a]

On the eylike 22 (now discontinued) you could change the ISO after the event in the software - but there was no exposure compensation slider

The current sinar 54 LV (virtually the same as the E22)  has an exposurse slider in software but no 'ISO' control

The earlier eyelike precisions had no ISO control (same chip?)

I think the ISO is just an exposure compensation 'tag'

My main point is for users rating thier back at 800 to 1600 ISO to consider the results they are getting

S