Is the ISO you set the one you really get?

Flash In The Pan

Flash In The Pan

Suspended / Banned
Messages
27,793
Edit My Images
Yes
An interesting article from luminous-landscape.com comparing data from DxOMark which asks the question "are camera manufacturer's increasing the iso settings without the photographer's knowledge?".

Discuss :D
 
I never trust DxO - their raw conversion is so poor, I don't think they are any authority on raw personally.

DxO figures derive exclusively from their interpretative of raw data... do we trust their interpretation.

I think DxO marks are more willy-waving..ie "my sensor is better than yours" - when in reality no one uses just the sensor..but rather than whole system. Seems a bit of a pointless mark really...
 
puddleduck said:
I never trust DxO - their raw conversion is so poor, I don't think they are any authority on raw personally.
Click to expand...

What was interesting to me was that they arrived at different results for cameras that have (supposedly) identical sensors (D80/D200 and D300/D300s)........
 
Does it really matter as long as the resulting images are correctly exposed?
Unless you use a hand-held light-meter and don't check against the histogram, that is...

I can see where you're coming from regarding the upper end though, but negative film was just the same: 125ASA (the 'old' ISO to you noobies...lol) B&W film always behaved better when exposed at 100ASA...
 
This article really annoys me - camera manufacturers give settings in ISO, not sensor gain. ISO is supposed to be a well defined standard based on the response of the sensor (be that film or digital), while gain on the sensor is anything the manufacturer wants it to be (if the manufacturer wants a sinusoidal gain curve then it would be weird but well within their rights to make it so). It appears that LL are complaining because manufacturers aren't using a uniform (linear) gain to iso conversion. Who cares? If they also add an aperture or even, for some reason, focal-length dependent correction to the iso then once the outcome is a correct exposure, they are doing the right thing!. We equate ISO to noise because that's how film upped the ISO by increasing grain size and how digital usually does it by increasing the gain on the sensor, but ISO is not a noise standard - it's a sensitivity one!
 
Flash In The Pan said:
What was interesting to me was that they arrived at different results for cameras that have (supposedly) identical sensors (D80/D200 and D300/D300s)........
Click to expand...

Indeed - despite the sensors being identical.

So that probably tells you all you need to know re: DxO sensor scoring.
 
even though i've never really thought much about it, it does make sense.

who knows how much light loss is caused by anti-aliasing filters and other microlenses on the sensor? even though a camera might have an identical sensor as another in the range, do we truly know if they have identical coatings and microlenses, much less electronics and firmware/microcode?

i guess, as has been mooted, does it matter? and does it matter that your camera knows this and does a bit of wizardry of the kind as to give you an illusion your lenses and sensor are behaving as you expect rather than how they really do?

actually, as long as it helps me take a better photograph rather than a worse one, i'm not sure i worry about it.
 
Last edited:
Whats the point? So the camera might raise the ISO slightly to compensate for light loss. You dont know its done, you cant see any negative effects, so whats the issue?

Is the writer of that piece going to time the shutter to make sure its exactly 1/250 of a second.. Not 1/252 of a second, gosh no, that would be a travesty.

I think he needs to stop worring about trivial issues and actually get out and use his camera.
 
I looked at the Dxo mark figures for a few cameras and they clearly don't bear much relation to real life image samples from the cameras, so I take their figures with a hefty pinch of salt.
 
Like others here I'm wary of entirely theoretical comparisons using figures from DxOmark, the most telling part about this article is that there isn't any sample images which demonstrate the problem they're talking about. In practical tests, neither of the conclusions they derived from the raw data appear to be correct which I think Andy over on DPR demonstrated with actual photos which showed larger aperture lenses still let in more light than moderate apertures.

If the problem is as serious as LL are making out, then surely it shouldn't be difficult to provide samples of it in action but the fact they haven't says a lot.

John
 
Some of the American forums are getting quite agitated about this, with hundreds of posts. There may be something in it, but I'm not convinced the source data is robust and it's well known that there is no standard for ISO anyway. But it all appears to work pretty much as it should, even if it would be nice if we knew what was actually going on.

Either way it doesn't seem that significant, other than it's a bit of slight of hand by the manufacturers. But if you want to go down that road, what about zoom focal lengths that fall quite a long way short of the claimed range, focus 'breathing', and f/numbers that are sometimes quite a way adrift of actual transmission?

Not to mention natural optical vignetting which often shows one or two or even more stops of fall-off around the periphery. In the context of that, a third or half a stop seems neither here nor there.

At the end of the day, one of the great beauties of photography is that you can simply take a picture, and the net effect of all components in the system is revealed. In this case, I've not seen any pictorial evidence that is likely to unsettle the judges. Just a load of numbers and a bunch of often half-baked theories :shrug:
 
Ignoring Dxo for a moment, different manufacturers have been shown by multiple test sites that they will ship cameras with widely varying ISO settings. You can see the appeal. Showing that brand X has the same noise at ISO 1600 as brand Y does at ISO 800 is the sort of thing that would imply that brand X is better.
 
Flash In The Pan said:
:thinking:

ISO...International Organization for Standardization

:shrug:
Click to expand...

As I understand it, ISO refers to film sensitivity. I'm not sure exactly how it's worked out but for the sake of illustration let's say a certain level of light required to deliver 18% grey.

Digital sensors work differently in that the starting point is not 18% grey but the maximum amount of light they take take before blowing/clipping, and 18% grey is worked back from that. And with different sensors and different amounts of gain applied, that falls in different places according to the dynamic range. I think this is maybe behind the 'calibrated for 12% or 15% grey' claims.

The result of this is that a sensor can be said to have pretty much any ISO rating you like, according to how much amplification is applied, and the numbers that manufacturers apply are kind of arbitary, almost what they can get away with rather than a measure of any fundamental characteristic.

Don't quote me on that, but it's something along those lines. This article from DPreview attempts to make more sense of it http://blog.dpreview.com/editorial/2009/09/sense-and-sensitivity.html
 
ISO stands for International Standards Organisation.

That it also has a 'standard' for film speed/digital sensor sensitivity is by the by...
ISO sets standards for everything from washing machines to aero engines and screw sizes.
Like the old British Standard kite-mark... you still find some products marked BS5750 or whatever...
 
Arkady said:
ISO stands for International Standards Organisation.

That it also has a 'standard' for film speed/digital sensor sensitivity is by the by...
ISO sets standards for everything from washing machines to aero engines and screw sizes.
Like the old British Standard kite-mark... you still find some products marked BS5750 or whatever...
Click to expand...
Have the standards been ratified for digital as well? I know there are different standard subtypes for types of film (colour, B&W, negative)
 
Take a look at the still life shots on imaging-resource for the Panasonic G2 and the Olympus EP1. You will see that if you look at the exif data for each camera at the same ISO settings the panasonic shutter speed is faster in each case. This is because when a G2 reports its sensitivity at e.g. 800 it's actually quite a bit more sensitive than that but when the EP1 reports at 800 it's actually quite a bit lower.

Disclaimer: ISO sensitivity data obtained from DXoMark which actually shows more than an entire stop difference between the EP1 and the G2 i.e. the sensitivty of a G2 at ISO 400 as actually more sensitive than an EP1 at ISO 800. All rather confusing but interesting at the same time.

EDIT: Canon, Sony and Nikon also all report higher ISO sensitivities when compared to the ISO standard.
 
Last edited:
I think that as long as everything looks good with the final shot then it's not a problem.
You can read too much into statistics and DXo Mark figures.
 
You must log in or register to reply here.
Back
Top