>2009/1/30 Chuck Norcutt
>That's interesting. DPReview has been reporting actual ISO test results
>for a long time. DXO says the Canon 5D's ISO 100 is actually 92.
>DPreview reported it as 125. And, of course, Canons says it's 100.
>Who's right? I doesn't make any difference until I have to get out the
>external meter for flash.
I came across this discussion elsewhere a couple of weeks ago, which may be
relevant:
-----------------------------------------
> I have a hard time trying to reconcile DXO results with real world
> testing by Phil and many other reviewers.
The discrepancy between DxO's results and the manufacturer's ISOs is very
straightforward to explain, and it's all down to how metering and exposure
works.
Basically the conflict is between two ways of measuring ISO. One method is
to determine the exposure required such that, if you spot meter off an 18%
grey card and then take a picture of it, the grey is rendered at a luminance
of 50% in the resultant image file. I'd argue this is the method most
relevant to how the majority of photographers work, and if you test cameras
this way they will all give essentially the same answer (always bearing in
mind that ISO 12232 allows a tolerance of 1/3 stop either way, probably as
much to cater for manufacturing variations as anything else).
However DxO use a different method, which is based on the point where
highlights clip to white. This does not necessarily give the same answer as
the grey-point method due to tone curve differences between cameras, and the
difference between the two methods generates the discrepancy in DxO's
results. In fact it turns out that the difference actually reflects the
highlight dynamic range of the camera; the lower DxO's "true" ISO, the more
higlight DR you've got.
Perhaps the easiest way to consider this is to think in terms of taking two
images of a high dynamic-range scene, using identical shutter speeds and
apertures but with and without Highlight Tone Priority turned on. Visually,
both will be correctly exposed, and by the grey card test, both shot at the
same ISO. But the highlight clipping point will be shifted by a stop between
the two, and therefore using that definition they will have been shot at
ISOs one stop apart, with an ISO 200 HTP shot considered to be ISO 100.
Now this doesn't actually mean DxO's method is wrong, as they are only
concerned with RAW data not output image files, indeed it's the only way
they can compare RAW output on an equal footing. But it's purely about
technical comparison of RAW data, not about using the camera.
--
Andy Westlake
dpreview.com/lensreviews
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|