Prev: Appropriate colour choice for a P&S, "puke" green
Next: Quantum dot image sensors
From: RichA on 8 Jul 2010 22:46
You've seen them. Those little 1/2 to 1" pictures of photo segments
magazines use to "show" you the ISO performance of a camera. They are
worthless. In many cases, 100 ISO looks like 1600 because the scale
is too small, and the reproduction simply awful. So why don't
magazines show some other aspect of the test other than this? It's a
waste of paper and effort and only makes web-based test sites that
much more appealing.
From: Me on 8 Jul 2010 23:38
On 9/07/2010 2:46 p.m., RichA wrote:
> You've seen them. Those little 1/2 to 1" pictures of photo segments
> magazines use to "show" you the ISO performance of a camera. They are
> worthless. In many cases, 100 ISO looks like 1600 because the scale
> is too small, and the reproduction simply awful. So why don't
> magazines show some other aspect of the test other than this? It's a
> waste of paper and effort and only makes web-based test sites that
> much more appealing.

No, I haven't seen them.
Most web based tests are flawed too, showing either 100% pixel mapped
view without taking into account scaling of the whole frame, they use
some arbitary in-camera NR and sharpness settings (jpeg) or do not use
an optimum raw converter for the camera's raw files. Exposure accuracy
and the ability to recover highlights so that optimum exposure can be
used also have a large impact, as does pattern noise in shadows, both
mainly ignored by self-appointed "expert" sites like DXO. Comparing
different formats at same ISO (using same f-stop & shutter speed) is
also potentially seriously flawed for the purpose of assessing real
world low light performance. In the end, there's much more
disinformation out there than useful information. YMMV.
From: John Navas on 9 Jul 2010 13:11
On Fri, 09 Jul 2010 15:38:20 +1200, in <i165k0$md8$1(a)news.albasani.net>,
Me <user(a)domain.invalid> wrote:

>On 9/07/2010 2:46 p.m., RichA wrote:
>> You've seen them. Those little 1/2 to 1" pictures of photo segments
>> magazines use to "show" you the ISO performance of a camera. They are
>> worthless. In many cases, 100 ISO looks like 1600 because the scale
>> is too small, and the reproduction simply awful. So why don't
>> magazines show some other aspect of the test other than this? It's a
>> waste of paper and effort and only makes web-based test sites that
>> much more appealing.
>
>No, I haven't seen them.
>Most web based tests are flawed too, showing either 100% pixel mapped
>view without taking into account scaling of the whole frame, they use
>some arbitary in-camera NR and sharpness settings (jpeg) or do not use
>an optimum raw converter for the camera's raw files. Exposure accuracy
>and the ability to recover highlights so that optimum exposure can be
>used also have a large impact, as does pattern noise in shadows, both
>mainly ignored by self-appointed "expert" sites like DXO. Comparing
>different formats at same ISO (using same f-stop & shutter speed) is
>also potentially seriously flawed for the purpose of assessing real
>world low light performance. In the end, there's much more
>disinformation out there than useful information. YMMV.

No kidding. The only real way to find out is to make your own images.

--
John

"Assumption is the mother of all screw ups."
[Wethern�s Law of Suspended Judgement]
 | 
Pages: 1
Prev: Appropriate colour choice for a P&S, "puke" green
Next: Quantum dot image sensors