DxOMark's measured ISOs vs. manufacturer ISOs

Started Mar 13, 2013 | Discussions thread
Jack Hogan Veteran Member • Posts: 6,983
Re: DxOMark's measured ISOs vs. manufacturer ISOs

Mjankor wrote:

Jack Hogan wrote:

Mjankor wrote:

Well you would generally have more tonal range, less noise, and therefore a better image, yes? How is that a bad thing, as a photographer? First time I've heard of a photographer running a higher iso than needed for the purpose of "photographic intent". Tell me more.

Of course you are right in general. And now we are no longer comparing cameras. This is exposure strategy with one single camera in a real world situation.

You are shooting indoor sports and you have determined that ss could not be any slower than you set it because of blur, and you could not make aperture any wider because of your desired dof. You also could not raise ISO above 3200 because otherwise you would clip certain highlights that you would instead like to retain some detail in. Keeping Exposure fixed, could you lower the ISO?

The answer is camera specific (reread the link in my last post), keeping in mind that some in cameras (Canon's for instance), lowering ISO in such a situation and fixing brightness in post will result in an overall noisier final image*.

Jack
* That's not the case for your EM-5, though, with which you could easily dial back to ISO 800 (or even 400) and recover it in post with virtually no noise penalty in the final image (1/6 of a stop at 400) but with the greatly added benefit of 2-3 stops better highlight headroom. Take a llook at this great chart by Bill Claff to figure out how to set ISO.

So, applying that to the IQ tests between the G3 and the OM-D, if you have the same test scene, why wouldn't you shoot it at the same iso and same exposure with both cameras, thereby resulting in a significantly better image from the OM-D compared to the G3 and better representing their performance in real life?

Choose your objective: real world 1 camera set-up decision or fair 2 camera comparison as many things being equal as possible?  As I mentioned earlier, once you understand what is involved, if you want to see how two sensors would perform in a similar situation, just read the DxOmark charts: it's all there fair and square.  On the other hand some other sources tend to be much less open and sometimes end up comparing apples to oranges

Post (hide subjects) Posted by
tko
tko
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow