Micro 4/3 Oly OMD v. Oly EP3 - ISO comparison

thanks Armando ... though I do get irritable about this constant emphasis by bloggers and other reviewers on how cameras perform at high ISO ... if I replace (or supplement) my E-P2 with another m4/3 body, it'll be because the new one has much greater DR and much better noise performance in the shadows at ISO 100 or 200, rather than because it generates acceptable noise at 3200 ...
 
I wish I knew where to look for it, but as I understand it, today's sensors can't be optimized for both ends of the spectrum - low and high ISO. Since "the marketplace" places the emphasis on high ISO low ISO suffers a bit. And that is why some sensors start at 200 ISO.

Anyone know more about this?

I'm with you PDH - I'd rather have a clean low ISO with strong dynamic range. Lately, I've been shooting a lot with tripod and I don't need the high ISOs. Also, I don't mind grain/noise at higher ISOs! But still, maybe someday we can have it both ways!
 
Definitely agree on the low ISO stuff, it's much more important for me! Olympus say the noise difference between ISO 100 and 200 is negligible so they used 200 as the base level, but it will still raise the noise floor and reduce DR. Besides, for long esposure shots, a low ISO value allows for longer shutter speeds without resorting to a ND filter. Bring on the ISO 50!

Having said that, it looks like about a stop and a half difference in noise and sharpness performance - impressive!
 
What annoys me these days is the stupid megapixel race, people who know no better say "ohhhh how many megapixels is it?" as if that's the be all and end all of a camera :-(
 
Improving high ISO does not mean that there are no advantages for low ISO. By improving high ISO the dynamic range at low ISO is improved, too. One can reduce noise by improving the readout noise, for example, which would give us cleaner dark areas at low ISO.
 
I was under the impression that the Panasonic G3 with the latest 16MP sensor was actually noisier at low ISOs than the equivalent 12MP m4/3 cameras, although I don't know if the extra resolution cancels out the difference when you view the images at the same size.
 
When the K-5 came out, (and the K-r) everyone was banging on about how great it is that you can have an ISO in the 52k region. WHAT? I tried it under circumstances when I would maybe want to use a high ISO and it was HORRIBLE. I try not to go over 3200 (and only if I must) and generally stick between 100 and 400.

Basically I stay as low as the camera will allow.
 
Yep, cramming more megapixels onto the same size sensor isn't the right way to go, IMHO of course.

It is not that easy. Sensor technology advances and therefore it is possible that newer sensors are better even with more pixels. For example, there is no doubt that the new 16MP sensor is better at high ISO than the old 12MP. However, if sensor technology is the same, a sensor with less pixels will be less noisy at 100%. But don't forget that we are looking at different magnifications, because the whole image of a 12Mp sensor at 100% is much smaller than a image of an 16MP sensor at 100%. Thus looking at 100% is the wrong way to compare two images of different sensors. If all other things are equal, more pixels will contain more information and will be better.
 
It is not that easy. Sensor technology advances and therefore it is possible that newer sensors are better even with more pixels. For example, there is no doubt that the new 16MP sensor is better at high ISO than the old 12MP. However, if sensor technology is the same, a sensor with less pixels will be less noisy at 100%. But don't forget that we are looking at different magnifications, because the whole image of a 12Mp sensor at 100% is much smaller than a image of an 16MP sensor at 100%. Thus looking at 100% is the wrong way to compare two images of different sensors. If all other things are equal, more pixels will contain more information and will be better.

Well put!

This is the big flaw in comparing cameras of differing resolutions by using 100% crops. For want of a better term, it is comparing noise and resolution "per pixel", not looking at the image as a whole.
 
Back
Top