There is some evidence, I think, that the OM-D applies some non-defeatable NR even in raw files, which could well explain why the noise in the OM-D sample image looks so much smoother.

A comparison shot at equivalent apertures will tell us more, as we'll be able to see how the two cameras handle detail in the shadows. If the OM-D raw is applying smoothing to the raw files, it should show up. To see that, though, you'll have to shoot both cameras at relatively smaller apertures (e.g., f5.6 and f11), or focus on the shadow detail instead of the foreground.

It might also be interesting to see how each looks after some NR is applied in LR. The 5D files typically clean up very well.

It's the other way around actually, which makes the results even more surprising.

__________________
Current Gear: EM-5 | E-PM1 | 14-150mm | 20mm f1.7 | Rokinon 7.5mm f3.5 | Fujian 35mm f1.7

For Sale: 14-42mm R, $90 gifted/shipped

There is a long (and typically annoying) thread on DPR right now in which this question is debated. One of the DPR guys reviews the DPR test shots of a couple of cameras known to apply NR in raw, a couple known not to do so, and the EM5. Based on that, he makes a pretty convincing case that the EM5 doesn't appear to be applying NR to raw files.

Actually The E-M5 got one stop more exposure based on the EXIF.

E-M5 was at ISO 800, 1/60s and f/4
5D2 was at ISO 200, 1/30 and f/4

If we lowered the ISO of the E-M5 to 400, it'd need a shutter speed of 1/30s to get the same exposure.

Still, if we assume that ISO 800 on the E-M5 is equiv to ISO 400, that's still one stop more sensitivity on a sensor that's 1/4 the surface area of its comparison. Impressive.

I do, however, have a sneaking suspicion that there is some undocumented light-loss through the 12-50 at certain focal lengths that does not correspond to the aperture (DoF is on point, but light saturation is not).


Big question to the OP, was the E-M5 set to auto-gradation?

__________________
E-M5 -- E-PL1 -- PL-25 -- m.12-50 -- m.17 -- m.45 f/1.8 -- m.14-42 (non-MSC) -- m.40-150

I glanced at that, and gave up. But I'm not sure any of DPR's studio test shots can be used to reach comparative conclusions like that. There just seems to be a ton of variability in their tests.

But either way, some comparison shots of the OM-D and 5D2, shot at the same ISO, "equivalent" apertures, and focused on the shadow detail will be interesting.

I hope auto-gradation doesn't affect the raw file, which the OP says he used.

Auto gradation applies a 1/3 stop reduction in the exposure to preserve highlights, but otherwise shouldn't affect the raw image. On the E-M5 sensor I find this to be conservative and I apply +1/3-2/3 exposure compensation to bring it back to or just above the normally metered exposure value.

__________________
Nic
E-M5, GH1
Flickr Photostream
>> Insert quote here <<

My findings are that the OM-D is freakisly amazing at high ISO situations.

What I do find is that when the subjects are at a farther distance with my 25mm f/1.4 landing a shot is no where near in the high 90% range. The 5dmk2 does a superior job of landing every photo while the contrast focus of the OMD falls behind. Comments such as this means sharpness is compromised since the camera can potentially front and back focus. The OM-D is a phenominal camera for serious hobbyists or casual use. For professional applications caputuring human emotions with high accuracy the dslr is still unbeatable.

The Raw file of the 5dmk2 can be pushed and pulled alot. The OM-D file is not that rigid but you can feel the difference as you change the exposure in LR4.1. The 5dmk2 is much more flexilble for serious paid sessions.

i do feel the 5d classic lags behind a bit in ISO performance if you compare it to the OMD. The 5dmk2 is still a different beast that I feel much more confidence in AF. Noise wise the OM-D is serious contender to many full frame camera's out there. I would say the OM-D puts virtually all of the dslr crop sensor to shame when your comparing them to static subjects.

I use the OM-D to supplement for the love of photography. Its a tool for many applications but if I have to do serious shooting I cant drop the dlsr gear at home.

Yes, indeed --that was a typo!

David

No, you're not crazy. This is actually what I would expect. The 5DII is notoriously bad in terms of shadow noise when you try to lift the shadows. This is one of the very few problem areas for Canon in terms of image quality, but it's a big one if you need the flexibility to lift shadows in post, and even their newest camera the 5DIII isn't that great in this area--better than the 5DII but still not there. So the E-M5 should definitely do better in shadow liftability than the 5DII and likely even the 5DIII.

I've only had my E-M5 for 2 days, but from what I've seen they're neck-and-neck at ISO 3200. The amount and size of grain is about the same with very minor differences perhaps ever so slightly in favor of the 5D. However the 5D has some banding in dark areas that is nonexistent on the E-M5 so I'd give a slight edge to the E-M5 because of that.

Center AF point, sure. Outer points not so much.

Amazing, isn't it? The naysayers really need to take a closer look before dismissing it. The fact that I no longer have to reach for my 5D for image quality reasons when I'm shooting at ISO 3200 is a huge benefit.