E-PM2 is better than GH4 in true low light?

mpgxsvcd wrote:

Everyone talks about shooting in low light. However, I don't think they really know what low light is. I shoot Astro-photography with Micro Four Thirds. Most of the time we are dealing with minutes for shutter duration instead of fractions of a second.

I have an Olympus E-PM2 that I had modified to remove the stock filter to let more Ha light in. Removing the filter worked well for objects that are primarily Ha light only like the Horse Head nebula. However, the modification really doesn't help much for objects like M27 that span the full visible spectrum.

I also have a GH4 that is unmodified that I typically use for Moon and planet videos. However, I hadn't tried it for AP yet. I decided to take 12 frames of the exact same object(M27 Dumbell Nebula) with both cameras and compare them.

I expected the noise to be less for the GH4 because it is supposed to have added cooling for the video modes that it has. It was extremely warm and humid that night but I used the exact same ISO, shutter duration, scope, filter, and number of subs for each camera. The details of that are below.

I think it is pretty clear that the E-PM2 has much less noise after stacking. I am not sure why this is happening but it definitely means that the E-PM2 is a great choice for Astro Photography. Olympus was selling the camera with the kit lens for $170 last week. Most people drop $200-$300 on a T2i for AP.

Let me know what you think.

Twelve 2 minute ISO 3200 images with an 800mm F4.0 Newtonian and a Light Pollution filter in my backyard. Stacked in Deep Sky Stacker with Flats, Darks, and Bias images.

Olympus E-PM2

Panasonic GH4

Olympus E-PM2

Panasonic GH4

The E-PM2 uses a Sony sensor while the GH4 uses a Panasonic sensor, however the difference in low light scores on DxO isn't that great. The GH4 ranks a little higher overall in total score but the E-PM2 ranks a little higher in low light.

The GH3 uses a Sony sensor, so I'd be interested in how the GH3 compares to the other two.

Also, I assume you stacked RAW images. If it was JPG, then differences in settings can affect the noise. I'd like to see one out of the 12 images (unaltered) that you stacked for each camera. I presume both cameras had the long shutter NR activated as well.

p.s. Sony sensors are known for their low light capability, although I'm leaning toward there being some other factor accounting for the result.

Also, are you sure that removing the IR filter wouldn't give the E-PM2 some advantage in this test. If you're not using an IR pass filter, then removing the IR cut filter allows a camera to take in more light, both in the IR and some wavelengths of visible light. I don't know if that would have a major effect on this test or not.

mpgxsvcd wrote:

You have some good questions. Let me try to answer as many as I can.

These images were produced by stacking the RAW images together with flats, Bias, and dark images from their respective cameras. In camera dark frame subtraction was not done because it was done in the post processing stacking procedure with dark frames that I took with each camera after the light frames were taken.

I believe that temperature plays the biggest role here. Remember these are not just high ISO images. They are also long shutter duration images. The heat build up will affect the ISO performance especially on a hot summer night in the south.

I wanted to find out which camera produces the better stacked image. The single frame image from the GH4 might actually be better than the single frame OE-PM2. However, both of those single frame images is terrible when compared to the stacked image. Stacking works. Trust me.

The stock filter that takes out UV and IR light plus the narrow spectrum of Ha light was removed and replaced with a filter that blocks UV and IR light but let's in Ha light. The filter only affects light that gets in and not the noise. Since this object doesn't have a great deal of Ha light the difference in filter should be negligible.

Here are the single frames from each camera. These are the JPGs from each camera. Sorry I don't have access to the RAWs for both cameras at the moment so I can't post those.

GH4 Single Frame 2 minutes ISO 3200

E-PM2 Single Frame 2 minutes ISO 3200

I understand better now, but not fully because I've never done stacking. I'm assuming you took one dark frame for each camera after the 12 light frames. Does that work equally to taking alternating light and dark frames (i.e. 24 frames for each camera)?

Also, I highly doubt that heat would affect the GH4 more than the E-PM2. The GH4 is a tank with extensive heat sinking (as you mentioned for video performance). I just can't see how the E-PM2 could move heat away from the sensor better than the GH4, given that the IBIS doesn't allow for good heat sinking.

If your findings are correct and you made no procedural mistakes, then it's far more likely that it's the Sony sensor in the E-PM2 that handles this type of procedure better than the Panasonic sensor in the GH4.

Because the GH3 uses a Sony sensor, I'm wondering how it would perform on this test.

mpgxsvcd wrote:

Beach Bum wrote:

I understand better now, but not fully because I've never done stacking. I'm assuming you took one dark frame for each camera after the 12 light frames. Does that work equally to taking alternating light and dark frames (i.e. 24 frames for each camera)?

Also, I highly doubt that heat would affect the GH4 more than the E-PM2. The GH4 is a tank with extensive heat sinking (as you mentioned for video performance). I just can't see how the E-PM2 could move heat away from the sensor better than the GH4, given that the IBIS doesn't allow for good heat sinking.

If your findings are correct and you made no procedural mistakes, then it's far more likely that it's the Sony sensor in the E-PM2 that handles this type of procedure better than the Panasonic sensor in the GH4.

Because the GH3 uses a Sony sensor, I'm wondering how it would perform on this test.

Actually I took about 16 Dark frames for each camera(More Darks than lights). Those Dark frames are averaged together to make one master dark frame for each camera. This works as well or better than taking in camera Dark Frames.

I am not sure why the E-PM2 performs better than the GH4 in this test but it does. If you want to get into AP then my opinion is that you can't beat the E-PM2 even with the GH4. I have the GH3 as well. I could test that one. However, I haven't seen a big difference between the GH3 and GH4 light frames so I doubt the GH3 would fair any better.

Also I tested the GH2 and the GF1 for this object before as well. The GH2 was terrible. The GF1 was actually ok up to ISO 1600. I like the fact that the E-PM2 is ok up to ISO 3200 though.

Sony sensors are widely considered the best, although the DxO scores of modern Panasonic cameras are on par with Sony. The GX7, GH4, and E-M1 all use Panasonic sensors.

The GH3 was the only Panasonic MFT camera to ever use a Sony sensor. The E-M5 and E-PM2, as well as many other Oly cameras prior to the E-M1 used Sony sensors.

The reason why the GH3 would be a good tie breaker is because it's the same body as the GH4 but with a Sony sensor. So, if it performs similarly to the E-PM2, you can conclude it was the sensor. If it performs similarly to the GH4, then you can conclude that there's something unique to the GH. body or Panasonic's processing.

It should be noted that the GH3 doesn't use the same sensor as the E-PM2, but they're both Sony sensors from the same year, so it stands to reason they would perform similarly.