Own An E-M1? For Low Light, Is It Better To Increase The ISO Or Just Underexpose And Adjust In Your Editor?

Indoor test subject under incandescent light; E-M1 w/12-40mm f/2.8 lens; ISO 200; 0.5 sec. @ f/8 (click to enlarge)

This image was properly exposed and shows no appreciable digital noise.

(I threw in the 30 year old Kodak Gray Card just for kicks.  Obviously, it has faded over time.  Auto white

balance was very close but I tweaked it using the ColorChecker Passport grayscale.)

If you own an Olympus E-M1, do you know if you will you get a better image by raising your ISO when needed or keeping your base ISO and then brightening your image to taste in your image editor?  I was wondering that myself so I decided to find out.  I did the same test with my X-T1 which you can find here.

I wanted to test my E-M1 to see what differences I could expect by raising my ISO as necessary for any particular conditions as I normally would or just underexpose the image at ISO 200, then brighten the image in Lightroom.  I set up two tests, one indoors and one outdoors.  I placed the camera on a tripod, turned image stabilization off, put it on manual exposure, set auto white balance, set it for raw file format and manual focus.  I planned on testing ISOs from 200 to 6400, the range which I would normally use the camera.  I don't normally use JPEG format nor do I use ISOs greater than 6400, so I didn't test that file type or settings higher than ISO 6400.

In order to keep this blog post relatively short, I will only post the images from the extreme, comparing ISO 6400 correctly exposed and ISO 200 underexposed 5 f/stops.  The results I saw seem to somewhat apply to these intermediate images as well as the extremes.  I will explain the differences I saw in the intermediate ISOs tested.

Disclaimer:  These tests are informal and non-scientific and only apply to my particular sample of the camera and under these particular conditions.  Your camera and/or conditions may vary, so just take the results for what they are as one particular example and not valid for all cameras and all conditions.  I don't have any sophisticated testing instruments.  I can only visually look at the results and assess what I see.  I'm just a photographer trying to better understand his camera and how it reacts in different situations and want to share what I find.

ISO 6400 correctly exposed for the light level; 1/60th sec. @ f/8 (click to enlarge)

This image was 5 stops underexposed then brightened to match the properly exposed image in Lightroom 5.7.1.

The camera was set for ISO 200 but the exposure required ISO 6400 (click to enlarge)

1/60th sec. @ f/8

ISO 6400 properly exposed (click to enlarge)

ISO 200, 5 stops underexposed, then adjusted to match correct exposure in Lightroom 5.7.1. (click to enlarge)

As you can see there is some difference in white balance, digital noise, contrast and resolution of fine details.  Any minor differences in white balance, contrast or saturation can easily be matched in your image editor.  Even though the digital noise level is slightly more in the underexposed image, when normally working to reduce the noise in your image editor the differences are slight and won't make a practical difference in the final image. 

However, and this is an interesting however as I did not see this in the X-T1 tests, the histograms are very different showing "combing" or a loss of data in the lower tones with the underexposed image.  This did not appear in the properly exposed image nor in any of the images in the exact same tests with the X-T1.  Since in this particular scene, those lower tones are mostly black, you won't see any loss of data in the image itself.  But what about other images?  

Here is the outdoor scene.  There are similar results, however, the 5 stop underexposed image shows a bit more contrast and the auto white balance shows a red shift which gets greater as more underexposure occurs.  Other than that, the results are similar to the indoor scene.

ISO 200

ISO 6400 properly exposed; color noise reduction only applied (as default in Lightroom) but no luminance noise reduction
(click to enlarge)

ISO 200; 5 stops underexposed then adjusted to match ISO 6400 exposure in Lightroom 5.7.1. (click to enlarge)
Again, only color noise reduction applied.  The explanation as to why I applied it is in the text.

With a comparison of ISO 400 and 200 (one stop underexposed), in this particular example, the underexposed image actually looks better to me.  Go figure?  The digital noise is tighter and there seems to be just a tad more detail.  For brevity, I have not illustrated that particular test here.  In the ISO 800 versus a 2 stop underexposed image, they look equal to me in all respects.

At ISO 1600 and above, I see a definite red shift in the underexposed images.  This may be a function of the auto white balance.  In any case, the slight shift, which gets worse as you further underexpose images, can be neutralized in editing software, but it does occur.

As in the ISO 800 image, the ISO 1600 image versus the 3 stop underexposed image look about the same, but the properly exposed image seems to have a bit more fine detail compared to the underexposed image.  Just a tiny bit if I look really close, but nothing you would notice by just looking casually.  You really have to look closely to find that small detail.  I would say, for all practical purposes, they are the same.

At the ISO 3200 and 6400 settings versus the 4 and 5 stop underexposed images, I see a definite difference in noise and detail in the favor of the properly exposed images.  The ISO 6400 image versus the ISO 200 underexposed image is illustrated here.

In this outdoor scene, the "combing" and loss of data in the histogram extends farther toward the middle tones.  So there is even more loss of data than in the indoor test image.  That is not good.  I see that in every image down to ISO 800.  At ISO 400, the histogram in solid and shows no data loss.  Again, the histograms technically show a data loss, but I don't think the loss is something that one would notice in most images unless you were doing some heavy editing.  However, I wanted to mention it so you are aware of that fact.

From what I have seen from these brief, informal, non-scientific tests, I can only conclude that, unlike the exact same tests with my Fujifilm X-T1, there are differences between properly exposing an image at a high ISO versus keeping your ISO at 200, thus underexposing and subsequently brightening the image in your image editor.  This must reflect a difference in how Fujifilm and Olympus has designed their systems. At this point, I'm not sure why or what those differences may be, but the two systems react a bit differently when tested in this matter.  

The other major difference I found between the two systems is that there was absolutely no color noise (chrominance noise) usually showing up as red/blue blotches in the X-T1 images even in the 5 stop underexposed image.  I have never seen that in any other camera I have owned and tested.  I'll make a leap of a conclusion here, and I may be wrong, but I can only assume Fujifilm is applying color noise reduction (and it may be luminance as well as chrominance) to their raw files in camera and the user has no control over it.  I have left the default chrominance (color) noise reduction setting in Lightroom 5.7.1 intact for the E-M1 files for this test so I could make a valid comparison against the X-T1.  If I scale it back to 0, there is a lot of color noise in the E-M1 files as you would fully expect.  Since Fujifilm has already removed it, I can't adjust it out in the same way I can with the Olympus.

You decide if you reach the same conclusion as I have.  I'm sure there are some niggles in differences in conclusions, but overall, the images between proper exposure and underexposure are pretty close.  Looking at the images in their entirety and not pixel peeping, they look identical with a bit of tweaking in my editor. I really don't think it makes much of a difference as to whether you set your camera to properly expose at, say, ISO 3200 or underexpose at ISO 200, then adjusting the exposure in Lightroom or another image editor.  

So, don't worry about making a mistake and accidentally underexposing your images, go out and enjoy your gear and make some photographs!

I have performed the same tests for the Fujifilm X-T1 and posted those results in my last blog entry.  If you are interested, you can find it here..

Thanks for looking.  Enjoy!

Dennis Mook

Many of my images can be found at www.dennismook.com.  Please pay it a visit.  I add new images regularly.  Thank you.


All content on this blog is © 2013-2015 Dennis A. Mook.  All Rights Reserved.  Feel free to point to this blog from your website with full attribution.  Permission may be granted for commercial use.  Please contact Mr. Mook to discuss permission to reproduce the blog posts and/or images.