I have been using M4/3 for about 1/2 of my photographic work for almost 2 years now and I have seen and compared the results on my 24" high quality calibrated computer screen. Initially, I sent test files to my stock agency and they said they were fine. I have felt comfortable that the quality was there, but I have never actually compared enlargements side by side to see if any differences I could possibly see at 1:1 on a computer screen translates to any discernible differences in a print.
I use both a Nikon D800E and the Olympus E-M1 (I was using the E-M5 until December when I purchased the E-M1 but they both have the same sensor and same image quality) so let's compare formats numerically first and examine the differences before we talk about prints.
Full Frame
Sensor Size
24 mm X 35.9 mm in size (my Nikon D800E's sensor; other manufacturer's vary every so slightly)
7360 pixels X 4912 pixels or 36.15mp
(If we print an image at full resolution at 250 ppi, that translates to a native enlargement of 29.44" X 19.64". If we print at 300 dpi, the native enlargement would be 24.53" X 16.37".) If you want to go larger you have to either "upsize" the file or print at a lower resolution. Upsizing works very well and you can easily double your files size without much degradation in image quality with today's leading software.
The sensor is 861.6 square millimeters in size.
M4/3
Sensor Size
17.3 mm X 13 mm in size (as far as I know all M4/3 sensors are exactly the same size no matter the manufacturer)
4608 pixels X 3456 pixels or 15.92 mp
If we print at 250 ppi, that translates to a native enlargement of 18.42" X 13.82". If we print at 300 dpi, the native enlargement would be 15.36" X 11.52".)
The sensor is 224.8 square millimeters in size.
[For your reference, APS-C sensors are either 23.5 mm X 15.6 mm or 366.6 square millimeters (Nikon & everyone else except Canon), and Canon's are 22.5 mm X 15 mm or 337.5 square millimeters.), which is why Nikon has a 1.5 crop factor and Canon has a 1.6 crop factor. Canon's sensor is actually slightly smaller. The difference, however, is statistically insignificant and won't make any practical difference in image quality or enlargeability.
Full frame sensors are 2.35 times larger than Nikon DX or APS-C sensors.
Full frame sensors are 2.55 times larger than Canon APS-C sensors.]
The difference is clear. The full frame sensor is over 3.8 times bigger than the size of the M4/3 sensor and at least 2.35 times larger than APS-C! The size, number of pixels and pixel pitch differences in these sensors should produce an easily noticeable difference since it isn't even close. But is that the case? Hmmmm.
Last week I decided to make two identical images, have professional enlargements made and see for myself what kind of actual difference there is good sized enlargements. Now, this was one test to ascertain practical differences under best circumstances. This was not a high ISO test or worst conditions test. So, give it the weight you decide is appropriate.
Here is how I set it up.
Both cameras were:
1) I picked a subject with lots of detail
2) Attached the cameras to a large, very sturdy Induro 410 tripod
3) Cameras were leveled
4) Both cameras set to ISO 200 and RAW
5) Both set on aperture priority exposure at F/7.1 & F/8 (Panasonic 12-35 F/2.8 for Olympus; Nikon 24-120 F/4) (my mistake as I intended for both to be at F/8, but there shouldn't be any difference)
6) both lenses were set at 50 mm or equivalent
7) image stabilization turned off, in the camera (and lens) for the Olympus and on the lens for the Nikon
8) Exposure delay set for 2 seconds set in the Nikon to eliminate any mirror slap (no mirror in the Oly)
9) The first image made with each camera was of a Colorchecker Passport to later match exposure, white balance and color
10) Several exposures were made with each camera
11) In Lightroom 5, both images were adjusted to make the best possible image I could, just as I would in any other case. Color and density of the images were matched. The color profile was that which I made using the Colorchecker Passport.
12) Both images were slightly cropped to match for a 16' X 20" enlargement
13) Both images had my standard "pre-sharpening" as would any other image I edit in Lightroom
14) Both images were exported as JPEGS, sRGB colorspace at full resolution and 250 dpi, as required by MPIX for printing, with no additional sharpening
15) Before exporting, I removed one small detail in one image so I could tell the difference between the two as otherwise, the images appeared identical
16) Both images were then sent to MPIX for 16" X 20" enlargements, with no correction adjustments being made, as is the option on their site.
17) I sat back and waited to see what the results would yield!
(Let me know if you see any flaws in my testing process. As a note, I don't think it is valid to test raw images right out of the camera without any editing as no one, no one, uses a raw image right out of the camera as a finished image. I believe it is better to make the best image you can from each raw, then compare those. After all, I would be making the best images I could if I were using them for submission for stock or any other purpose.
The same goes for JPEG images. Why? I don't believe it is valid to compare JPEGs across manufacturer lines as each manufacturer sets its own default parameters for sharpening, contrast, saturation, etc. An image from one manufacturer could easily look better from an image from another manufacturer as the first may set hight default sharpening, contrast and saturation, making the image right out of the camera look better. You can adjust the JPEG parameters in your camera to match another, if that is your wish. It is better to compare after you have taken what comes out of the camera and edited it to make the best image you can.)
Friday evening, the box with the enlargements arrived by USPS. I anxiously opened the box to find two professionally made and well packed enlargements. MPIX did a nice job as they always do. Since it is winter and the sun had gone down, I first looked at them under the bright, fluorescent lights of our kitchen. But I wanted to see them in a bright setting before making any judgments. Saturday morning, I looked at them again in bright, direct sunlight to maximize the illumination so I could discern any differences while examining both prints.
Results
After looking at the two images under the kitchen fluorescent lights, I could see no discernible difference between the two images.
After looking at the two images under direct sunlight, I still could not see any discernible difference between the two images.
The color, contrast, saturation, brightness, sharpness and tiny details looked identical. In fact, they looked like two images from the same file. They were just as sharp on the edges and corners as they were in the center with the same level of detail across both prints.
Here are the two images.
I'm sure you can't tell any difference at these sizes and on the web. The top one is from the Olympus.
Here are 1:1 crops from both images.
The top image is a 100%, pixel level crop of the Nikon D800E image. Any difference you may see on the computer screen are not apparent in the 16" X 20" enlargement.
My conclusion is that this test indicates to me that under these conditions and an enlargement of this size, the two prints are, for all practical purposes, identical. I know this was taken from near the center of the image, but the edges and corners are identical also.
Now, if you can find a difference in these two images, one has to look SO CLOSELY for any MINUTE difference, that it really makes no difference. If you have to look that closely, then you are not looking at the image itself, but you are looking at minutia.
I may perform some tests in the future with larger enlargements or higher ISOs. There should be a discernible difference with a higher magnification and/or the higher the ISO used. But, at this time, I don't know what that point is where a difference will be seen.
That being said and, looking at my history, only about 3% of my images are made at ISOs of 1600 or more. Ninety-seven percent are made at ISO 800 or less. Is there a difference at ISO 800? Don't know.
I'm satisfied that M4/3 will meet my personal needs and, has already been demonstrated by me over the past 2 years, will meet my stock agency's needs. They have no issue with either native files or up-sized files.
UPDATE: I conducted an additional test using 20" X 24" (50 X 60 cm) enlargements. Here is that post. I think you will find it even more interesting.
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 © 2014 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.
Nice comparisson. But you should work with equivalent F-Stop, 4 at mft and 8 at FF. THat would be better.
ReplyDeleteThen the DOF will be also the same.
You are correct in stating equivalent apertures between the two formats. However, I wasn't necessarily testing depth of field. I had focused on the side of the building and that is where I was measuring sharpness and detail. It turns out, however, that the depth of field from both cameras sufficiently covered the entire scene. So, it was a wash, as they say. All that being said, after looking at the images in Lightroom, I had failed to have the aperture on the Olympus set where I had intended. Thanks for the comment.
ReplyDeleteI do agree there is no difference, and that this is a very interesting and realistic test. I note that the print size you chose is slightly larger than the "native enlargement" size for the Olympus, whereas it is considerably smaller than the native enlargement for the Nikon. So both native images had enough detail for a 16x20 @250dpi. But if you chose for example, twice the magnification through cropping, then I would expect the Olympus image to show noticeably less detail than the Nikon image.
ReplyDeleteThe test was to enlarge both images to a size that would be large, but still would fit on a wall if one wanted to frame and display them–a normal use for most photographers. The test was not to enlarge them in equal magnifications. I wanted to see, if I made enlargements to 16 X 20 or 20 X 24 (in the subsequent post), from the two disparate formats and sensor sizes, would there be a noticeable difference? In this case, there is not a visible difference, even pixel peeping. I would have no use for an image, say twice the magnification or a D800E file, for a wall or any other normal use. So, I kept the enlargement size realistic for a use such as a framed wall display. Additionally, these files were saved at 300 ppi, which is my standard for my stock photography submissions.
ReplyDeleteHi Dennis,
ReplyDeleteVery interesting. I suspect that in your test the lowest common denominator is either when you exported the jpgs or at the printing and that is why the prints are the same. Had you tried to obtain the best from each format the result may have been different. Do you have details of what the print company does? What were the prints i.e. glossy/matt/paper/canvas? As I am considering buying an OM-D camera it is encouraging to see the quality available (a 16" x 20" print is the largest I am likely to want).
Bill
I 'm not sure I can concede that the prints may have been better if they were straight from my raw files. However, I used MPIX for the 16 X 20" prints and Bay Photo for the 20 X 24" prints. MPIX requires JPEGS with an sRGB colorspace so that is how I sent both sets, so all things would be equal. I can't print that large at home, so I sent the files out. I have to say, in 44 years of serious photography, many of those years as a professional and currently shooting stock, the 20 X 24" print is as good as anything I have seen. I'm pretty obsessive about my photography and I can't find anything to pick about in either enlargement. But remember, this is low ISO on a sunny day–i.e., the best possible conditions. A higher ISO or low light would have been much different. But I was really looking at best possible circumstances, not worst conditions.
ReplyDeleteThe whole point is that FF can allow you to stand much closer than m4/3. You are missing the entire point, which is FF means shallower DOF than m4/3.
ReplyDeleteI was able to identify the Olympus images immediately because they have better colours.
ReplyDeleteThis is bright sunlight and FF sensor does not use a fraction of its power. Pls compare the same pictures taken in dim moonlight with ISO >10000.
ReplyDelete-- M.L.
Any difference in colours is due to slight saturation differences in PP. I have a Panasonic GH4 and a Pentax K1, and I too can detect no IQ difference at base ISO. But the GH4 image quality drops dramatically against the K1 as ISOs rise.
ReplyDeleteSo I would always use a wider GH4 aperture when possible, and the greater DOF makes this possible.