I have very high expectations for image quality. In fact, I have high expectations for almost everything. I am sort of a perfectionist, but I've worked hard to be satisfied with less than perfection, with some success. I don't think there is anything wrong with having high expectations in every aspect of life. But that is just me.
In 95% of my M4/3 images, I am very satisfied with the quality. However, every once in a while, with my M4/3 gear, when I return to my photo lab and bring up an image in Lightroom to critically examine it for focus, noise, etc., I find myself disappointed. Your mileage may vary. This eagle image is one of the times. Also, I will say that I have never been disappointed with any image I've made with any of the full frame digital gear I've owned (Nikon D700, D800E, D810).
About the image and for some background, this image was made with an Olympus OM-D E-M1, Olympus Zuiko 50-200mm f/2.8-3.5 SWD lens, ISO 800 at 1/1000th second @ f/8. I also used the Olympus 1.4x tele-converter. I had previously tested this setup to be very sharp and have high resolution. In other words, there is nothing lacking with the camera, lens, or converter. Also, I made the image using the RAW format.
The gear was all locked down on my very large, heavy and sturdy Induro 410 tripod (not even fully extended) with a RRS B55 Ballhead. About as sturdy as one could get.
The image was manually focused using the 10X magnification in the E-M1's viewfinder and is in sharp focus. In body images stabilization was turned off. The image should be as good as it could get.
The light was 5/8 to 3/4 backlit from the left. The exposure was spot on and easily in between the sides of the histogram so no major adjustments needed to be made. The only tonal adjustment was to slightly lighten the head and body of the eagle so more detail in the feathers was visible.
This is a very sturdy setup. I believe the entire process showed good photographic technique. I'm not sure what I could have done differently to improve the image.
In the final image, the eagle was still too far away for the equivalent of 566mm (the combination should produce an equivalent of 560mm but the metadata shows 566mm equivalent) so I had to crop the image to make some sort of pleasing composition. I just didn't have anything longer to put on the camera. I wish I did. You see the resulting crop. It is not a great shot by any means, but it serves well to illustrate my point. I was hoping to make something aesthetically pleasing out of this, but it is more of a record shot, I guess.
At this point, I felt I did everything as best I could to maximize image quality. I'm not sure what else I could have done. Now it was up to the sensor to produce a high quality image.
Here are some additional images to better understand to what I am referring.
 |
| This is an out of the camera JPEG conversion as shot, no cropping. (click to enlarge) |
 |
| This is a 100% crop of the eagle. Notice the noise when you enlarge it. The pine needles, bark and pine cones are very sharp and detailed. No complaints with them. But the eagle doesn't look as detailed as it should and the head is very noisy. (click to enlarge) |
 |
| This is my best attempt at grabbing a bit more detail from the brown as well as white feathers and adding some noise reduction without losing any more detail in the latest version of Lightroom. Any more noise reduction obliterated the little that is left in the feathers. Should I be happy? The image is very sharp and well exposed but noise for ISO 800 should be better than this? (click to enlarge) |
I am now entertaining thoughts of ending my M4/3 experiment, divesting myself of the M4/3 gear and moving up to APS-C, which is probably the best all around format considering size, weight, price and image quality. I think it may be the best all around compromise format. Then I think how much I love using my E-M1 and my M4/3 gear and how it is only going to get better in the future. So, there is some internal conflict. When I have internal conflict, I don't do anything until I mentally resolve that conflict.
I'm willing to try the Fuji X cameras again, but I'm not sure anything has changed since I last tried them a year ago. From my previous tests, the Fujifilm X-system would probably be fine for anything but landscape work, in my opinion. However, things may have changed with Lightroom's ability to better demosaic the raw files as well as Fujifilm's continuous upgrade of firmware. I just don't know at this point. I'm also willing to look at Sony's cameras, but I'm not sure they have the fortitude to fill out the lens line in a robust and satisfactory manner.
Probably the best thing for me to do at this point is "chill," be satisfied with 95% and see what is on the horizon with Olympus, Panasonic, Fujifilm and Sony. If only Nikon would come out with a mirrorless 24mp D750-like camera with a 3 million dot EVF! I definitely would have to jump into that.
Your thoughts?
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.
Dennis, Isn't 95% good enough when you consider the size/weight benefit of the m43? Of course you want to be sure to have your FF gear around for the other 5%. My standards are lower than yours, but I experience the same results. Both my LX5 and my Sony 5R with not-so-good 16-50 kit are "good enough" 95% of the time because at least with me 95% of my images are frankly not much more than "snapshots" (family pix, record shots, I was here shots). For the other 5% I have my EM1 + 12-40 or 35-100 zooms.
ReplyDeleteEven if APS-C gave you a better result in your eagle shot (likely no more than one stop of better ISO performance) is it worth the weight. I just looked at the specs for the new fuji 16-55 F2.8 and it is something like twice the weight of the oly equivalent 12-40 F2.8 but otherwise seeming to be a comparable level of quality and features. Plus I am not convinced (at least with zooms) that the fuji are as sharp around the edges.
All that being said I have enjoyed greatly the ergonomics and haptics of the Sony a6000. The sensor is 1.63x bigger than the FT/mFT sensor but the sony, at 24mp has actually 1.5x more pixels. This means that each actual pixel on the a6000 is slightly bigger than each pixel on the mFT 16mp sensor. A light went off when I realized that even if I cropped the 24mp image to 16mp, it would still have covered a larger surface area of sensor than the 16mp on the Oly sensor. Hope I explained myself well enough here.
Anyway... just some more food for thought.
Oh, and if you are shooting in other than wide open, and you are shooting in Aperture priority, the difference in noise might not exist. For example, EM1 at F8 at 1/250th at ISO800 would be matched for DOF on an APS-C at F11 at 1/250th, necessitating ISO1600.
And finally... I'd hate to sell all those great mFT and FT lenses, when the next version of the OMD (EM-1mark2) will likely have better ISO performance... as there is no stopping technology.
Peter, I don't disagree with anything you wrote. As of now, I have no intention of selling any of my M4/3 gear. I really like using it, but I would really like 100% of the images I make with it to meet my standards. Now, that is not realistic, but that is what I would like.
DeleteI think you hit something I'm coming to believe is important. An APS-C sized sensor with 24mp may be where my mind is going. That may be my "sweet spot" for detail, size, quality and cost for cameras, lenses, etc. Sony to me is not yet viable as they don't have a lens selection comparable to M4/3. I'm hoping they keep developing the E and the FE line of lenses.
I've always set my expectations very high and have always been disappointed in myself, others or in things that don't meet those expectations. I guess that is how I rose to the top of my profession, twice, and came as far along in photography as I have.
Thanks for the comments. I'll be watching the industry to figure out what I'm going to do, if anything. More on Olympus and M4/3 in the next couple of posts.
Dennis
Dennis, I absolutely agree with your thoughts!
ReplyDeleteI'm through with similar thinking and acting: I owned/tried M4/3, Sony NEX, Fuji X, Canon plus Nikon APS-C and Canon plus Nikon fullformats. Again and again I hoped that compromizing where possible - but nope.
Finally I'm sure that not even APS-C helps my with my equalling expectations: I don't look for the vage chance of minimizing weight … I shoot Nikon Full. I carry, I "schlepp", I sweat – but now I'm constantely happy with the results!
Jens, from Germany
Sigh. I've been around long enough to have run into exactly these problems with 35mm film. I slowly mastered the medium, grew frustrated by either having to run very slow film on a tripod or not getting the quality I wanted, then moved to medium format (Fuji 69 and then Pentax 67), only then to move to digital 10 years ago and finally settling on m4/3 as my camera of choice. It's the same story with m4/3 as it is with 35mm: for 95% of all shots, it's the bee's knees, the cat's meow. But for that last 5%, it falls on its face and can't deliver.
ReplyDeleteYou know what? I've gotten past the point of worrying about it. I recently took photos at the office Christmas party, miserable lighting and exposed most everything at ISO 25600, using either a Nikon 85 f2 or an Oly Zuiko 180 f2.8, all hand held. It worked wonderfully. Could I have used a Sony A7S instead? Of course, but who's going to give me $2k for the body and a set of batteries so that I can take 800 pictures in the course of the evening, all candids with no one realizing they were being photographed, capturing the spirit of the party and making all involved happy.
APS-C is simply too close to m4/3 to make that much of a difference. Moving (back) to FF would make a difference. I have too many Oly and Panny lenses at this point to be willing to move to the Sony A7 and its variants, especially since moving to that FF sensor places enormous demands on lens quality.
Nope, I'll stick to my m4/3 setup (EPL2, EPM1, EM1 and GF2, with 9 BCL, 14 f2.5, kit zoom, 45-200, along with E510, E30, 12-60, 50 macro and 70-300. I usually only have time to get out and take photos when I am travelling and that means it needs to be portable. I put the EPL2/VF2, EPM1, EM1 and GF2 along with 9 BCL, 14, kit, 45-200, then add to that the 70-300 4/3 and a 600 f8 Series 1 Solid Cat via adapter, all in a Kata 3N120. I literally can't do that with any other system (okay, chargers and the like are not in that bag). Hiked Zion, Bryce, Arches, Yellowstone and Yosemite with that setup and yes, it was heavy, but yes, it was doable. Even got nice hand-helds with the 600 on full stabilization.
For serious landscape work: consider a Gigapan robotic tripod head. I've dedicated my 4/3 gear to that project. Mosey on over to www.gigapan.com and search on the gigapans by JohnF: not all are wonderful, but I've sold a few (largely of New York) and they've been printed at 2x6. Not feet, meters. 1200 DPI even...works.
Right tools for the job: you had 566mm equivalent and it wasn't enough to avoid cropping. A 400 f6.8 Leica "trombone" would have given you 800 equivalent without cropping. That's why I have the 600: for the times you need that reach, it's the only way to go. :-)
No offense, but I don't think it was system failure; I think it was photographer failure. What did you expect when you cropped a 4/3s image that severely? You have to use a m4/3s system within its limitations. You can't make it walk on all fours. If you can't live with those limitations and are willing to carry the weight penalty of a full-frame system, then all I can say is better you than me! Or perhaps Fuji will make enough changes to provide what you're looking for.
ReplyDeleteDave, I expected better than I got. I've cropped many other images, with even smaller sensors, in half and the results were better. As I said, this seems to be an anomaly which raises its ugly head only on occasion. I'm not sure why. Most of the time, no problem.
Delete[This may be a second entry for this comment, and I may have made the first one as a "reply" to Mr. Jenkins, which it is not. My internet connection is behaving oddly right now.]
ReplyDeleteDennis--
Tsk, tsk. All this discussion of sensor size when the problem was that your model hadn't shampooed his (or her) gritty headfeathers for a week or two. It's grit, not noise, and the E-M1 captured it perfectly!
More seriously . . . my first thoughts when I clicked up the photo was "wow, look at the great detail in those sunlit pine trees." But then the eagle. I've been pondering, in addition to the photo, two sections of your description of the problem:
"The light was 5/8 to 3/4 backlit from the left. The exposure was spot on and easily in between the sides of the histogram so no major adjustments needed to be made. The only tonal adjustment was to slightly lighten the head and body of the eagle so more detail in the feathers was visible."
"So what is the problem? Noise. Using a good exposure I shouldn't be getting the amount of noise I was getting in the eagle at ISO 800. I really shouldn't be getting any appreciable noise at that ISO. Let me state up front that I think this is an anomaly. I only see this on occasion and I don't know why."
I guess it would be good to at least be able to know in the field when the anomalies will rear their noisy little heads. Here, I'm wondering whether the anomaly arises when looking for detail in a shadow area which we "know" is "really" white? From the camera's point of view, the white head was in shadow, lit by what appears to have been a broad blue sky, so the head, properly exposed ("spot on," as you say), should be a dull bluish grey, and it is. So if we want it to be white, or at least significantly whiter, and to bring up the detail, we use an adjustment brush in Lightroom. Pulling up noise may be like pulling up noise in larger shadow areas by too extreme increase via the LR sliders. (Am I right that shadow exposure can be brought further up without noise in photos taken with full-frame sensors?)
So I'm wondering what would have happened if you had exposed the original image for the expected/desired white of the eagle's head? The rest of the image would have been overexposed, but too much to have been dragged down in LR? (My question is related to the idea of ETTR, as well as to the kind of exposure adjustment we need to make in snow scenes.)
The snag the eagle is sitting on must also be noisy, but it's exposed the way we expect for a shadow area, so we don't care. If it had been an osprey rather than an eagle sitting on the snag, would the problem have arisen, at least so obviously?
I don't claim to know what I'm talking about, but you asked for our thoughts! (And I appreciate your always-thoughtful posts.)
Walter Foreman
namerof@uky.edu
Try DXO Optics Pro 10 to clean up that noise and detail. It solves most M43 problems nicely, if you underexpose the highlights a bit and use only the "Prime" noise reduction and "Lens softness" features at their default settings. The generally higher sharpness and QC of affordable M43 lenses tends to even out resolution compared similar APS-C lenses, whose sharpness tends to be a bit more variable.
ReplyDeleteOh, and I might mention that when new firmware is imported, it inherently overwrites the old ROM instructions and the values of the configuration registers. This is virtually a given if you want a clean and reliable upgrade. Rechecking your configuration after an upgrade is as much to be expected as resetting your camera to your normal ISO and other configuration settings after changing the settings for a specific circumstance.
I' familiar with DXO Optics Pro and the software has intrigued me. From what I had read, their noise removal engine is not equalled. I may try it. Thank you for the suggestion.
DeleteDennis, I don't think it is an anomaly.
ReplyDeleteIt is a low contrast situation with the eagle's head showing as a faint grey. Even your Nikon 810 would struggle with it - especially if the lens is not equal to the superb contrast of the Zuiko.
See: http://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=nikon_d810&attr13_1=oly_em1&attr13_2=canon_eos5dmkiii&attr13_3=canon_eos5dmkiii&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=800&attr16_1=800&attr16_2=800&attr16_3=800&normalization=full&widget=1&x=-0.0001869246226460243&y=0.5073284934205087
Besides: Lightroom prefers a little more noise to loss of details as you can see when comparing the in-camera jpeg.
Also: Why did you use ISO800? Would F4 or F5.6 not have given enough depth?
Mike, I used ISO 800 so I could keep my shutter speed at a high level in case the eagle flew from the top of the dead tree. However, he was content to just sit there. I did see his/her mate fly in to the nest which was about 10 yards to the right of where this eagle was perched.
DeleteDennis,
ReplyDeleteI own a PEN E-P3, which is at the same time the best (digital) camera I ever owned, so the hardware is not exactly comparable. Still I have tried quite a lot to make the best out of available light situations.
My experience in terms of noise is, that there are differences relating to the "brightness" of a picture. What I want to say that a picture which is shot at a higher ISO Setting may be totally satisfying where it is brighter (high key Areas), but just muddy and noisy in the darker areas. You may see what I mean if you compare the brightly lit parts of the trees bark to the chest of the eagle.
My experience is that you could have avoided the noise on the eagle's face and chest by exposing the picture longer or even by chosing 1.250 ISO instead of 800.
That would have been at the cost of structure in the sky/Background, which might have become completely white.
An experience that I share is that the amount of noise doesn't seem to be completely foreseeable. There seem to be factors apart from ISO and brightness lying deep in the camara.
Greetings from Cologne, Germany
Thomas
Dennis,
ReplyDeleteI may correct my previous entry a bit. Choosing a higher ISO would of course not have reduced the noise as long as the whole picture wouldn't have been brighter. That, of course, would have affected also those parts of the picture which are already bright (bark, sky) and have led to a loss of structure there. The overall conclusion is that you could have avoided the noise on the eagle's chest but the rest of the Image would have suffered.
Greetings
Thomas