A Hands-on Field Report
Olympus is an interesting company. They have been making high-quality 35mm cameras since the early 1970's, when the OM-1 was introduced. This was the beginning of a professional grade camera system that lasted until very recently. The hallmarks of the Olympus line were high-build quality, small size and light weight, and exceptional lens quality.
Though Olympus never gained the market penetration among pros that Nikon and Canon did they served a discriminating niche market that was always very loyal, and who appreciated the quality products that this small manufacturer offered.
Olympus also always danced to a different drummer. Beginning in the early 60's they produced a line of half-frame 35mm cameras that many photographers (myself included) embraced because of their ultra-small size and high build quality. These were today's equivalent of the high-quality digicams that many photographers, including pros, use for non-critical applications.
Enter the E1
In fact, the digicam arena has been where Olympus has been focusing its attention for the past several years. Their E-10 and follow-up E-20 models have been among the most popular high-end digicams for the past couple of years.
In late 2002 Olympus announced the new Olympus E-1 camera and system, the first 4/3 format camera. Now, a year later cameras have started to ship, and with the kind assistance of Vistek, Canada's largest professional imaging equipment dealer, I have had the opportunity to do a hands-on evaluation of a full production camera.
Before diving into my impressions and test results, let's understand what the 4/3 format is. Firstly, it's pronounced "four thirds". The name has historical roots, which aren't terribly important, but it's worth noting that the aspect ratio of the format is indeed 4/3, less rectangular than traditional 35mm's 3/2 format, and similar to medium format 645 in aspect ratio.
4/3 was announced by Olympus and Kodak at Photokina 2002, and now a year later these are still the only companies backing the new standard, with Olympus shipping their new E-1 system and Kodak making the imaging chips. Fuji has also "expressed interest" in the new standard, but that's about the extent of it. In addition to being a new digital-only format the 4/3 standard applies to lens mounts as well. If another company were to support the standard then their lenses and those from Olympus would be interchangeable— something that one can only hope for, but I doubt that we'll actually see.
Yellow = 35mm full frame film or digital (Canon 1Ds / Kodak 14n)
Red = Reduced-frame digital (Canon 10D & 300D / Nikon D100 & D2 / Fuji S2)
Green = 4/3 format ( Olympus E-1 )
Fig. 1 above is a quick sketch showing the relative size of full-frame 35mm, reduced frame digital, and the 4/3 format. It's clear from the illustration that the 4/3 chip is smaller than the sensor size found on current reduced-frame cameras. It is significantly smaller though than traditional full-frame 35mm.
Of course chip size has little to do with pixel count. The E-1 is a 5MP camera, whereas most of its reduced-frame competitors are 6MP, but that misses the point because it could just as easily have a high pixel count by making the pixels themselves smaller.
By way of comparison, the Kodak sensor in the E-1 uses 6.8 micron sized photo-sites (pixels). The Canon 10D, has 7.4 micron pixels, while the Canon 1Ds has 8.8 micron pixels. These are all dramatically larger than the roughly 3 micron sized pixels found in most digicams. For those that haven't been paying attention — pixel size is one of the most significant determinants when it comes to image quality. All other things being equal, the larger the individual pixel size the cleaner and less noisy the image produced, and low noise is one of the great advantages of digital over film.
So now we see the conundrum. Make the pixels smaller and you have higher resolution from the same surface area. But, you make the image noisier and introduce other aberrations. The laws of physics and optics are the limiting factors here, not technology. So by choosing to go with the sensor size that they have Olympus is betting that they can increase pixel count without adversely affecting image quality. This may be the case, but if it is possible it also applies to other chip fabricators as well.
Therefore there has to be a rational for the smaller pixel size and the potential limitations that it places on the 4/3 format. According to Olympus it's the ability to make the camera itself small and lighter, and also the lenses. Later in this review we'll see how well they have succeeded in this.
Only Kodak with the DCS 14n and Canon with its EOS 1Ds have yet produced full-frame 35mm digital cameras. Other manufacturers, such as Nikon and Olympus, claim that there are inherent limitations and liabilities in using lenses and lens mounts designed for film with digital sensors. The foremost of these concerns is that wide angle lenses in particular send rays from the edges of the lens onto the image sensor at an oblique angle. With film this isn't a problem. But, a digital image sensor isn't an almost 2 dimensional object the way a piece of film is. It has depth. Indeed each actual pixel site sits at the bottom of a small well. Light from the center of the lens shoots down into the well. Light coming in at an oblique angle, from the edges of the lens, doesn't get down to the bottom as cleanly, leading to vignetting and chromatic aberration.
The solution that some company's chip designers have come up with is to put what are called micro-lenses above each pixel site. These serve to refocus the light rays so that they are better able to enter the wells and illuminate the silicon at the bottom.
So. Who's right?
Damned if I know. What I can tell you is that I have now (November, 2003) shot some 10,000 frames with my full-frame digital Canon 1Ds. Many of these have been with the Canon 16-35mm f/2.8L and also the Canon 15mm f/2.8 full-frame rectangular fisheye. Have I seen vignetting? Yes, occasionally, but not more than I recall seeing on film. Have I seen chromatic aberration. Yes, a bit, but nothing that is too bothersome, and it usually can be corrected easily in software. Again, likely not a whole lot more than would have been seen on film.
Parenthetically, I have also recently been testing full-frame medium format backs and have not experienced any of these issues in this format either, even with the widest angle lenses available.
If you look down the throat of an Olympus E-1 what you do see is that the relative size of the lens mount opening and the size of the sensor are such that it will likely be much easier to design retrofocus wide angle lenses that deliver optimum performance. Will the benefit be more theoretical than real? I can't say. I'll leave it to more technically rigorous reviewers than me to answer that one.
Olympus E-1 with 50-200mm f/2.8-3.5 lens @ ISO 100
Open the box and pick up an E-1. One's first impression is of quality materials. The second impression is that it feels right. The way a camera comes to hand is critical to its usefulness as a photographic tool. Many factors come into play, and some of them aren't immediately obvious. But, if you've spent enough time using a wide variety of cameras over enough years, when a company gets it right you know it right away, and Olympus has gotten it right with the E-1.
The design approach that Olympus took with the E-1 owes little to tradition. I have the feeling that the engineers started with a clean slate, and it shows. But the shape and positioning of the human eye and hands are a constant, and so the E-1 isn't dramatically different than cameras that have come before. An experienced photographer will be able to become familiar with the controls in short order. I was a little less enthusiastic though about the placement of some of the controls. Many of the function selection buttons are strewn around the top and side panels of the body in a seemingly random manner. In the time I had the camera for testing I never really became familiar with their placement, though I'm sure that over time I would have.
The camera's surfaces have just the right textures to ensure that one's grip is secure. Areas where one holds the camera are properly textured, and even with sweaty palms the right hand grip offers necessary support. It also is deep enough for even the longest fingers, and the gap between the grip and lens is wide enough so that one can use the camera comfortably when wearing gloves. Part of the reason for this is that unlike in a camera body which owes its original design to the need to handle film, the E-1 places the lens to the left of the body. Nicely done!
Olympus is touting the E-1 as a professional camera. Be that as it may, they certainly have taken the needs of the pro to heart when it comes to control actuation. Almost every single control has an interlock. In other words, to make a change one must press a button and turn a control wheel. This means that settings will not be made accidentally. But, this is not the case with the focus mode control on the front-left of the camera. It switches from Continuous, to Single shot to Manual focus with a non-interlocked lever. Wearing gloves I found on several occasions that day that I had accidentally moved this lever to the wrong position, and in the cases where I expected Autofocus mode and it was accidentally in Manual focus mode, I missed some shots. Back to the drawing board — for this lever at least.
Olympus rates the body as "splash resistant". This is because all body openings have rubberized seals and tethered caps. From the look of things I wouldn't hesitate to use this camera in the rain, snow or in dusty and sandy conditions.
On the top panel there is a canted and anti-reflection-coated monochrome LCD settings screen. All the information one needs is there, and easily viewable, except for the ISO setting. Like many camera the E-1 doesn't show you the current ISO without pressing a button. I find this terribly annoying, because one of the strengths of working digitally is being able to change ISO between frames. I do it all the time. But, without being able to tell at a glance what the last ISO setting was I often found myself shooting at the wrong speed. Maybe it's because I have become used to the always-available ISO read-out on the Canon 1Ds, but I bitch about this on my Canon 10D as well.
Moving to the back of the camera one encounters the colour LCD. This is a bright and as sharp as any I've seen. I found the menu structure a bit unorthodox, but not something that one wouldn't get used to. I shrugged a bit when I saw that the camera ships with a removable plastic screen cover. Who uses these things, and why? After thousands of hours of use in some pretty nasty environments I have only ever seen a few small scratches on the screens of cameras that I use, and even then, so what? I don't get it, but if it appeals to some users, that's fine. At least it's removable.
Olympus E-1 with 14-54mm f/2.8-3.5 lens @ ISO 100
It's in the Cards
The E-1 takes all types of Compactflash cards, including Microdrives and even the latest 4GB Microdrives from Hitachi. The card door has a twist lever interlock and the cards load from the side. The opening is quite wide and the cards eject smartly — in fact maybe a bit too much so. Watch that the card doesn't pop out onto the floor.
The camera can be set so that a post-exposure image review is available. But, as with the Pentax *ist D that I reviewed last week, there is no post exposure histogram available. What is it with camera makers? Don't they get it? The in-camera histogram is the most important exposure tool since the TTL meter was introduced in the 1960's. Setting exposure properly for digital makes a huge difference in image quality, yet by leaving it out Olympus (and others) are denying photographers a critical tool.
I'm convinced that camera engineers aren't photographers, otherwise we wouldn't see such gaffs. I can see it now — a Sushi bar in Tokyo, some time in early 2002. Engineers from Pentax and Olympus are sitting around having a late evening intramural saki....
Pentax Engineer: Looks like we're not going to display a post-exposure histogram on our next DSLR. Marketing says that some photographers want it but the rest of us in engineering don't see why it's important. Do you?
Olympus Engineer: Nope. Me neither. I take pictures of my kids on birthdays and holidays and have never seen the need. My wallet-sized prints always come out fine.
Pentax Engineer: My snapshots from our summer vacation were great too. The hell with it. It requires at least 4 lines of new code. With debugging it might take a whole day to complete. I don't know about you, but we've got to ship before Christmas 2003. Save a day here, a day there. It adds up. You know how it is.
Am I being too harsh? I don't think so. I don't know of a single professional photographer working digitally that doesn't regard the histogram as the most important exposure tool available after the light meter. (If you aren't aware of what a histogram is or does, please read my tutorial Understanding Histograms. Also also please read Expose to the Right for a better appreciation of why they are such important tools.)
Fortunately there is a work-around. Set the camera for at least 10 second review and then when the image appears press the green Display button. This will call up the review display a couple of seconds later. Now press the Info button. If you've programmed the Info button to display a histogram it will now show up. This takes two button presses and a few extra seconds, but it does solve the problem.
While I'm on a rant, here are two more related complaints. The histogram that is displayed is coarse and blocky, making it difficult to interpret. Olympus engineers should look at the histogram displays that some other manufacturers can produce, for guidance. While they're at it, they should make it state-of-the art by separately displaying the three colour channels.
The second (and last) of these items is that the image review display can be set to display several different types of information screens. I found the programming very confusing. I read and reread the owner's manual page on this several times but it never worked the way it was supposed to. It seemed as if it would with one image, but then not again with that same image at another time. With a fresh one it would. Totally strange, and eventually I gave up. Either there's a bug in the firmware or my sample camera was faulty.
The E-1's autofocus works relatively well. It only has three points but I rarely had difficulty locking on when the subject was contrasty. In low light there's a built-in infrared assist beam that seems to do the job. At least at moderate distances. I did find that the autofocus wouldn't lock as well as some other recent cameras when the subject lacked contrast.
Virtually every exposure made with the built in metering was spot on. I tried the intelligent metering (named ESP for some reason), averaging and spot and all did a fine job.
Update: Due to an editing error my comments of the E-1's viewfinder was accidentally omitted from the publication of this report. They now appear below...
I had expected the E-1's viewfinder to be less bright than it was. But in comparison with the Canon 10D it appeared to be comparable, though a bit smaller due to the reduced frame size. There's little to praise or criticize here.
Battery & Accessory Grip
As with most contemporary camera there is a medium sized removable lithium ion battery in the grip. I didn't have enough time with the camera to do a rigorous battery depletion test, but it appears to offer very good shooting life.
The strange thing is that the LI battery used in the accessory grip is a different size. I don't find this to be a sensible solution. On the Canon 10D, for example, the same battery fits into the camera as well as the grip. In fact the grip takes two of them. This makes both working with and without the grip straightforward. On the Olympus if you've been working with the grip and then want to reduce bulk and weight by removing it you'll need to have the original small battery handy as well.
The battery grip otherwise does the job well, improving vertical handling as well as doubling battery life.
Olympus E-1 with 50-200mm f/2.8-3.5 lens @ ISO 100
The E-1 does not have built-in flash. This helps to keep the size down but some photographers may miss it. I don't work with flash often, except for wildlife where I use a large external flash, often with an Extender for distant reach. But I'll often use the built-in flash on a camera that has one as a fill-light, and find it quite handy to have at times. Even the $6,000 Hasselblad H1 medium format camera has a pop-up flash, so they aren't considered of use only to snap-shooters. Many pros like and use them as well.
The Shutter and Mirror
The shutter release on the E1 is smooth, and functions just as it should. The sound of the release and the mirror mechanism is muted. Buttery smooth, to coin a phrase. I was pleased to note that the mirror lock-up can be set to an actuation delay of from 2 to 30 seconds. Terrific versatility.
The camera can be set to record RAW files, TIFF files and JPGs with two different compression settings as well as a number of different sizes. When RAW is shot one can elect to also record a JPG at the same time.
The provided software is a mixed bag. It's not really terrible, but definitely not as good as it should be. There are two programs; Olympus Viewer and Olympus Studio. Viewer is basic browser that is also capable of viewing and converting Olympus RAW files. Studio is a somewhat faster and slightly more sophisticated version, but is provided as a 31 day trial. After that it dies and you have to purchase it for $150.
I promise not to rant too much about this, but Nikon took heat a few years ago for charging extra for its RAW converter, and now we see Olympus doing the same thing. This wouldn't be so bad if the product was any good, but it isn't. Products like Capture One from Phase One and Camera RAW from Adobe run rings around it. Unfortunately Capture One doesn't currently support the E-1, but Photoshop CS does.
If you choose to use either Viewer or Studio you'll find that while browsing and actual RAW conversion isn't all that slow, every function appears to take a couple of seconds to be recognized. These programs have the feel of something written in interpreted Basic. On a positive note they installed elegantly and smoothly, unlike the software from another camera that I tested recently (no — not the Pentax) that took 30 minutes to install, gave no options, and loaded up my computer with some 50 MB of drivers and DLLs that I didn't want or need. End of Rant #2.
Oh yes. One more thing. Like the Pentax *ist D, the Olympus E-1 produces files at 72 PPI, with no obvious way of producing something more practical prior to importation into Photoshop. Why? Who needs a 27" X 35" file at 72 PPI? Nuts! (Maybe too many meetings at that Tokyo sushi bar).
I almost forgot. The software has no settings memory, so every time you do a conversion you have to make all of the settings afresh. Come on Olympus! This isn't rocket science. Please work with companies like Adobe and Phase One and get your RAW format properly supported. You, like most camera manufacturers, haven't a clue about how to write software. Leave it to those that do. And whatever you do, don't try and turn it into a profit center. At least not until you can produce fast, robust, full-featured code.
Inside the 4/3 Box
I believe that Olympus and Kodak have boxed themselves in with the 4/3 format. Here's why.
Let's begin with two premises that most people can agree with. Everything else being equal, larger imaging chips produce bigger prints, and larger pixels are cleaner and have lower noise than small ones. Now, let's consider the current chip manufacturing situation. The larger a chip gets, the more expensive it is to make. By the time you get up to a full-frame sensor (24 X 36mm) the chip alone may cost as much as an entire high quality camera. This is because of two factors. Imaging sensors are produced on 6" wafers, the same as any other semiconductor. This means that while you may get a half dozen large chips on one wafer you may instead get dozens, even hundreds of smaller ones. Since wafers have a fixed cost to manufacture, bigger necessarily costs more.
Another issue is yield. Even the best fabrication plants can't make every device perfectly. This means that some percentage are thrown away, increasing the costs of all the others. If a few chips on a wafer out of dozens need to be discarded it doesn't add much to the overall cost. But if one or two out of just a handful are bad it has a real economic impact.
Now, with Chip Fabrication 101 out of the way, it is also true that chip manufacturer prices are steadily dropping. One only has to look at the roughly 50% drop in digital SLR prices during the past 2 years to realize this. The Canon 300D Rebel has a 1.6X factor chip and retails with lens for under $1,000. Such is progress. A sideways glance at the price of LCD monitors will also demonstrate Moore's Law in action. A few of years ago LCD screens were ridiculously expensive and only available in small sizes. Now one can buy large LCD wall screens, and prices, on the smaller ones at least, are pushing traditional CRTs off the shelf.
We can now look at the flaw in the logic behind the 4/3 format. While larger chips (up to full-frame) are more expensive than ones of the 4/3 size today, this won't be the case for long. Increasing production volumes along with technological advances will bring the price of large chips downwards at a steady pace. Will smaller chips always be less expensive? Of course. But will the differential be enough to make the downside of using a smaller chip worthwhile for very long? I doubt it.
So, we have someone that buys into the 4/3 format in late 2003 or early 2004. They also buy several lenses for this format. But what happens in 2005 and 2006, and onwards? We will undoubtedly have imaging chips ranging from a 1.5X factor to full frame 35mm that don't cost all that much more, and you can be certain that companies like Nikon and Canon will be making cameras that use them, and which can utilize the huge existing inventory of full-frame coverage lenses available.
Anyone owning 4/3 format lenses then will have no escape. They will be limited to using cameras with a 2X magnification ratio because their lenses are unable to cover a larger image circle. If we assume that the price differential between small and medium sized imaging chips is going to decrease, then a 4/3 based camera will always suffer from smaller images or lower image quality by comparison, because while the number of pixels can be increased (this is accomplished by making the pixels themselves smaller), by making them smaller image quality is reduced. It's just physics. Anything that Kodak does to the 4/3 format chip can also be done to larger ones, so the differential will remain.
It seems to me that history is about to repeat itself. Olympus was the champion of the failed but elegant little half-frame format of the 1960's, and now appears to be heading down the same path. A shame really, because the E-1 is a very fine camera in many ways, and deserves better than to be built around a format that, like half-frame, may turn out to be just a footnote in the history of photography.
I have chosen to do a comparison of colour accuracy against the Canon 10D. In part because I have one handy, but also because it is a prime competitor at this price point.
The frame immediately above shows Chris Sanderson holding a Macbeth colour chart. The descriptions that I provide here are of what I see on a calibrated and profiled monitor. The comparisons are against the Macbeth colour chart as viewed under a D6500 Ott-Light True-Color fixture. (These sRGB JPGs may or may not show you what I describe, depending as well on the adjustment of your particular monitor. Go by what I say, not by what you see above).
Both frames are crops, but neither has had any tonal or colour adjustment done to them in either RAW conversion or in Photoshop, other than Gray Balance.
The most obvious difference is in the rendition of yellows. The Canon is too saturated while the Olympus is not quite accurate, but better. The Olympus also does a more accurate job with the blues but the reds are better rendered by the Canon. Orange is also more realistic on the Canon but neither do a very good job on flesh tones (too warm).
In any event, the Olympus turned in a credible showing in this regard, and neither camera can be faulted in terms of accuracy of colour reproduction.
ISO and Noise
I did some extensive high ISO and long exposure noise tests. In the interest of space (and reader fatigue) I won't bore you with them. The Olympus is competitive up to ISO 400 with just about any other camera, but can't hold it's own against the Canon 10D at higher speeds. Few cameras can. Similarly with very long exposures. Whether this will be an issue for E-1 owners will very much depend on the type of shooting that they are involved in.
The Olympus E-1 is an excellent camera and the start of what promises to be a well designed system. It has a few flaws, but what product doesn't? Olympus is a very progressive company and I'm confident that as time passes the line will be nicely filled out with differing and improved models, and more lenses.
One area where I hope they concentrate some effort is by adding some form of image stabilization to their long lenses. With Canon offering an extensive line of IS lenses, and Nikon introducing new ones (albeit at a slow pace), Olympus will find that many photographers, especially sports and wildlife shooters, will miss the availability of stabilization.
As for the future of the 4/3 format itself — well, I've made my thoughts known and won't rehash them. If I'm wrong that will be great, because it will mean that the photographic community will have another strong player and more choices.
My thanks again to Vistek, Canada's largest professional imaging equipment dealer for making this equipment available for testing.
Update — Nov. 29, 2003
DPReview is one of the most popular sites for technical reviews of digital cameras, and rightly or wrongly their opinion carries a lot of weight with purchasers. They have just published their review of the Olympus E1 and it contains some strong criticisms.