The Olympus OM-D series E-M1 is the first Olympus Micro 4/3 camera targeted to professional and advanced amateur photographers. It was introduced in 2013 and seems to have been well accepted by advanced and professional photographers. It includes highly configurable settings, a relatively large number of physical controls, metal-alloy body and weather sealing. A "Pro" series of Olympus Micro 4/3 lenses designed to match the capabilities of this camera (but usable also on all other Micro 4/3 cameras) is slowly being introduced by Olympus. So far this series includes only one available model, shown in the figures, and three more planned.
The E-M1 is a bit wider than an E-M5 equipped with half of the HLD-6 battery grip, but the E-M5 in this configuration is a bit taller than the E-M1. Volume is about the same. Weight of E-M5 with half grip is 528 g versus 497 g for the E-M1 (yes, the E-M1 is slightly lighter). This is a fair comparison because I cannot securely hold the E-M5 in my right hand without this grip. There are third-party grips available for the E-M5, but they force the right index in an uncomfortable position to reach the shutter release button and front dial.
Holding the E-M1 is a pleasure, except that the built-in grip is a little short in the vertical direction for some people's hands. I found a solution in a third-party Arca-compatible L-bracket, branded iShoot (above) that extends the height of the body just enough to provide a home for my little finger. With this bracket, the E-M1 is probably the best combination of comfortable holding, reasonable weight and ease of attachment to a tripod head among all the cameras I owned or used. Cosmetically, this bracket fits very well, and looks more like an integral part of the camera than an add-on. The vertical portion of the bracket can be removed by unscrewing a single bolt. The battery door at the bottom of the camera still opens fully, and movements of the rear LCD are unimpeded. The rubber doors on the left side of the camera are more difficult to open than without a bracket, but can still be operated reliably.
This bracket has on its right side an attachment slit for a camera strap (above). It is probably meant to attach a hand or wrist strap, but I use it for the neck strap instead. The result is that the camera sits against my chest with the lens pointing downward and the rear LCD facing upward. This reduces the risk of banging the front lens element against other people in a crowd and getting it dirty, as well as scratching the LCD by rubbing it against buttons or zippers of my clothing. When I lift the camera to my eye, the strap remains out of the way on the right side of the camera. Operation of the card door is also easier now that the strap does not hang across it. It would be possible to remove the vertical portion of the bracket and use its bolt for the same purpose on the left side of the camera.
The E-M1 improves on the E-M5 in several significant respects. Numerous reviews are available if you are interested in a complete rundown of all features, for example on dpreview. Therefore, I don't need to repeat these reviews, and can concentrate on the improvements that are especially significant to me. The following list is therefore not complete, but only a selection of what I think is most important.
A combination of phase-detection and contrast-detection autofocus is the single most important improvement of the E-M1.
DSLRs (and modern film SLRS before them) contain a number of mirrors and prisms. The main reflex mirror sits in front of the sensor and directs light to the optical viewfinder, where the image is righted by an optical pentaprism that gives the camera its typical shape with a viewfinder "bump" atop a brick-like body. The main mirror is semi-transparent and some of the light passes to a secondary mirror located behind the main one. This secondary mirror reflects light to a single-chip device carrying multiple autofocus sensors. Exposure sensors are also present, but their placement can vary.
The sensor has specialized sensels that record only light coming from one specific direction, and therefore a set of complementary sensel-pairs records an image coming from the left direction and one coming from the right. Typically, phase-sensitive sensels are arranged in a line (with one subset of sensels "looking to the left", the other looking to the right), and therefore the output consists of two lines, or one-dimensional images. If the subject is correctly focused, rays coming from one region of the subject and passing through different regions of the lens (for example the right and left sides of the lens) are focused on the same region on the autofocus array of sensels. The two images recorded by the two subsets of sensels in this case seamlessly overlap. If the subject is out of focus, instead, the two images are shifted sideways (i.e., out of phase) with respect to each other. The camera electronics detect the difference between "left-looking" and "right-looking" images, and compute the direction and amount of offset between the two lines. Based on this data, the lens is refocused in the necessary direction and by a computed amount. The process continues during focusing, until perfect focus is achieved, typically in a fraction of a second. Overshooting the focus position normally does not happen, because focusing speed can be slowed down as the system detects that perfect focus is approaching. This is the basis of phase-detection autofocus. In poor lighting conditions, with slow lenses, or with subjects that lack detail, focusing may results in focus "hunting" back and forth.
Once autofocus is achieved, the main and secondary mirrors are lifted, the shutter opens and the image is recorded. Advanced cameras may recognize during autofocus that they are tracking a moving subject, and continue to focus predictively during the time taken by the mirrors to lift and shutter to open.
Usually, a phase-detection autofocus sensor has rows of phase-sensitive sensels that can detect a phase shift in the horizontal direction. This works best with subjects that have points or vertical/oblique lines of high contrast. This type of sensor cannot detect phase shift in a subject that only has horizontal, parallel lines of contrast. Advanced phase-detection autofocus sensors have also vertical columns of sensels, to sense phase differences also in this case.
Since phase-detection autofocus relies on light passing through opposite sides of the lens, low-speed lenses do not provide light pathways sufficiently diverging from each other. Generally, phase-detection autofocus can work reliably with f/5.6-6.3 or faster lenses.
Mirrorless cameras, as their name implies, have no main and secondary mirrors. Until a few years ago, these cameras could only perform contrast-based autofocus, which relies on the detection of sharp borders between light and dark areas of the image recorded in live view by the same sensor used to record images. This method provides no indication of the required direction of focusing, so the camera has to try one direction first, and if focus get worse it reverses and focuses in the opposite direction. There is a fifty-fifty chance of starting in the wrong direction on the first try. Autofocus may have to reverse again if it overshoots the correct focus and has to backtrack. Sometimes it overshoots and reverses multiple times if the subject does not have areas of sharp contrast. A further problem is that the live view image is recorded as successive, discrete frames rather than continuously, and focusing may have enough time to overshoot between frames, especially in poor illumination that increases the time interval between frames. On the other hand, contrast-detection autofocus is in principle not limited by lens speed, and with subjects that contain a lot of fine detail it can potentially be more precise than phase-detection autofocus.
Contrast-detection autofocus has substantially improved over the years, but not to the point of equaling the speed and immunity to hunting of phase-detection autofocus. The Olympus E-M5 is an example of good contrast-detection autofocus. The main limitations of contrast detection (no indication of which direction of refocusing is needed, and how far is perfect focus from the starting position) seem to be unavoidable without some type of phase detection. Short of putting a mirror back in the camera (which was done in some Sony cameras), until recently there seemed to be no way to add a phase-detection sensor to a mirrorless camera.
The Bayer sensors used in most digital cameras have two green-sensitive sensels for each red- and blue-sensitive sensel. Therefore, one might think of using some of the "extra" green sensels for a different function, specifically, phase detection. This leaves a "hole" in the recorded image, which can be made to disappear with the same type of in-camera processing used to make defective sensels and "hot pixels" disappear. As long as there is enough processing power available, and the sensor has a sufficiently high total number of sensels, even entire rows and columns of missing pixel data can be made invisible. In the E-M1, 6.25% of the sensels are dedicated to phase detection. This does not mean that the image produced by the sensor is 15 megapixel instead of the declared 16, because the missing pixel data is interpolated from neighbor sensel data. In addition, phase detect sensels are always spaced 4 sensels away from each other, both vertically and horizontally, and as a result only 1 out of 8 green-sensitive sensels is "missing". In the sensor of the E-M1, phase detect sensels are overlaid by a prism-shaped microlens instead of a convex one like ordinary sensels. There also seems to be no color filter on autofocus sensels.
The E-M1 has phase-detection sensels that look left or right, but not up or down. Therefore, it cannot autofocus by using cues from horizontal features of the subject, and the camera may need to be inclined sideways or turned by 90° to autofocus on very unusual types of subjects that have no vertical or oblique features. Just remember this in the unlikely case it should happen to you. The trade-offs of this type of sensor are well worth the improvements in autofocus performance. The E-M1 can autofocus with 4/3 lenses (via an adapter) almost as fast as Olympus 4/3 DSLRs do, while all earlier Micro 4/3 cameras are painfully slow with these lenses. As a whole, the E-M1 cannot yet compete with professional-class DSLRs in autofocus speed, reliability and tracking capabilities, but seems to be better than many consumer-class DSLRs. It is also competitive with some of the professional DSLRs of a few years ago.
This does not mean that the E-M1 will autofocus at blinding speed with any lens and in any light condition. With the Olympus 60 mm macro, for instance, there is still quite a lot of focus hunting and racking the lens back and forth (although normally not more than once). The same thing happens with all DSLRs I used, which is quite natural because in a macro lens the focusing range is so wide that, even with focus-limiting activated, the subject is so badly unfocused that neither phase-detection nor contrast-detection find anything to lock onto, and the camera must rack the lens all the way out and in before it can finally find something to focus onto.
A limitation of the E-M1 is that it cannot use its most peripheral phase-detection sensels with 4/3 lenses, perhaps because these lenses are designed for a higher distance between rear elements and sensor, and therefore phase-detection sensors that, for instance, "look left" and are placed near the left edge of the sensor see no light coming through 4/3 lenses. This is quite possibly a limitation of 4/3 lenses, rather than of the E-M1. The E-M1 has also other restrictions on the use of phase autofocus in certain operation modes, which may be due to limitations in processing power and, in future cameras, could be removed by multiple processors working in parallel.
Newer types of sensors, some of which are already used in Canon cameras, place a pair of complementary phase-detection sensels, or even four of them (one horizontal pair and one vertical pair) in place of a single green sensel. These sensors increase the number of autofocus sensels and make their output potentially usable for imaging in addition to autofocus. I expect fast improvements in imaging sensors with built-in phase-detection in the near future, with the potential of matching or exceeding the performance of dedicated phase-detection sensors of DSLRs. When this happens, there will not be any point in using the new sensors in large, heavy, mechanically and optically complicated DSLRs, which eventually will be replaced by simpler and cheaper mirrorless designs. Nikon and Canon should better take notice before it is too late.
A consequence of phase-detection integrated in the imaging sensor is that it is no longer possible to place an anti-aliasing filter in front of the sensor, and therefore anti-aliasing has to be done by in-camera post-processing, in connection with demosaicking. This is another example of shifting more and more camera functionality from hardware to software, and of the increasing importance of powerful in-camera data processors. As a result, "raw" image files may receive more and more in-camera post-processing, making them more and more different from the original concept of just storing the unaltered data collected by the camera sensor.
Autofocus types in the E-M1
It is not possible for the user to choose which AF type is used by the E-M1. The following is my current understanding of which AF system the E-M1 chooses to use in different circumstances:
More on Olympus E-M1 versus E-M5
As far as I am aware, the E-M1 does not lack any of the E-M5 features that I normally use. The E-M1 also uses most of the accessories available for the E-M5, including batteries. Exceptions are the HLD-6 battery grip, third-party L-shaped camera brackets, and some of the camera plates for attachment to a tripod.
All changes from E-M5 to E-M1 are improvements. However, there are a few points that could have been improved in the E-M1 but were not addressed, for instance:
Both the E-M1 and E-M5 have a low-vibration shooting mode (called Anti-Shock by Olympus) that electronically starts the exposure after a configurable delay of one to a few seconds from the initial closing and opening of the mechanical shutter. This feature allows shutter vibrations to die out before the exposure starts, and is especially useful in astrophotography, tripod photography (especially with long lenses or when a remote control is not available), photomacrography and photomicrography. A present to E-M1 owners came with firmware version 1.2, in the form of an electronic first-curtain mode that completely avoids the mechanical shutter movements before the exposure, and therefore eliminates vibration. To set this mode on the E-M1, set the shutter delay to 0 seconds in the Custom → Anti-Shock menu. A limitation of this new setting is that it works only for exposure times of 1/320 s or shorter, which substantially reduces its usefulness. So far there has not been a comparable update to the E-M5, although this feature has been introduced in firmware updates of at least one lower-end Olympus camera.
The maximum length of time-lapse sequences is much greater in the E-M1 (999 shots, compared to 99 shots in the E-M5). It is now feasible to shoot reasonably long sequences to display as high-frame-rate animations.
The E-M1 has a flash hot contact at the front of the camera, missing in the E-M5. The same voltage limitations apply as for the hot shoe contact.
Several of the physical controls of the E-M1 have been repositioned with respect to the E-M5. This is both an advantage and a problem. The slightly wider camera body and the built-in grip on the right side allow the controls on the top and rear panels to be less crowded, but this is not the only change. Button 1 is now easily accessible and can be found without fumbling. The Playback button is much easier to operate with normally-sized male fingers, once you you learn its new position.
The HLD-6 battery grip for the E-M5 duplicates the front dial and shutter release button already present on the body of the E-M5, which is a necessity but makes the controls even more crowded. The built-in grip of the E-M1 avoids this duplication and consequently is less crowded with controls on its top panel. Curiously, the rear dial of the E-M1 and E-M5 have no button at their center, which would be an obvious place for a comfortably large button. The small button at the center of the E-M1 mode dial locks the dial against accidental turning. I did not miss such a feature in the E-M5, and question its real usefulness. A configurable button doing something more useful could instead be placed here.
The on-off switch has been moved to the left side of the top panel. This is near the place where a similar lever was located in the OM-series film SLRs, which is the reason for this design choice (however, the lever in the OM has additional functions). In the E-M1, the on-off switch is easier to use with either hand than the awkwardly placed one of the E-M5 (near the bottom right of the rear panel). Olympus should probably take a hint from Nikon DSLRs in this respect, where the on-off switch is a rotary dial around the shutter trigger button, easy to operate with the right index in shooting position. It saves a lot of time and avoids having to move either hand away from their normal shooting position. Some Panasonic Micro 4/3 cameras place the on-off switch around the mode dial, which is placed near the shutter release button and is easier to reach than in the E-M1.
The mode dial in the E-M1 moved from the left to the right side of the top panel, and in place of the mode dial of the E-M5 there is now something that looks like the film rewind knob of the old OM cameras. It is round like a dial, carries two buttons, but it does not turn. Other controls have been moved without apparent reasons, for instance the button that switches between eyepiece and LCD screen. In general, controls are easier to operate than in the E-M5, but initially it takes a little time to find them if you are used to the E-M5.
Olympus E-M1 versus E-M10
The E-M10 is currently the latest addition to the Olympus OM-D line. The E-M10 is slightly smaller than the E-M5, lacks weather sealing and phase-detection autofocus, and is meant as a simplified and cheaper model than the E-M5. It is still a capable camera, and a valid alternative on a smaller budget or where a physically smaller camera is desired but an eye-level viewfinder is still necessary. It is not fair to compare the E-M10 with the E-M1, because these two models are simply in different classes. The E-M10 could also be a choice as a second Micro 4/3 body in a kit based on the E-M1, unless one expects to switch cameras seamlessly and continue shooting without a thought. Both the handling of the body and the placement of the controls are different, so these two cameras are not really interchangeable. The E-M10 is definitely not recommended for photographers with big hands.
Olympus E-M1 versus DSLRs
The number, size and comfort of use of physical controls in professional-class DSLRs is still superior to the OM-D series, in spite of the improvements of the E-M1. The buttons in the Nikon DSLRs I used in the past are twice the diameter of the E-M1 buttons. I can fit the entire tip of my finger on recessed buttons of Nikon DSLRs , as opposed to having to extrude the tip of my finger, like toothpaste being squeezed out of a tube, into the narrow holes housing some of the recessed buttons on the E-M1 and (especially) E-M5. On the E-M1, the Play button has this problem, and the Fn1 button is also a bit too flush for comfort.
The E-M1 is larger and heavier than most other Micro 4/3 cameras, most of which are just too small for comfortable holding. However, the difference in size and weight between the E-M1 and a typical DSLR with metal-alloy chassis remains dramatic. A typical Micro 4/3 kit with high-quality lenses weighs about half the corresponding kit of an APS-C DSLR, and less than half compared to a full-frame kit. Additional, dramatic weight savings can be made by using Micro 4/3 consumer-level lenses in cases where top-class optical performance is not essential.
Olympus E-M1 in a two-body kit
An Olympus E-M5 or E-M10 remains a choice as a second body in a kit based on an E-M1. Disadvantages are the differences in placement of controls and the much smaller body of the E-M5 and E-M10, difficult to hold in one's hand without an additional hand/battery grip. The main advantage is that the weight and size of a second body are not very noticeable in a normal camera backpack. Two Micro 4/3 bodies weigh substantially less than a single full-frame DSLR body like the Nikon D4.
A second E-M1 would be the obvious choice for a second body if allowed by one's budget, since it adds very little weight and size with respect to an E-M5 or E-M10, and eliminates the need to cope with different control positions and software configurations. This solution may be best for action and event photographers who need to frequently and quickly alternate between two very different focal lengths (e.g., telephoto and wideangle).
In some cases, the E-M1 could be accompanied in a kit by a totally different Micro 4/3 body, for instance one of the Olympus Pens or Panasonic smaller bodies without eye-level viewfinder. This might be useful for street photography and a few other situations where a camera that looks like a point-and-shoot offers psychological advantages, or where every gram of weight counts.
The Olympus OM-D E-M1 is a significant improvement on the E-M5 (and E-M10) in most respects. New autofocus capabilities and a larger number of better placed controls are the main advantages over the E-M5. It is slightly larger and heavier than the E-M5 but significantly more capable and comfortable to handle. The E-M1 takes up essentially the same space and weight of an E-M5 equipped with half of the HLD-6 grip.