In September, 2005 Olympus surprised us by announcing a new entry-level SLR, the E-500, just ten months or so after the previous model, the E-300, was introduced. What was so surprising was not just the timing, but the fact that the E-500 was an entirely new construction rather than an evolutionary step. In particular, Olympus seemingly gave up on the flat-top design with side-swinging mirror, reverting to the "classic" pentaprism-shaped hump instead.

Now, here we go again. Just three months after the E-500 arrived to the stores, Olympus announced yet another model to be available within a month or so, the E-330.

Live electronic viewing

This is the single feature most used by Olympus in their E-330 marketing campaign. It is also, I believe, the most controversial one.

Traditional SLRs cannot provide this functionality. The swinging mirror between the lens and the light sensor (or film) either reflects the light-forming image towards the finder screen, or, moving out of the way, allows the light to get to the sensor — but not both at the same time.

SLR cameras could, in principle, be switched between optical and electronic viewing (Canon EOS D20a, a specialized astrophotography version of the D20, can do that in a very limited and kludgy way), but in a costly and complicated way. The shutter would have to stay open during electronic preview, then close for sensor flushing, re-open for the actual exposure, close again to terminate it, and open again for viewing. Additionally, high-performance sensors used in SLRs are not best suitable for live preview, and for many reasons.

Still, the E-330 is not (contrary to some uninformed claims you can read) the first SLR with electronic preview. The E-10/E-20 SLR from Olympus had this feature, back in 2000, and the Sony D700 a year earlier. These cameras did not have interchangeable lenses, but, still, they were true SLRs ("single lens reflex") nevertheless. They were using a semi-transparent prismatic mirror to redirect a fraction of light from its path to sensor towards the viewing system. In the film domain this approach has been originated by Canon with their Pellix SLR model forty years ago.

The E-10/E-20 was a very good camera at its time, and it is still capable of delivering good results. For a number of reasons it had to have an in-lens shutter (this implies limited fast speeds), and its sensor had to be a compromise between viewing and image recording. I was hoping Olympus would revisit the concept.

A typical use of electronic preview is in taking pictures from a high or low angle, where using the SLR finder is inconvenient or impossible: for example, over the heads of a crowd, or from the ground level. In such situation the usefulness of this option is greatly enhanced by a tilting monitor.

One group of people who will be especially happy with the live preview feature are those who do underwater photography. Until now, this application was basically impossible with an SLR, as you cannot use the optical finder when the camera is in its underwater housing. Now it will be just fine, especially with Olympus announcing the dedicated underwater enclosure, PT-E02, for the E-330, and with the more readable, larger monitor. (Of course, a non-SLR camera would be as good for this purpose...)

Another use may be the infrared photography: without live preview you have to take the IR filter off the lens to frame the picture, only to put it back before releasing the shutter; a real hassle.

In a quite unexpected move, in the E-330 Olympus implements not one, but two mechanisms for live image preview, vastly different and sharing only the final element: the LCD on which the preview is displayed.

Live View — Mode A

The image-forming light coming out of the lens is first reflected by the side-swinging mirror (A). This is a mirror like in all SLRs, except that in other cameras it redirects the light rays upwards, and swings up. While in the viewing position, the mirror blocks the light from the light sensor (I), in front of which we additionally have the focal-plane shutter, and a number of filters, including the anti-dust cover; not shown here.

Actually, the mirror (A), as far as I understand, is also semi-transparent on a part of its area, letting some light through to another, smaller mirror (not shown), riding piggyback on its back side and providing light for the autofocus sensors. While this is an extra mechanical complication (one more thing which may go wrong), the piggyback mirror trick has been in general use in AF cameras for years.

The real image of the scene is then formed on the groundglass screen (B) which must be placed at exactly the same distance from (A) as the sensor (I), so that the viewed image (especially its focus) corresponds to the one which will be recorded.

Up to this point there is nothing new in the design: all other SLRs form the image on the screen (B), to be then viewed by the eyepiece (F). In the E-330 this is facilitated by three other mirrors: (C), (D), and (E); in some cameras a solid glass prism is used for that purpose. Note that a three-mirror solution needs that all mirrors are precisely aligned, otherwise viewing accuracy will be affected.

Again, the E-300 uses exactly the same three-mirror approach and it works just fine. The difference is in the third mirror, (E), which in the E-330 is semi-transparent. While it reflects most of the light towards the eyepiece (F), it lets some of it (20% or so, I would expect) through, with yet another mirror (G) directing it towards the viewing sensor (H). Obviously, there must be some lens between (G) and (H) to form another real image on that sensor's surface; that lens is not shown in the diagram.

(Another diagram by Olympus indicates that the multi-segment exposure-metering sensor, not shown here, is located behind the mirror (G), which would indicate that this mirror is also semi-transparent.)

All of this becomes much simpler at the moment the picture is taken: the mirror (A) swings away letting all the light to the sensor (I), and making all other parts irrelevant. This is a good news, as all complexity described above affects only the viewing process, not the actual image recording.

Two imagers instead of one may sound like a major expense, but this may depend. Instead of procuring a custom sensor for this purpose, Olympus decided to play the advantage of scale factor, and use the same 8-megapixel CCD chip which is used as the imager in their newly-announced Stylus 810 camera. Frankly, I was a bit surprised, as pushing around this amount of information in real time requires some CPU muscle and uses power. Oh, well, the designers had all the data and I have none, so I must trust their choice here.

One of the reasons for choosing the Stylus imager for viewing was that it has a low-light boost mode, combining the signal from a 3×3 block of pixels into one. What you effectively get this way is an 0.9 MP imager with nine times the sensitivity; this may come handy, as, remember, the semi-transparent mirror lets only 20% of the light towards that sensor. I wouldn't be surprised to learn that this mode is used all the time. By the way, this is a 1/1.8" chip, which means that it effectively measures 5.4×7.2 mm, less than half of the linear size of the main imager.

Some Web sources mention a 5 MP or even 0.8 MP sensor dedicated to the live view; the information I'm giving is quoted after an interview with Olympus representatives.

Another price we pay for the luxury of real-time electronic preview is the sacrifice of some light at the semi-transparent mirror (G). The image seen through the eyepiece (F) will be correspondingly darker. Let us hope the 20% (my guess) loss will be barely noticeable; after all, our senses are scaled logarithmically. Using a lens one stop wider in aperture doubles the finder brightness, much more of a change, and a few more glass-to-air surfaces in a more complicated super-zoom may easily cause a similar light loss, in spite of multicoating. Therefore I'm not really concerned about the 20% loss.

A common way to increase image brightness in the optical finder would be to make its screen (B) more transparent, so that what we are viewing is, partly at least, an aerial (as opposed to real) image. This would, however, allow our eye to accommodate itself to images out of focus, and the finder-based manual focusing would become much less accurate. Actually, I wouldn't mind that: manual focusing of the tiny finder image is, anyway, problematic at best. I believe this was the approach used in the E-10/E-20, where, in spite of much smaller sensor and screen size, the viewed image was larger and brighter than in any sub-$2000 digital SLR I have handled (I have seen statements to the opposite in some Web articles, but the writers never bothered to compare; I did). I wish Olympus would have chosen this way — but this could raise a lot of fluff...

Early previews of the E-330 (see the list at the end) list complaints about the quality of live preview in low light: it may just get very dark. There is a boost feature, allowing to increase the brightness, but it is reported to suffer from graininess and flicker. We are not sure whether this is a problem with pre-release cameras, or a built-in "feature". As ugly as this may sound, it does not worry me a lot: with the optical viewing I would see a need for live electronic preview only in special situations, when all boils down to framing.

Interestingly, the electronic preview from the dedicated sensor covers only 92% of the image area (I am not sure if this is actually area, or linear percentage, in which case the actual area coverage would be 84%). This is probably supposed to hide any misalignment problems, and, really, I would rather take a 92% live preview than none at all.

Well, if you think now you know how the E-330 does real-time preview, you are only half-right. There is also...

The second live preview: Mode B

This is even more unexpected, for me at least, than the live preview described above. Looks like the good folks at Olympus decided that this feature may be not good enough for critical applications, so they decided to add yet another feature, referred to as "Live Preview Mode B", with the one already described being "Mode A".

In Mode B the main mirror (A) remains out of the way when the image is previewed, so that the light from the lens goes directly to the main recording sensor (I). The shutter remains open, and the main sensor signal is used to generate the image shown on the LCD monitor. This has both advantages and disadvantages, both built-in into the very concept.

On the plus side, the previewed image comes from the same sensor which is used to generate the recorded one. There is no risk of misalignment in the (A)..(H) path to cause any geometric inaccuracy of the preview. (Actually, the Mode A LCD shots I have seen in one of the reviews were slightly tilted, which shows that alignment problems are real.) The preview is also generated with full pixel resolution, so it can be enlarged on the screen to help in critical manual focusing — more meaningful than the tiny image in the optical finder.

What you are giving up in the B mode is the availability of all the metering (exposure, white balance) information during the preview. The sensors performing those functions are placed after the main mirror (A), so this is obvious: the mirror is out of the way.

Now, what is happening when you shoot a picture in Mode B is quite complicated. The shutter has to close to allow the main sensor (I) to be flushed. The main mirror (A) gets back in the position, cutting off the live preview. The circuitry for exposure and white balance metering does its job, adjusting the camera's settings. Then the mirror goes away again, the shutter opens, and the main sensor records the image. This is terminated by the shutter closing again, and the image signal is picked off the sensor to be recorded to a file. As soon as the sensor is done, the camera gets back into Mode B live preview state: the shutter opens again.

All of this takes some time, and therefore induces an extra lag, a second or so, between the release button is pressed and the picture taken. This lag would have been even longer if not for the fact that Mode B is designed to work only with manual focusing: the AF is not activated in the short instant the mirror is "down" (actually: in the light path).

This means that the Mode B preview is suitable only for critical, tripod-based work (mostly macrophotography); but that's fine with me, as it is most useful in such applications.

Obviously, it would be very nice to have a real-time brightness histogram preview in both Live View modes. This, however, was not included; most probably because of software development cost. If and when the technology matures, I will expect to see this feature.

The new sensor

There is one more impact of Mode B preview on this camera: FFT (full frame transfer) CCD sensors are not really suitable for preview, which requires long exposures to light. This is why Olympus decided to use a MOS (metal-oxide-semiconductor) sensor in the E-330, not the (excellent) Kodak CCD used in the previous models. (The new chip is made by Panasonic, i.e., Matsushita.)

The sensor type is referred to as Live MOS, whatever "live" may mean here. It remains unclear how, if at all, does this type differ from CMOS sensors used, among others, by Canon in their SLRs. Intrinsically, CCDs have less noise (before applying any filtering to the image, that is) and higher dynamic range. CMOS sensors, however, are less expensive to manufacture, consume less power, and image readout is generally faster.

This alone makes the E-330 an entirely new camera, compared to the previous E-System models, and I am not sure how the sensor switch may affect the image quality. Olympus refers to the Live MOS as delivering results "comparable to" those of the FFT CCD, and it remains to be seen what "comparable to" means in the marketing lingo.

The early previews do not include any image samples. A few samples have to be posted on a number of Olympus Web sites; here is a link to the Olympus Japan E-330 image sample page. The samples published currently by Olympus clearly show how good the Olympus glass is (these were shot with use of the 50 mm F/2 macro and 11-22 mm zoom), the noise seems to be well under control (at least at ISO 100), but I would wait with voicing any opinion until I can shoot the same subject with the E-330 and, say, E-500, side-by-side.

The new sensor also has a slightly lower pixel count than the CCD used in the E-300 and E-500: nominally 7.5 (as opposed to 8) megapixels; in terms of "binary" megapixels (2²⁰), the numbers will be 7.03 (E-330) and 7.62 MP (E-300/E-500). I have seen voices complaining that it is a "step back" from those cameras, but hold on: what it boils down to is that the native image resolution in the E-330 is 3136×2352 pixels, compared to 3264×2448 pixels in those cameras. This translates into less than 2.9% difference in (linear) print size, all other things equal. Come on, this is not even worth mentioning!

According to Olympus (and quoted after Jason Busch) the new sensor sports three times larger photo-sensitive area than a typical CMOS one. This should be reflected in better image quality, I would expect.

Now, with the major issues covered, let me go into some other detail on the new camera.

Putting all said above aside, the E-330 looks and behaves like a hybrid of the two preceding models. It inherits the best of each parent, throwing in some new features.

Inherited from the E-500

From the mama (E-500), the E-330 gets the excellent control system, including the direct access to almost all settings from the well-designed Control Panel, without a need to dig into the menu system.

It also gets the gorgeous LCD monitor, 2.5" (64 mm) diagonally, made by Sharp. A number of cameras boast nowadays monitors of this size or larger, but most of them limit the resolution to 130 k pixels (or rather RGB sub-pixels), while the E-330 has a pixel count of 215 k, a visible improvement.

One reason I dislike the new Olympus P-series is exactly that: the bigger monitor has the same old low resolution. Olympus used to have what I believe was the best non-SLR line-up on the market: the C-5050Z, followed by C-5060Z and C-7070WZ, and now this line seems to be gone, with the P-series cameras being a far cry from those. Give me an updated C-7070WZ with faster AF, bigger and more detailed monitor, and some tweaks to the user interface, and I bring you my check again. The mass market does not seem to agree with my whining, though. Tough.

The electronic preview uses two (!) dedicated buttons above the monitor: one to turn it on, and another to switch between Modes A and B. I'm not sure I like that: this is a premium real estate, easily accessible with your thumb, and I would prefer to keep the exposure lock there, instead of moving it way down. After all, switching between live preview modes is not a function I'll be using often. These two buttons could be placed more out of the way, and the lock button would be easier to find with your eye at the viewfinder.

Hold on, what happened to two other buttons present on the top-right in the E-500: WB reference mode and focus spot selection? While I'm not missing the latter, reference WB is a feature I like to have easily available, and this is a throwback.

Taking that into account, I consider the controls in the E-330 to be a tad behind those on the E-500, while still better than on the E-300. With no doubt, the direct access, active Control Panel is a great advantage, and I believe all camera makers should borrow the idea from Olympus. (I also wish this could be retrofitted into the E-1 and E-300 by means of a firmware upgrade; this would be a terrific plus and a good marketing point for Olympus: "we care about you even after we've got your money".) From the E-300 the new model gets sturdier construction with the durable, crinkled finish I like so much, the more substantial metal inner frame, and, most importantly, the flat-top design, with the unorthodox light path to the finder, including the side-swinging mirror. More about this later, when I get into details.

The E-330 also retains dual memory card slots (CF and xD) present in the E-500. While this is not really important, I like this feature for some irrational reasons.

Most other specs, including the autofocus, exposure automation, shutter speed range and more, seem to be the same (or at least close) to those on the E-500 (and E-300), and I'm not going to bother you with the full listing. If you have to have that, check any E-330 preview from a respectable source.

Inherited from the E-300

Now, from the papa (E-300) the E-330 gets, first of all, the flat-top design, so much maligned in many reviews. True, it may not look as cute as a small body with a big hump on the top, but it makes better use of the space available in your camera bag. It also moves the built-in flash away from the hot shoe for external one, making possible to use both at the same time. Being able to use the external FL-36 in bounce mode, still providing some fill-in from the internal unit is a great plus; inability to do so was perhaps my biggest complaint about the E-500.

The flat-top layout, with the side-swinging mirror (discussed in detail in the section on live preview) leave more room to include the preview sensor into the equation. It may be a matter of necessity, not choice.

The body construction and finish also seem to be similar to those on the E-300. While the E-330 no longer has the funky metal top apron, it provides the same durable, crinkled finish I liked so much on the older camera (way ahead anything on competing models, including the E-500), and the padded thumb rest on the back. The metal frame inside the (otherwise plastic) body seems also similar to that in the E-300; more solid than the symbolic one included in the E-500. In general, a quality job, again.

All these similarities aside, the body was reduced in width, mostly by narrowing the gap between the handgrip and the lens. While the E-300 was 146×85×64 mm (W×H×D), in the E-330 it is 140×87×72 mm: just 2 mm taller, and 6 mm narrower. This should improve the balance of the camera in your hand a little bit (although the E-500 may still be better balanced). At 550 g (body only), the E-330 is heavier than the E-500 (435 g), almost the same as the E-300 (580 g). Well, a small price to pay for sturdier build.

A slight re-design of the top of the body can be also noticed. I'm not using my cameras as fashion statements, but the E-330 may, indeed, look a bit better. Here is the comparison:

Remember: beauty is only skin-deep, and what may seem hip now may look pathetic in a few years. Still, note the missing top metal shell and more rounded shape. The AF-assist light is missing (like the E-500, the new model may use the built-in flash for this purpose), and the mode dial is no longer tilted backwards (a nice touch on that camera).

New to the E-330

While we already went through the major news, there are some less dramatic changes for better in the E-330.

First of all, the LCD monitor can swing out (one degree of freedom), up to 90° up and 45° down. This makes live preview easier. Yes, a two-axis rotation would be nice, like on the '5060/7070, but it requires more room, necessitating larger body size. This is not really important, but I wonder if the 45° of tilt downwards is sufficient for some applications (like shooting over the crowd).

Second, hallelujah, we've got the eyepiece shutter (see the picture above). Olympus had to include it to avoid stray light disrupting the Mode B live preview, but I find this feature essential for all times when the camera is used on a tripod (and critical for any infrared work). Once again, I would have wished to have the eyepiece shutter in the previous models, regardless of the reason.

Still missing

Some of my complaints about the E-300 and E-500 still remain valid.

• No wired remote socket. Come on, using the (optional) infrared remote is not the same! Does the extra socket cost that much to make?
• No time-lapse photography. This is something a few of us need, and only rarely, but when we need it, we need it badly. We've seen this feature in the E-10 and E-20, and it can be implemented entirely in the firmware.
• The USB connection is USB 2 "Full Speed", which is not the full speed USB 2 is capable of, but rather that of USB 1.1. Taking the full advantage of the maximum possible "high speed" transfer rate would be perhaps impossible because of camera's circuitry limitations, but a speed increase of two or four times would seem realistic.

And I really, really hate this little, pesky piece of plastic, protecting the hot shoe from dirt and elements. I already lost six of those (replacements are available from Olympus at $7 or so apiece.)

Digital image stabilization? Nah!

The camera's specs at the olympus-esystem.com site, list, in the "Scene Select AE" section something named "Digital Image Stabilization", while a similar table at Olympus America refers, in the "Scene Select" section to "Reducing Blur". Are you thinking what I'm thinking?

Interestingly, among four detailed previews I've seen, two never list or mention this feature, while two refer to it as increasing ISO setting to avoid blur. Olympus does not refer to it in their press release, and this would be something definitely marketable.

Three possible explanations come to mind:

• Some kind of electronic image stabilization was originally designed for the E-330, but it didn't work well and was pulled off just before the press release (and before sending cameras to reviewers); specs sheets might have been prepared earlier;
• Some marketing genius decided to offer a special "scene setting" using higher ISO, and to call it "digital image stabilization", a clear case of misrepresentation. In a similar way I could add a "mechanical image stabilization" option which would display "please use a tripod" on the screen.
• Olympus, indeed, refers to the "digital image stabilization" in the recently announced Stylus (Mju) 810 point-and-shoot. This is applied to an already saved image file (!), but based on the camera shake information recorded off a built-in gyroscopic sensor while the picture was being taken. It may sound ingenious, but I'm skeptical: if there is no detail information in the original image, you will not create it in postprocessing, period.

My bet would be on the third explanation, but I cannot be sure at this moment; at least not before the E-330 manual becomes available, if not the actual camera.

A coincidence? The OIS (Optical Image Stabilization) system by Panasonic, compensating for the camera shake by means of dynamically adjusting a corrective element in the lens, also uses a gyroscope to generate the reference information. And Olympus and Panasonic are getting really close for some time...

Instead of conclusions

The merger of best features of the E-300 and E-500 would have made the E-330 a great camera on its own, without even live electronic preview. It could be seen as a less economizing alternative to the E-500. (To be true, I am still not sure how the MOS sensor will withstand the comparison with the CCD used on previous models, so I am saying that assuming that it will do a comparably good job. Olympus has a good record here.)

With the live preview Olympus enters the ground not trodden before. There is no such feature on any of the competing models. The implementation can be criticized for being a bit kludgy, the preview quality may be not at par with that on non-SLR cameras — so what? Take it or leave it; nobody else offers this feature on an SLR. Turn the thing off and you are where everyone else is.

On the other hand, the Mode A preview can be useful in odd-angle shooting; certainly better than nothing. I also believe Mode B may be more useful in critical macro work than focusing through the viewfinder.

The technology may be still immature, and I'm sure it will keep improving in time. Waiting for it to do so does not make sense: you will die of old age before you get the "real real thing". It is better to start taking pictures in the meantime.

I consider this camera a very attractive offering (as all other E-System models from Olympus), in addition to being quite unique. If you have an E-500 (or an E-300), there is no urgent need to upgrade, though (unless you badly need live preview); if you are thinking of moving into the SLR domain (or switching to Olympus from another brand), the E-330 seems to be well worth consideration. Still, the final recommendation can be done only based on images the camera delivers (this is all about pictures, remember?).

Update: a sibling, the Panasonic DMC-L1

The Japanese giant Panasonic (Matsushita) caused a stir at the PMA show, announcing its own SLR, the DMC-L1. It turns out that this model shares much of the technology (and even some parts) with the E-330, at the same time being an entirely different camera (this is not like Samsung re-branding a Pentax model).

The DMC-L1 has the same dual viewing as in E-330 and the same MOS imager, but it uses an entirely different image processing engine, and the camera's controls are very different. Actually, it looks more like a rangefinder Leica (or, say, a Voigtländer Bessa) than anything else, with external, mechanical controls for shutter and aperture, and elegant shape, both understated and functional. And yes, it takes Four Thirds lenses, bringing its own, Leica-branded standard 14-50 mm zoom to the table.

Web references

At this moment, in addition to official Olympus information and specs, there are a number of previews available, based on pre-production cameras. Here is the list.

• Olympus America offers the usual sales pitch but also specifications; I prefer the global Olympus E-system site, providing less fluff and somewhat more information.
• Previews on the popular digital camera sites: Steve's Digicams, Digital Camera Resource, Digital Photography Review, and Imaging Resource.
• A hands-on preview by Iker Morán on the Spanish site quesabede.com (in English), intelligently written and focusing on the live view feature.
• A hands-on preview by Alan and Mario on myfourthirds.com
• A very interesting interview with Ms. Smith Clemens and Mr. Pelkowski, Olympus Product Managers, posted at Digital Camera Info, and devoted exclusively to the E-330.
• A first impression by Terrence V. Smith of Malaysia, who seems to be more into picture-taking than specs-digging.
• A brief but meaningful and informed commentary by Jason Busch on his Digital Dingus site. (You can always learn something new from Jason.)
• A hands-on, in-the-field review by Danijel Turina on the Croatian fot-o-grafiti.hr site: critical and enthusiastic at the same time, and all from a photographer's perspective. To quote Danijel:

  [...] the improvements are partial but still enormous. When they work, they turn [the E-330] into a unique tool, a class of its own. When they don't [...], it is “only” an SLR.

• Evolution of the Live Preview in Digital Photography by K.J. Kabza at Digital Camera Info.