A couple of weeks with the Sony ɑ55
I was reading through this article on cnet and the following statement sort of jumped out at me:
"Many people like the optical viewfinder, but from a technological point of view we think [translucent mirror designs] are better."
"Better" in what way, and according to who? Before the Sony SLT cameras hit the market I had never seen an EVF that I liked, and since I own a DSLR with what is arguably the best OVF in the business — the ɑ900 — I found myself wondering about the direction that Sony will take with the ɑ900/850 successor. The following post details my experiences and thoughts about this new “wonder-camera”.
There is quite a bit of commotion on various forums about hints that Sony has already produced the last of the high-end cameras with pentaprisms. There is certainly a lot of interest in the ɑ33/ɑ55, but it seems that I am not alone in my willingness to pay extra for a top-notch camera with an excellent OVF. Many of those interested on pro forums like GetDPI.com are buying just to play with a new toy and have no intention of going all-EVF in their kit. I have already voted for OVF with my $$$, but the camera will not last forever and it will need to be replaced at some point. What is there for me to look forward to in the future of A-mount? From the statement below it seems that Sony has already decided that "serious photographers" will have to be satisfied with EVFs if they are to continue using their ɑ-Mount lenses:
"Based on initial feedback and our own study, we have decided to utilize this unique technology for an advanced model of the amount camera," Turo Katsumoto, head of Sony’s Alpha business, said at a press conference at the Photokina show here. "We are confident this camera will satisfy the needs of serious photographers."
I am very skeptical that such a camera will satisfy my needs and in the interests of forming an informed opinion I arranged to rent an ɑ55 for about a week in mid-October for a multi-day event where I planned to shoot with it alongside my ɑ900, both day and nighttime shooting. My intent was not to directly compare IQ but to evaluate how the EVF worked out for me in real-world circumstances. Unfortunately the camera shipped late and I was not able to use it for the event.
When it did arrive, the very first thing that I noticed taking it out of the box is that it is really a tiny camera. The first thing I did was started the battery charging. In the interests of giving the EVF system the best chance I started off with my venerable Minolta 85 f/1.4(most light to the sensor) and experimented with how it felt in the hands and took a couple of shots of the setup using my phone(sorry about the poor iPhone picture quality):
Now, I have fairly large hands, but not unusually large for my 6′ frame. I was unable to grip the camera with all of my fingers, my pinky ended up curling under the camera body. After about 15 minutes fooling with the camera that finger started to become pretty sore. I was able to mitigate this somewhat by supporting the lens with my other hand but I really feel that Sony should have considered adding a grip option for the camera.
The next thing that I noticed is that the power switch is close enough to the same location relative to the shutter button as the front control wheel is on my ɑ900. I can’t even begin to guess at how many times I inadvertently powered the camera down when I wanted to make an adjustment. I’m sure this is something that anyone could get used to but I for one prefer the power switch on the left as on the ɑ900.
Once the battery was charged up I installed it along with a 16GB SD-HC card. I then installed a generic Wimberly dovetail plate and noticed another gotcha: The plate blocks access to the battery/card door. If I was planning to own the camera I would cut the plate down or find a smaller one. The Wimberly plate is designed with a gentle curve in it that ensures that the ends of the plate engage the camera body with useful force though, so it might be necessary to tighten the shortened plate a bit more than you should otherwise have to. Also you would lose the safety bolt on the one side that is intended to keep the camera from slipping out of a loosened clamp. I suppose this sort of thing is to be expected on such a compact camera.
Once I started shooting I was immediately reminded of why I have stayed away from APS-C. My 85mm basically became a 127mm and despite the larger-than-usual-for-APS-C viewfinder I felt very cramped. That’s a very personal thing though, if you’re used to APS-C, especially the Sony cams with the pentamirror-LV system, it will be an improvement in size. I found that in my home at night it as very difficult to track a cat moving across the living room because of the extreme EVF lag under these conditions. Remember, this was with a f/1.4 lens. If you’re wondering whether I could even get a shot in these conditions, the answer is “Yes”. With the ɑ900 and a bounce flash, I can both frame and expose just fine. With the ɑ55, not so much. In low light it takes the sensor so long to form an image that the EVF only refreshes 4-5 times per second(in my living room that is; conditions vary of course).
Another thing that I noticed right away is that in good light the EVF appears to be very washed-out. In very bright light it gets so bad due to light entering the VF that it’s difficult to frame your subject. This may vary depending on the shape of an individual’s face, whether one is wearing a hat, glasses, etc. I heard the same thing from Gary Friedman after spending a day shooting with him where he was using the ɑ55. Certainly it’s better than an LCD like on the NEX, but at the same time it doesn’t solve the problem of shooting in bright light 100%.
Also, at the request of one of the other moderators on the Sony Camera Club Forum, I took a look at how well the EVF could pick out stars in the sky. Here is the response that I posted there:
Finally cleared up last night enough to try this. Actually after all of the rain it was VERY clear. Yes, I was able to see the bright stars in Orion through both the EVF and the LCD. Playing with DOF preview, they were very hard to pick out by f/5.6 and just about gone by f/8. With “normal” haze I would expect them to be just about gone by f/5.6.
Worth mentioning though, the display was gained way up, extremely noisy, and as a result the stars exhibited enough blooming that it was very difficult to get focus nailed, even with magnified LV.
Also, even with EVF brightness at minimum, after trying this I was blind in my shooting eye for about a minute and impaired for several minutes while my eye dark-adapted again. If I was trying to move around on a rocky hillside at this time it might have been a serious problem.
With initial impressions and basic tests out of the way, I went on to investigate one of my biggest concerns with the EVF, the ability to track fast action. I have had a great deal of success with OVF cameras, both film and the ɑ900. The question I wanted to answer was, would the EVF lag that I assumed was there impact my ability to tightly frame and follow fast-moving race cars as I did when shooting the 2010 Monterey Historics? The majority of those shots were taken at the entrance to turn 5 at Laguna Seca, where most of the cars are slowing to around 100mph. Since I couldn’t get out to the track for my testing I chose a point along a local freeway where traffic normally moves at 70-75mph. I positioned myself on the frontage road – Arastradero Road – near the middle of this view, at a spot where the fence was relatively low and I could shoot over it(I had to do essentially the same at Laguna Seca). I was shooting all of these in 10FPS mode, auto-review off. My first and second attempts gave some indication of what was to follow:
After reviewing the first set, I could see that the camera was showing me where the car had been, not where it was, so the car tended to creep off to the right of the frame. In the second pic, I tried to compensate and ended up overcompensating. This next pic was a few attempts later. Note that for this type of pic I try to concentrate on the driver; that’s the part of the frame that I want to be sharpest and usually near the center. In this next shot I tried leading a bit to give the vehicle room to “move into the frame”. This time I managed about 8 shots before losing track due to the lag:
In this shot you can see a bit of the back-and-forth adjustments that have to be carried out to keep the subject centered – and the number of lost shots from the sequence due to never really knowing exactly what is going to be captured:
Somewhere around the middle of my testing I managed to nail one set — but only one set. I was not able to reproduce this effort. By this time I had stepped back from trying to shoot over the fence because I had found that awkward and I had decided that I was no longer concerned with the fence in the foreground or with focus, I was only worried about framing — having said that though the camera focused on the subject pretty well:
One side effect of this need to readjust while panning is that if you are readjusting your panning speed you are no longer following the subject, so there tends to be additional horizontal blur introduced while readjusting. Here is one such example:
The above photo was #6 in this sequence. Not the best sequence, but it illustrates the problem, which is visible in essentially all of the photos taken while I was trying to adjust for the lag.
Now, some will say that this is something that you just have to get used to. While that may be true, I found the effort to be a real chore and IMO getting so few well-tracked sequences out of 54 attempts is a real deal-killer for me unless Sony can make several quantum leaps in the performance by the time I am ready for a new camera. Actually the real number was closer to 70 attempts but I only posted 54. Many of the ones that I didn’t post were interrupted by a full buffer, having been attempted too soon after one of the long sequences that I’ve posted. To give an idea of the number of attempts made, consider this image:
Note that I made a Photoshop action to compile these grids… and it still took quite a bit of time. If you disagree with these results, or are convinced I was doing something wrong or that you can do better, don’t just say so – post a link to your examples. If you want a copy of the PS actions they are here. The full gallery with all of my results is here.
The “Contrast Loss” that physics says must be present because of the design of the beam-splitting mirror is so slight as to be very difficult to measure. I shot a number of test shots with the 135mm f/1.8Z, 70-400G, and 28-35 f/4-4.5(three of the four ways I could get to 135mm and compare my sharpest lens with others) but even with the MF-check LV it was very difficult to get the identical focus point between shooting with mirror in vs. mirror out. Some folks online pointed out the focus errors that I had missed since I was directing my attention at the point where I had focused in the images, and the differences in focus were only apparent by looking at the rest of the image. I re-shot the tests and found that with variations in focus point — even when shooting with the greatest of care — make it very hard to determine if there is a visible difference or not. At a certain point I realized that could only see at most a difference of 2 or 3 on the RGB scale of 0-255. Well within the range of human error/metering error at the shooting stage. I don’t have the tools or the patience to dig any deeper. Kudos to Sony for doing such a great job of minimizing this contrast loss.
Overall, I think this is a fine camera for what it is intended to be — an upgrade for users of higher-end P&S cameras and/or a midrange DSLR-like camera packed with convenient features(especially for this price point). If you don’t normally shoot in contrasty conditions, extremely dark conditions where dark-adaptation is important or fast action where the EVF is a greater hindrance than help, this camera is worth a look for its great IQ and feature set. The 10FPS capability may be just what you need for situations where the subject is not moving too quickly across the frame or where you are happy with framing your subject very loosely to compensate for the EVF shortcomings. Picking up the camera in a store and looking through the viewfinder for 30 seconds is not going to tell the whole tale.
Note: The pros and cons listed below are solely my opinions and the fact that I found more cons than pros simply reflects what is important to me(minus some things I left out because they are rare at this price point, like dual control wheels). Different photographers will of course have different priorities. I did not delve into video, in-camera HDR/Pano/etc. I am not a professional camera reviewer, but even if I were I would still suggest that you get your hands on the camera and make your own assessment. If you have any doubts at all going in, consider renting one.
- Very compact and lightweight (maybe too much so for some)
- Very quiet shutter
- No mirror slap makes for sharper images when handheld at low shutter speeds
- AF is snappy and responsive with most small lenses, though a bit underpowered with heavy lenses like the 1st-gen Minolta 80-200 f/2.8 APO.
- Live histogram and other displays/grids in EVF is a bonus
- Protruding eyepiece handy for those with large noses
- EVF lag makes tracking fast subjects a challenge and a chore.
- EVF limited DR blows out highlights that will in fact be captured.
- EVF limited DR obscures facial expressions when attempting backlit portraits with fill-flash
- Small grip and overall size makes handling with any but the smallest lenses awkward.
- Power switch is in a bad location(IMO), too easy to turn off when your intent is to spin the control wheel.
- Unable to access battery or SD card with Wimberly standard dovetail plate installed(though you could cut a plate down).
- EVF all but unusable when shooting UniWB
- Flash as focus assist beam is very annoying for subjects and often results in subjects thinking picture has been taken and moving before the shutter fires.
- EVF is a bit grainy for MF, though if speed is not an issue MF check magnification is available. Not so helpful for MF candids, handheld macro, etc.
- EVF color tearing due to sequential refresh
- EVF has a severe negative impact on dark-adapted vision(amateur astronomers will know what I mean by this, it can take 30 minutes or more to dark-adapt again)
- Shooting menu not as streamlined as quick-nav on ɑ700/850/900
- Protruding eyepiece surround may fall off if carrying camera against your body with a strap – gets hung up on clothes when stooping over.