Review by Sean Chan - Website. © 2006.

Basic Assumptions

Keeping in mind that I'm a beginner, my review will be based on my (lack of??) photographic experience. Also, note that how well suited a camera is to a photographer depends on his/her usage. For example, I don't do sports photography, and I'm learning landscape/scenic subjects. What I do enjoy and think I'm getting close to getting right is non-posed portrait photography... by this I mean capturing images of people]in their "natural" state. And that's what I use my camera for mainly.

Auto Focus

45 point system

I noticed many people who are not familiar with the EOS-3's 45 point focussing system ask, "Isn't this confusing with all those points out there distracting you??". Well, actually all those points are invisible in the viewfinder until you select one with your eye or with the dials (or if the camera selects it/them - Automatic Selection). You get a red LED lighting up at the required point. They are superimposed on the viewfinder. So, no it's not distracting. Also, for those who are not familiar with ECF, it does not light up everywhere you look at. It lights up only when you press the shutter halfway. That way it knows when and where to focus.

So, does one really need 45 focussing points?? Well, I think the obvious answer to that is "No". Is it a nice-to-have feature?? This gets a little bit more detailed because the answer depends on your usage.

Firstly, to have 45 points, ECF is think is a little more crucial because can you imagine manually selecting 45 points with the dials?? I don't think so. In my opinion 45 points without ECF is useless. Secondly, if you get ECF working, is 45 points more useful?? Well it can be if your subjects move a lot - in theory ECF was designed mainly for sports/wildlife where subjects can move in all directions erratically.

To quickly summarise - No I don't find it particularly useful, in fact I use CF13-1 which cuts the number of points down to 11 and that's good enough for me and my kind of photography. The AF area, however, covers a large proportion of the viewfinder and has reduced my recomposition greatly. Psychologically, you grow accustomed to the AF and don't recompose so much because you know there's an AF point there.

 

Auto Focus Speeds and Limitations

AF of lenses - USM Technology

What can I say?? Before USM I told myself "no biggie"...so what if it's a little quieter or faster?? Seriously though now that I've got 2 USM lenses, there is no way I'm going back to loud and slow AF. Now even Nikon and Sigma has their version. I've been told that USM alone has managed to siphon a lot of Nikon sports photographers to Canon. Now Nikon's trying to win back some of those lost clients with the AF-S technology (I won't discuss the USM vs AF-S price/performance issue...that's for the pros). All I can say is try before making a judgement. Note that the EOS-3 is not compatible to all non-Canon AF lenses. Some models of Tokina and Sigma AF lenses have to be re-chipped to be compatible to EOS-3. If your 3rd party lenses are under warranty, this re-chipping is free. If not it will cost you Just be aware of that fact. Tamron's are ok.

AF on EOS-3 body

My gripes about this later. But first AF on the EOS-3 is faster than AF on any other EOS bodies. However, to me and my kind of photography, it's a non-issue. We're talking about a couple of milliseconds faster here. For sports photography it's important though.
Ok, back to the 45 focussing points. Canon specs says all 45 points have horizontal sensors (i.e. detect vertical/diagonal edges) of lenses to f5.6. Seven of these sensors (in a vertical formation) have cross sensors (i.e. detects vertical/diagonal/horizontal edges) enabled when you have a lens f2.8 or wider aperture. The central sensor is the most useful having cross sensors to f4, or horizontal sensors to f8.
In practice, what does all this mean?? You put a 50mm f1.4, you have all cross sensors/horizontal sensors enabled - you can AF any orientation of edge. If you put on a zoom like the 28-135mm f3.5-5.6 IS USM the story changes. From 28-75mm the open aperture is f3.5-4.0 so you can use the central sensor for all orientation of edges. Above that, the aperture is f5.6, you have no more cross sensor capability. You can use any of the 45 points to AF vertical or diagonal edges. "So what??". Well, I take pictures in low light, usually largest aperture of f3.5 or f5.6 and shutter speeds of 1/4 sec - 1/20 sec (Image Stabiliser enabled depending on focal length) with this 28-135mm lens. Believe it or not, it hunts around a lot with AF because it can't lock on. In the same scenario with the 70-200mm f2.8L, it locks on immediately. Obviously cross sensors picked up the edge that the horizontal sensor couldn't. Before my 70-200mm f2.8L, I cursed and swore at this limitation. But now It's not so bad...I've calmed down.

My sore point is, to have full practical use of AF, the EOS-3 cries for large aperture lenses,primarily "L" luxury lenses - i.e. $$$. Nikon has taken a better approach, cross sensors to f5.6. This means that you can use cross sensors even on majority of their consumer range zooms. Hope Canon learns here. Sure Nikon sacrifices horizontal sensor AF at f8, but how many people will be using that? I know sports/wildlife will be using teleconverters on 300mm f4's, etc. but to me that's the minority of the market...sorry to be harsh here.

Oh yes, one more thing...I performed another test with Canon's EOS-3 AF sensors vs. EOS-RT's. Believe it or not, EOS-RT's a wee bit more sensitive than EOS-3 in low light. I could AF some subjects with the EOS-RT that the EOS-3 couldn't.

 

Eye Control

This is probably the most talked about feature (by non-Canon users) in the Canon EOS cameras.
The EOS-3 is my first ECF camera - so I cannot do any comparison to Elan IIE/50E or A2E/5.

A very quick history about EOS-3 and ECF when I first got it back in December, '98, I wore glasses then. ECF took about 2-3 weeks before started to work consistently. During that time I calibrated it 20-30 times till it didn't want to calibrate anymore. Then my glasses broke - back to square one.
I stared wearing contact lenses next. ECF took only a 5-6 days to get to where I got with my glasses in 3 weeks. Obviously distance of eye to viewfinder was a major contributor to this improvement. Even though it's working relatively well for me, it's not good enough for fast moving subjects. Technique plays a very important part in accurate ECF.

As such ECF is not something that works straight away for everyone. You have to be patient and calibrate/re-calibrate over and over again till it works fine. Then it's not only that the camera learns your eye, but you have to learn how to place your eye in the same position every time if you want accurate ECF.

Here are a few pointers to have successful ECF (may not work for everyone) :-

1. Relax your eye, don't squint
2. Practice keeping your eye still, like looking at your lover right into his/her eye...I'm serious!!
3. Keep your other eye open, just to keep your eye muscles more relaxed
4. Practice putting your eye to the same exact spot on the viewfinder. Don't tilt your head; you may forget this when you're leaning out on a railing to get a shot. Basically, your head and the camera should be perpendicular.

That's basically it!! Calibrate it as you would take a photograph.

By the way, with CF13-1 where you cut down to 11 focussing points, ECF works even better. I guess this is due to the separation between the points are greater and the camera is able to pick up the point differences easier.

 

 

Metering

Boy, there is a lot to be said here especially about the underexposure, so I've reserved a separate sub-section for it. But firstly, I will go through the different metering modes...with the basic assumption that metering works perfectly.

a) Evaluative

Let's face it, what is a metering system?? It uses some sort of method of calculating what it thinks to be the "perfect" exposure for a particular scenery and tells you the "correct" shutter and aperture. So it's merely a guide. The different kinds of metering system (evaluative, 3D matrix, centreweight, etc.) are just different methods to reduce the risk of incorrect exposure...but the risk is still there.
Having said that, the results of my photos seem to indicate higher chances of "correct" exposure with the 21 segment evaluative than with centreweighted. So, this evaluative metering system is doing its job. But in comparison to the EOS-RT's 6 (or is it 5 or 7??) segment metering, I'd say that it's not that much better. Maybe out of a roll of 36, with EOS-3 I may get all correctly exposed, whilst the EOS-RT may be 34. This isn't much to me considering I've got crap composition :-) But to a pro who may shoot rolls and rolls a day, the 21 segment evaluative metering may save his/her day.

 

b) Partial

Seems to be the same as all the other EOS's. No biggie.

 

c) (Multi)Spot

My biggest attraction to the EOS-3 is its multi-spot metering capability. I have been absolutely pampered by the Olympus OM-4Ti with this feature, I vowed that the next AF SLR I was going to get must have this feature, which the EOS-3 did. Seriously, use this with care and it blows 21 segment evaluative metering away, at the cost of time.
Just like the OM-4Ti, the EOS-3 metering is right next to the shutter button. So convenient to press and meter various points you like, and the camera will average it out for you. Unlike the OM-4Ti, the EOS-3 meter will take 8 different spots only - whereas the OM-4Ti will drop the first reading and cycle through the spot queue.
The one thing I found confusing with EOS-3's multi-spot is the reading for each spot. The OM-4Ti is different because it's Av only and shows a bar of shutter readings. Each spot is marked against a shutter speed setting. So you know definite values. The EOS-3 marks each spot against the exposure compensation bar and displays the averaged reading in the viewfinder. So you have to manually calculate yourself what the actual exposure was for each point. Mind you too the scale is only +/- 3 F-stops. beyond that, it doesn't show. But this is a minor issue that one will get used to after time and time of usage. Besides, all you really care is the final averaged reading anyway.
Just remember, no matter how useful and flexible multi-spot is, it can only do so much in to deciding the best shutter/aperture combination. It cannot solve all situations. You are going to encounter certain scenes where it's just too contrasty to get everything in the viewfinder - there's only so much your film latitude can handle.

 

Flash Photography

I won't bother telling you the terrible luck I've had with flashes. Let's just say that with my earlier 3rd party flashes, I got washed out overexposed shots which disappointed me terribly. Needless to say, I was dying for a Canon flash for the EOS-3 - none better than the 550EX.

 

Recently I finally got my wish and got my 550EX (and the 70-200mm f2.8L) and took some shots at a wedding. I was shocked with the results - underexposure. And this is after I've had my camera meter tuned up for flashless underexposure problems (talked about later). Anyhow, I performed a few more tests and it appears to be -1/3 to -1/2 F-stop flash underexposing. Easily (temporarily) fixed with Flash Exposure Compensation (FEC) dial. The tests involved various distances, various camera modes, including CF14-1 (which stops auto flash reduction - e-mail me if you don't know). I've yet to complete my tests but with FEC, so far ok.

Let's just say flash technology is where I have to relearn or try to re-understand the concepts. I came from a background where you needed an ISO/distance table, and set the aperture appropriately - known as Manual flash (as I understand it). Then with the OM-4Ti came TTL-OTF, supposedly clip it in and point and shoot at whatever aperture I like (OM-4Ti's aperture priority) but disappointing overexposure results. Similarly with my EOS-3 and 3rd party flashes. In short, I was using these SLR's and flashes like P&S (switch the camera on in P mode, clip the flash and snap away). Noooooo....it had to make my life difficult. So after my flash underexposure problem, he tried using my system. He's coming from a Nikon background. Which is different from E-TTL. So there's a shift in 'flash technique' again. We used to understand that if either you or the camera knows the subject distance, then the flash power will be set accordingly (either automatically or manually) and everything in that distance will be 'correctly flash exposed'. Of course there's a difference in the way the different brand handle flashes too in various metering modes. The Nikon's Av mode allows you to set your aperture but keeping the shutter at 1/60 for flash sync. Canon's Av mode allows you to set your aperture but your shutter goes way down to meter ambient light. So no more handheld - need a tripod. If you want to have 1/60th shutter and aperture control, M mode is the only way.
If only it were true. Now they talk of linking flash to particular segments of the camera's 21-segmented meter. It knows where your subject is and will meter your flash at that point. So youcan't point, AF/AE lock and recompose because the sensor will be pointing at somewhere else already and the flash will be wrong!! Complicated isn't it?? Ok, now I know I can't recompose easily. Then there's CF14-1 to worry about. Canon cameras/flashes are intelligent enough to detect Ev 10 or above and cuts down the flash power automatically as though you want to do "fill-in". But it may not be what you want!! Anyhow, I've decided to forget all and just assume the camera knows best.

 

Underexposure

Body Underexposure Problem

Ok, I must have chucked such a huge fuss about this that almost everyone knows about this underexposure problem. However, I am not the person the first person who noticed this, someone else brought it up in the EOS mailing list a month before I spat the dummy. When I found out my camera was underexposing I remembered this guy's post and did further testing which confirmed his & my 'fears'.
Let me state, that this is not a major technical problem (I can just hear the EOS mailing list members laughing at me thinking that I've contradicted myself). There is a simple technical fix available from Canon Service, which takes less than a day for them to do, though the job queues may be long. Therefore, if you really like the EOS-3, don't let this underexposure problem discourage you from buying it. With the camera fixed, it works beautifully.
What my gripes are back when I first found it till now, is that it is inexcusable for a manufacturer to come up with a "top of the range" (ok, second from the top, but close enough) camera and have a simple fault such as not being able to meter correctly. It sorta reminds me of my dad's new Land Rover with his driver stereo speaker not wired in. Simple problem with a simple fix but completely inexcusable. I have my personal views of all this and it basically ties down to $$$ and profit making, etc. I shan't discuss such issues.
Anyway, as the story goes, some EOS-3's underexpose by about 2/3 F-stop. This is most noticeable with chrome/slide film. I discovered this initially when I went around shooting with my mentor. He used his F5, I used my EOS-3 (it was originally intended to be a test of "composition"). We compared slides and mine looked like crap next to his. When this was discovered, I remembered the post made in the EOS mailing list and we promptly tested the F5 vs the EOS-3 metering. Both cameras were unloaded, he had a 100mm F2.8 macro and I had my 28-135mm. To simplify we both picked F5.6 with my camera at 100mm. Under evaluative, centreweighted and most importantly spot metering my camera was picking shutter speeds faster than his. Using the exposure compensation dial we narrowed it down to about 2/3 F-stop underexposure. Small allowances had to be allowed for differences in lenses and spot metering sizes. Further tests of my EOS-3 vs my OM-4Ti confirmed that 2/3 F-stop was the difference.
Next stop was to Canon...I called Canon Service in Sydney and they confirmed one other report of underexposure in Australia. I can only attribute this to low volume sales in Australia, and not many sold would have noticed it or cared. They told me no software fix or firmware updated needed, just bring it to Canon Service in Perth and they can tune it up. And so I did. I left the EOS-3 body with them and they did a brief test and wrote down "underexposure" in the jobsheet.
Almost 2 weeks had passed, and nothing from them. One day a call came in from Canon and thetechnician wants to see my 28-135mm F3.5-5.6 IS USM lens!! I brought it in and he took me to the back room. There he stuck a 50mm F1.8 lens, placed it in front of an Ev meter, opened the camera back and stuck a light sensor at the back of the shutter blades. At various Ev settings, he snapped away. There was a digital readout showing what the sensor was detecting through thecamera lens/meter. The readings are not constant but varied between -0.35 to -0.55 (presumably in f-stops). The technician was trying to convince me that my EOS-3 was underexposing by -1/3 f-stop (I call this "playing down the problem"). He then proceeded to stick my 28-135mm in and ran through the same tests. Depending on Ev and focal length picked, my lens was underexposing from -0.4 to -1.0!! Anyhow, I told him I wanted a +2/3 F-stop tune-up. He 'reluctantly' agreed, and in the jobsheet, put down +2/3 F-stop correction at customer's request. I was not impressed with such a report because it made no references to an underexposure problem.
But I wanted my camera back quickly so I didn't get them to fix the job description.
The explanation for the differences in underexposure range is due to inaccuracies in "open aperture" of lenses. When a lens aperture is fully opened, the lens tells the body what it's meant to be - f1.8, f2.8, f3.5, f5.6, etc. However, it's often inaccurate where the actual aperture is really f1.9 or f1.7, f2.7 or f2.9, f3.4 or f3.6, f5.5 or f5.7. But the camera then measures the light and calculates the shutter on that aperture value given by the lens. These variances make minimal differences to final exposure if the meter is correctly tuned. But if the meter is not accurate, tuning it can be difficult.

New Flash Underexposure Problem

My newly acquired 550EX flash appears to be underexposing too. I have now taken about 4 rolls of it and with a proper understanding of E-TTL technology, I believe I'm using the flash correctly. The EOS-3/550EX flash underexposes in between 1/3 to 1/2 F-stop in flash E-TTL mode, and EOS-3 in all program modes. The custom function CF14 (which enables/disables auto flash reduction) has no effect on this problem. However under flash manual mode, it exposes correctly.
I rang Canon today and they will have a look at my camera system next week.

 

 

Viewfinder

Although Canon isn't renowned for eye-relief, I personally have no problems. Even when I used to wear glasses. Similarly, with all the ravings of the Nikon F100 being a bright viewfinder, I don't find the EOS-3 to be unusably dark. Face it, cameras these days are so close to each other, one's slightly brighter or darker than the other isn't going to make a great difference to your shot...it's just a "nice to have".
What about all the viewfinder information?? It info at the bottom which tells you shutter/aperture, flash, ECF, focus, and (flash) exposure compensation is easy and intuitive. I would wish the had put the program mode and metering mode in there too (like the Nikons do). Very often I forget I had left the camera in spot metering, or in Av (then using it with flash without a tripod), etc. The vertical right hand side has a exposure compensation bar - in fact, 2 of them. One for exposure compensation and the other for flash exposure compensation. At the top/bottom of each bar are tiny little arrows to show if your shot is over/underexposed. These are so small they're easy to miss. I'd rather look at the flash confirmation LED.
As for the bar, it's a matter of getting used to. Canon users will find it fine but as mentioned before for multi-spot metering, you have to manually calculate what the previous spot exposure was. Overall, I think this vertical info is too cramped.

 

Handling

Apart from multi-spot metering. This is the other primary reason why I selected the EOS-3 over other cameras. Although the F100 wasn't out yet when I got my EOS-3, I had held the F5 and knew that the F100 wasn't far off from it. Someone else wrote that the F100/F5 has lots of levers and buttons all over the place...making it look cluttered. I entirely agree. The buttons and levers deters me from picking it up. Whereas EOS-3's clean look is inviting. Since then I have picked up and held the F100 and although the actual weight balance and grip size/mould feels good, I'll stand by my comment on the buttons and levers. As such if I had to do the choice all over again, I'd still pick the EOS-3.
The issue of "what's more durable and what's not" is to me not an issue at all. Nikon gives the feel of durability, whereas EOS has the feel of plastic. But look how far we are into technology. What you see isn't necessarily true anymore. The same can be said for what you feel. The lack of feel in durability doesn't mean it's less durable. I've read reports of F5's being damaged from a simple fall, I've read the same from EOS-1N's. It really depends on what angle it falls at, the "where's" and the "how's". I won't even say that F5's are more durable than EOS-3's. But please don't ask me to test. I'm not about to scratch my EOS-3 for your sake :-). All I'll say is this. If I were to be shoved into a wall, and my EOS-3 gets sandwiched, apart from scratches, I'm sure it will survive. Overall, I believe it will most knocks and some falls. Why?? I think the polycarbonate will absorb the shock and cause the camera bounce instead.
I also want to say a little more about the ease of accessibility of features. The multi-spot metering button for example is so conveniently located. But easier than the EOS-RT's partial metering (which has caused me to accidentally shoot unwanted frames). Also the back thumbwheel is extremely easy and convenient (much more convenient than the F100/F5's thumbdial). All in all, it is only matter of getting used to. Now that I've used the camera for so long, I find I don't make simple mistakes anymore. This is where I'll stress again that comfort in a camera system is probably very important when selecting body to invest in.

 

Conclusion

Am I glad that I got this camera?? Despite the problems of underexposure, which can be fixedeasily..."Yes I'm glad I bought it". Do I regret not getting a Nikon F100 or Canon EOS-1N?? The answer is "No". I don't forgive Canon for making me go through the pains of fixing the exposure problems because I paid a lot for the camera and expected it to work right out of the box. Instead it made me look like a fool where P&S could have obtained better exposures than my EOS-3. But now that it's fixed (or getting there with the flash), it's much much better.
The weight of the camera is my only concern now but really, do I have any other choice if I wanted another camera with all these features?? Nope. So I'll just have to live with it. But this is also why I'm still keeping my Olympus OM-4Ti system because it's extremely light. You'd be surprised how weight can affect one's mind...my EOS-3 with 70-200mm F2.8L makes me rush my shots because of the weight.