Canon EOS 1D mrk4: First Impressions

Canon EOS 1D mrk4: First Impressions

Canon 5Dmrk2, Canon 1Dmrk4 and Sony EX‑1 — all shooting video simul­ta­neously in our kitchen…

First, let me underline: this is not a test. I got to play around with the latest Canon flagship for c. 48 hours over the weekend. Shots some stills (about a thousand or so), shot some video (c. 30 min), played around with the settings, trying to figure out if this is such a great camera everybody is buzzing about.…

I focused on couple of issues, things which matter me the most, things where I have hoped to see some impro­ve­ments and which would make my job easier. I am sure somebody else in different line of photo­graphy would focus on something totally different, but this is what I found important and significant.

For those who do not know my work, it consists of mainly sports events (both national sports and major events such as olympics, wc, etc.) and also lots of multi­media work lately. I have shot Canon for over 20 years, and I am pretty familiar with their lineup of products. I guess I have a reputation as some sort of a techno geek in this country (which I consider not totally merited…) — at least judging from the calls I get from friends and colleagues — daily — asking about all kind of technical stuff — be it cameras, videos, Macs…

I presently shoot with Canon 1Dmrk3’s and Canon 5dmrk2’s. My present video favorite is the Sony XDCAM EX‑1 (see image). So that is the background I compare the mrk4 with.


First, I was interested in the video capabi­lities of this new baby. It shoots full HD (1920x1080) as well as 1280x720 and 640x480. Both the NTSC and PAL rates i.e. 24,25,30 and the smaller resolution in 2xFPS i.e. 50 and 60 fps. With the 5D one of the major flaws in Europe has been the lack of 25 frames per second, as this means always recounting the framerate.

To make it short, the camera performs beauti­fully. I love the ease of operation (consi­dering, it is a DSLR shooting video, not a videocamera). I won’t go into what it can do, you can find it elsewhere.

My way of shooting was that I shot thru the viewfinder as if I were shooting stills (focusing, framing, exposure), and to engage video, the only thing you had to do is to hit the FEL-button. To stop, hit it again — or to stop and leave the videomode, to push the set-button.

While you’re in the video mode, naturally all the focusing, exp. correction etc. is done thru the viewfinder which BTW is so much better than the mrk3 (which, as such, doesn’t tell that much… ;-))

But I thought it might be interesting to compare how it relates to my favorite tool for shooting high quality video and that is Sony XDCAM EX‑1 — and compare these two with Canon 5d mrk2 (30fps, but will have a new firmware in Q1 enabling it to shoot 24 and 25 fps fullHD video).

So I rigged them all up in our kitchen, attached more or less same focal length lenses and set them to record the same scene (which was me cooking dinner for the family…)

Two interesting notions: the ambient light was just enough for the EX‑1 — iris fully open — to have good dynamic response. To achieve the same illumi­nation for the Canons, I had to set the ISO to 2000 and exp. to 1/30 and 4.5 for the 5D — and 1/30 5.6 to the 1Dmrk4.

This means that a) either the 1Dmrk4 is c. 2/3 stops more sensitive or b) the 5Dmrk2 is 2/3 stops worse — depending how you look at it. Image quality from both cameras was impeccable, totally in par with EX‑1, even better, maybe… I am careful not to make too much judments about the quality as the mrk4 I was using is still pre-production model and with an early firmware (5.8.5).

I raised the ISO to 6400 (stopped down accor­dingly) and image quality was definitely accep­table for you basic ENG-work; yes you could see the increased noise, but it did not get disturbing. Had I raised gain in the EX‑1 accor­dingly, the changes (for the worse) would have been more noticeable and disturbing. Now, you can easily do the math: consi­dering the Canons were (depending on the lens you’d choose) maybe 2 stops closed to begin with and you could raise the ISO at least 2 stops without having too much trouble with the noise… we are talking video­graphy in circums­tances where previously.… well, at least it would have been difficult and definitely looked noooo-isy.

One of negative surprises I got is also related to this: maximum shutterspeed 1/30 i.e. no slow shutter possi­bility. Don’t know the technology inside the Canons that well, but I could imagine, that especially the mrk4 with two Digic 4 — processors could handle that if it were intro­duced later via firmware update (EX‑1 has slow shutter option, naturally)

One could compare iso’s, shutter speeds and apertures numerically naturally as well, but it would make very little sense as all the three cameras I was playing around with had differing sensor sizes and thus the effects on DOF would produce different results on each model.

I trans­formed all the Canon video thru MPEG Streamclip (from Square 5 software), EX‑1 with Sony’s own XDCAM transform, and cut with Final Cut Pro 7. Worked like a charm, smooth workflow — even on a laptop (Mac BookPro 17″). With 5D you still have to do the framerate conversion and that does slow you down . So we are all waiting for that firmware…

I thought of publishing the result… but I don’t think so, in the end. First of all, me cooking pasta for the family in shorts.… not a pretty sight. More seriously, the results are not suited for accurate compa­rison, as I used a bit different lenses on the cameras (although more or less same focal lenght). And thirdly, the 1Dmrk4 is pre-production and one should not base judment on that.

Having said that, I can add that all three cameras performed in par with each other and it would be hard if not impos­sible to pick a winner between them. But there was something about the 5D image that I just loved… but could be totally my imagi­nation. I have to hear what DSLR gurus such as Philip Bloom and others will have to say on the subject.…


It was a nice surprise to notice how well the built in mic picked the ambient sound. Not that you could use it for anything per se, but it is enough material to sync the audio obtained with an external recorder (such as Zoom H4n) using plural eyes software. There is a miniplug, manual controls, etc.… but I am not a great fan of audio within a DSLR. The afore mentioned Zoom for instance does not cost you a fortune, but we are talking absolutely superb sound quality here. Possi­bility for mixing, external mics, phantom, etc.

I am sure the built in recording function works nicely, but as I said, it is not part of my workflow.

High ISOs

I only shot jpegs as Adobe has not updated their converter (Photoshop and Lightroom) to accomodate this camera — and I simply did not have enough time to start playing around with Canon’s proprietary software. As this was only testing, jpeg would tell me all I needed to know — and much faster.

To give you a baseline: in 1dmrk3, shooting an evening event my highest accep­table iso is typically 2000 or 2500 asa. 3200 was too grainy to my taste, consi­dering you only achieved 1/3 of a stop.

Well, my new upper limit is 12800 — even 25 600 asa. Totally accep­table for your basic news/sports work. The two highest ones — reaching over 100 000 asa tend to generate too much noise to my taste. But then again, I ‘m sure there are cases when these can be used and they prove themselves useful — also consi­dering that the file produced by the camera is rather large and there is room for downscaling and getting rid of some of the problem.

I tested with really bad lighting (a dance class and a practise ice-rink). I was waiting for good perfor­mance and I was not let down in my expec­ta­tions. When in the past “my indoor sports default” was 1600 asa, it is easily now 6400 — even more. In other terms: compared to mrk3 the mrk4 has 2–3 stops more sensitive sensor for your accepted noise­level. And those of you who come from mrk2; the diffe­rence between mrk2 and mrk3 was of the similar ballpark, maybe a bit less.

Canon has claimed that the jpgs coming out of the camera are more “ready” than with the previous model, and based on my short trial run I agree. But with mrk3 I shoot 95% raw so I am not an authority on this one.And when it comes to jpegs, what I would like to have — and to my unders­tanding the camera does not do — is a post cropping possi­bility — in the camera. That would be a nice feature when working with extreme deadlines and shooting directly to a webpage or a gallery which gets published without any editor in between.


This is a tricky one. After the turmoil with 1Dmrk3 this is a hot potato to say the least. Canon says they have totally redesigned the autofocus system. They say it is totally different — and I agree, you feel it immediately. Is it better? Well, depends what you are after.

First, let me warn you: I am totally happy now with my 1Dmrk3’s. They are blazingly fast and accurate — that is, now, after three major repair cycles, mirror boxes changed — the last one only half a year ago (and yes, you hear the discontent there…).

But, they really work. They react so much faster than my 47-year old reflexes. No complaints there — and not hiding behind my age — but point being: I analyzed my images from an event some weeks ago and each focusing error was mine, not the camera. The trouble with mrk3 is that it can be too fast for your average user. It reacts to sporadic movement (the only camera­model which does that), which can be a puck or the ball or sudden movement of the active focus point to to background, spray of water, whatever…

1Dmrk4 has some intel­ligent algorithms built into it. If there is sporadic movement — or mathe­ma­tically signi­ficant change in the focus — it will take another measu­rement before it accepts it. At least if I have understood the manual correctly.…

In practice: the focus is very stable and smooth (and we are talking AI, naturally). I shot in an ice-rink focusing on one skater while others got between me and my target. Unless the obstacle was in the line of focus for a prolonged time, the camera continued following the appropriate target as soon as it was in sight again. Mrk3 would have jumped focus immediately to the obstacle.

The most stable and accurate autofocus up to date. I absolutely agree, 100%. But… or is there a “but”?

What does this mean in practise? Frankly, I don’t know. I’d have to shoot half a dozen hockey games with the camera to see what this means in terms of keepers and no-goods… I am sure a lot of people will praise this new focus system, but I am sure that there will be also voices (young sports photo­graphers, with their reflexes like Usain Bolt — at least they might think so…) who might think that it reacts too slow. (At the age of 20 I would have complained… but then it was manual focus all the way…)

Just as 1Dmrk3 can be adjusted in terms of sensi­tivity and speed, so can 1Dmrk4 — and I had not time to do too much fiddling in those adjust­ments. So I am very careful not to make too bold state­ments to one direction or another…

And also, let me put it this way: if there had not been this trouble with 1Dmrk3, I would not be this careful about this. But — there was and thus — I am. More testing needed, but I have to return the camera tomorrow, so for my part it won’t be now.


So, 1Dmrk4 is lot of money. Do I want it? Yes. Do I buy it? Yes. Why? High iso’s and superb video capabi­lities incor­po­rated into the same, solid profes­sional tool.

Would I buy it just to replace the focusing of my mrk3? No, as I am totally happy with them, they work like a charm. Would I buy it if I just wanted to shoot video? I don’t think so… I’d wait for the firmware of 5D — or get a videocamera separately (depending what you are after: Canon HFS10, Sony EX‑1, Sony EX‑3… and what your budjet is, naturally).

But: as a profes­sional tool for what I do — and what I like to do — I have to say: it seems to be the best piece of gear I have ever seen.

(UPDATE: see my later posts of the perfor­mance of the production model Canon EOS 1Dmrk4 in the Four Hills Tournament)

Leave a Reply

Your email address will not be published.