Canon EOS-1D Mark IV camera review

Canon EOS-1D Mark IV (photo from the Mark IV page at Canon Europe)

Update July 19, 2010 – I’ve added some impressions after two CFL football games at the bottom of the page. One was a ‘high noon’ 1pm start under full sun where the Mark III was typically a disaster. The second game was under beauty light (low, direct setting sunlight) for the first half and stadium lights the second half. Overall verdict is that the camera performed very well!

January 28, 2010.

The Canon EOS-1D Mark IV is here in Canada, finally! I’ve added updates inline with the original review, and will continue to do so as I gain more experience with the camera. Here’s what I’ve found so far compared to the Mark III, which I’ve used since mid 2007:

(in no particular order)

– The 920,000 dot LCD display is very nice and I really wish Canon would have included it in the III. It’s possible to zoom very deep into images and check actual sharpness with much less doubt than the 230k Mark III screen. The new tempered glass and filled air gap definitely display a much richer, deeper image. There is much less light reflection off the surface of the display. The viewing angle is also improved with much less of that silvery effect in dark/black areas and far less color shifting as well. Update 2010.01.31: Was on a sunny ski slope for a few hours and the antireflective measures taken by Canon do make the display easier to view, but the screen itself could still benefit from being brighter. Nose grease on the cover glass seems to be a bit more difficult to wipe off.

– Live view benefits tremendously from the added resolution, allowing much, much more precise manual focusing. The live view AF function, that uses contrast detect for AF right off the sensor is surprisingly snappy, when there is good light. I wouldn’t want to use it for critical time sensitive situations, but it can be more accurate than the regular AF system when working with very fast (think f/1.2) lenses with very shallow depth of field.

– In live view the last shot image will appear as it normally does for a couple seconds and it’s now possible to review shot images without having to leave live view (i.e. when you press the playback button the shutter does not close and the mirror does not drop back down like with the III. Just touch the shutter release and the camera will resume the live view feed.

– AI Servo AF: Only had a chance to test it on cars driving past the store at about 50 km/h with the 70-200 f/4L IS. In hindsight I realized I still had the IS on, which I would want to generally turn off for action shots, and after having looked at the sequences, it might have affected some images with a bit of IS ghosting (when the IS doesn’t adequately stabilize an image). However, AF acquisition and tracking felt very snappy, but the results were not 100% perfect. There were some instances where one frame or even several were significantly off. I would really have to test it against the III in a side-by-side shootout, but the initial impression is that the IV is giving better in-focus percentages with moving objects (at least cars).

Update 2010.01.31: Shot over 1300 frames of skiers with the 600 f/4 IS (IS off) using primarily the center AF sensor (which gives cross-AF with an f/4 lens) while experimenting with a few of the IV’s custom function settings. In particular I focused on setting III-2 AI Servo tracking sensitivity at slow just as I have with the III (which worked well shooting football, etc. where the direct line of sight on the subject would often be broken momentarily by another player), III-4-1 Continuous AF track priority in conjunction with III-8-1 or 2. Based on these settings I have some concern about the IV’s AI Servo capability. There were enough bursts with multiple soft frames, predominantly front focused, but also some back focused. Explanations may include some user error due to my inability to keep the AF point exactly on the same spot, though that was the point of enabling focus point expansion so that I could allow the subject to move around a bit in the frame. Some images might have subject/camera movement softness, though I was shooting at 1/2500, 1/3200 and later even at 1/5000 to try to minimize this chance. It could be that the extra resolution of the sensor is amplifying motion blur more than I was used to seeing from the III or IIN. The lens might require slight micro focus adjustment, though there were also significant numbers of acceptably sharp frames… A significant number of skiers wore solid dark or black clothes and the sun was from the side, often resulting in backlit like conditions. Initial impressions are that the camera has some issues with flat/backlit/low contrast subjects, just as any AF system will. I still think it’s too early to determine with certainty that there’s a serious problem, but the IV’s AI Servo, based on how I had it set up and was using it, wasn’t an instant success. It looks like I’ll need to spend a bit of time figuring out this aspect. One aspect I will have to evaluate is III-2, which as mentioned, I usually have set to slow. For something like skiers, where there is little risk of a momentary obstruction, results may improve at normal or fast. Some sequences had consecutive frames that were focused significantly in front of the skier. If III-2 is set to normal or fast it should reevaluate and correct incorrect focus shifts more quickly, if I understand Canon’s logic behind this setting correctly.

Update 2010.02.14: I’ve been meaning to update again, but have been busy. I had the chance to go to Canon Canada and borrow their CPS loaner 600 IS lens and compare it against mine on the IV. It pretty much confirmed that while both lenses are similarly sharp, mind is consistently front focusing on the IV. It explains to some degree the bad results I got of the skiers back a couple weeks ago. I’ve since done a number of tests with the 400 2.8 IS, and have had better results than with the 600. So, I think the 600 & IV will eventually go to Canon for a proper calibration. During AI Servo testing with the 400 I decided to basically start from scratch with the core AF related custom function settings (III-2, III-3, III-4 and III-8). Results were not hugely different between the various custom function combinations, but I’m leaning towards leaving these settings at the default unless specific situations benefit from some tweaking. What I did still see though was some sign of “AF blips” during extended tracking sequences where the focus was more often than not slightly in front of the subject. BTW, Canon has a fairly informative, though hard to find booklet called AI Servo AF Custom Function & ISO Speed Settings Guide (click on the photo below to download the pdf) with tips on how to optimize the various AF related custom function settings.

Update 2010.03.31: Sent the IV and 600 IS to Canon for calibration. Apparently the lens required focus calibration, but it also needed optical re-centering. It’s something I had suspected was an issue even when I was using it on the lower rez Mark IIIs. In certain situations with sharp, high contrast tone transitions, there would sometimes be a very slight, soft edge. The softness was more evident when using the lens with a teleconverter, which makes sense. In any case, I had a chance to use the IV and 600 combo for a running race and am generally pleased with the performance now. AF is much more consistent and images are sharper. So, in this case it appears most of the blame can be laid on the 600 and not the camera.

– General AF accuracy seems to be good. I tested the camera with a couple f/1.2 lenses and here too it worked quite well, also with off-centre AF points. However, with shots of people moving around in normal conditions, the camera/lens combos did at times have trouble keeping focus lock. But, given the very shallow depth of field at f/1.2 at relatively short working distances, I’m willing to accept some soft images because the camera has to work very hard to find and keep the point of focus (I couldn’t really do any better myself).

Update 2010.02.14: Shot an event in very low light (ISO 3200 1/100 f/1.2) and found that there were a number of times when the IV either didn’t want to lock focus or was very slow to do so. One Shot definitely worked better here (as I would expect), but still was slower than I would have liked.

Update 2010.02.25: But I think it has a lot to do with subject contrast (lack thereof). I did a side by side test with one of my Mark IIIs in ambient light around ISO 3200 1/100 f/1.2 – both cameras AFed similarly in One Shot. Both had no trouble locking focus on subjects with reasonable contrast. But AF in both wasn’t very fast. Typically the AF would move relatively fast until it got near the correct distance, then took about half a second to finally decide and light the AF confirm light. AF assist points were disabled for both. Non-center AF points behaved similarly on both cameras.

Maybe it was still too bright, so I tried it at ISO 3200 1/30 f/1.2 light level and both had similar difficulty locking on lower contrast subjects. This time the center point on both did a bit better. While both behaved similarly, AF was slow and in the darker test conditions sometimes hunted, but generally locked onto subjects with good contrast.

In this respect I’m not sure there is much difference between the III and IV at this level of ambient light.

Based on the IV review at canonrumors, the reviewer felt the IV was inferior to the 5DII in very dark ‘candle light’ conditions of ISO 12,800 1/30 f/1.4….

So I adjusted the ambient room light to this brightness and yes, the IV displayed more AF indecision than the III. Two times out of three the III would lock AF, though it by no means was a very fast process. In contrast the IV hunted more, AF drive speed seemed slower and AF lock was less often than the III with the same, relatively high contrast subject. And it helped greatly if the lens was already prefocused close to the correct distance. Both cameras couldn’t lock onto medium and low contrast subjects at this light level.

An interesting twist: Live view contrast detect AF in the IV at ‘candle light’ level worked great! Medium to low contrast subjects the IV’s phase detect AF couldn’t lock on were not a problem for live view AF.

BTW, ISO 12,800 1/30 f/1.4 is an EV of -1, which is right at what Canon states is the IV’s threshold for AF operation.

– Other than AI Servo performance, everyone seems to be very fixated on how well the IV works at high ISOs, especially in comparison to the D3s. For some good side-by-side tests in real life situations, there’s a very nice, well thought out review here by Unique Photo with lots of files to download and compare. Based on this and other comparisons on the web, I already knew that the IV won’t match the D3s out of the box at ISOs higher than around 3200, at least not with a fair bit more noise reduction work. But that’s OK with me, as I agree with Canon’s approach of increasing sensor resolution considerably over the III while still remaining fairly hands-off in regards to ‘under the hood’ noise reduction. I appreciate being able to have more control over the fine balance between image detail and noise suppression, though I do realize it will likely mean more work for some high ISO images.

– There has been debate on forums whether one even ‘needs’ ISO 12800, etc… and to a degree I can see the point. While experimenting with the camera, there were situations where my guess at the correct indoor exposure was often off by 2-3 stops, on the hot side, meaning the images would have been well over exposed. But we’ll see… I know for sure that some venues, such as shooting night football at Ivor Wynne Stadium will definitely benefit from being able to work at ISO 6400. As it stands that place is now ISO 3200 1/400 @ f/2.8 and gaining another stop will finally allow freezing game action and should improve the keeper rate significantly just by gaining a decent shutter speed.

– Is ISO 12800 usable? For me I feel it is, but then I’m not put off by relatively high luminance noise like some others are. It’s also important to keep in mind that at the native 16MP resolution, one is in effect dealing with an 11×16″ print. For most ‘normal’ uses, the file will be greatly downsized and therefore should be even more usable in light of the very high ISO.

– Button response. In the IV white paper Canon states that the stroke and feel of some of the buttons has been improved. I can feel this most with the AF-ON, exposure lock (*), and AF point selection buttons. On my IIIs, those buttons feel quite spongy whereas on the IV there is a more distinct point of resistance that gives way when the button is pressed (if that makes sense). I think it will mean better feedback when using gloves. The AF-ON button, particularly the one when in landscape (horizontal), has never been quite right for me. It’s better on the IV, though I still find it a touch too far to reach and prefer using the * button instead (might also be a habit from 15+ years using previous 1 series cameras). The shutter release doesn’t feel different from the III. I still find it very easy to fire two frames with the camera set to continuous high, especially with the vertical shutter release.

Update 2010.02.14: The AE lock (*) button on my IV has become quite soft. I prefer to use it instead of the AF-ON button to activate AF because that’s what I’ve been used to using since the dawn of the EOS-1 series. I guess old habits die hard. Anyway, after pressing it for a while to activate AF (such as during a hockey game), it will stay stuck down, sometimes enough to continue to drive AF. It’s definitely a problem, but Canon Canada says they don’t have IV parts yet and can’t fix it for the time being. In the meantime I’m trying to get used to the AF-ON button.

Update 2010.03.02: Got an email from a reader of this review who reports the same mushiness with his Mark IV’s AE lock (*) button after only 1200 actuations.

– Sound (camera’s noise level). Offhand the III and IV sound the same, but they’re not. The IV has a slightly sharper and unfortunately slightly louder sound. It also has a slight metallic echo, like as if someone is hitting a hammer against a steel pipe. It’s not dramatic, but it kind of sounds a bit cheap compared to the III which does not exhibit the echo. This is a bit disappointing, only in the sense that it slightly cheapens the impression of the otherwise solid camera. But, if that’s what it takes to produce a more reliable AF and mirrorbox system, so be it… More disappointing though is that the ‘silent’ shooting setting isn’t quieter than the III, nor is it when in live view.

– AWB: Auto White Balance, as reported by others, seems to handle artificial lighting better than previous 1 series cameras. It’s definitely better under tungsten lighting at getting a somewhat more neutral rendition.

– Buffer: While it’s a bit down the list, it’s actually one of my top concerns about the IV (as stated above, these points are in no specific order). Throughout the 1D series history, if one wanted to shoot sports action in RAW, the buffer limit was always a problem. I often wondered why it was possible for Nikon to offer $500 buffer upgrades (effectively doubling internal memory) but not for Canon. With every new 1 series camera, resolution and/or bit depth increased (except between the II and IIN), effectively canceling out the corresponding buffer capacity increases, generally capping RAW capacity in the range of 25-30 shots at lower ISOs and the low 20s at higher ISOs. 25-30 sounds like a lot, and it can be, but imagine an extended action sequence where the running back starts out at the 30 yard line, breaks through a number of tackles, had a nice looking run, scores a TD, then celebrates with teammates. At 10 fps one has effectively 2.5-3 seconds, yet the play may take 5-10 seconds, plus the aftermath, etc… So one has to be somewhat picky about what to shoot. Take a chance and get the nice action at the beginning of the play, or hold out somewhat hoping that it might develop into something better? Anyway, too often I found myself at the buffer wall. With the IV I was able to get off about 25 full rez RAW only frames at ISO 800 using the Sandisk Extreme IV CF card, which is OK. And buffer write time was at least about as fast as the III, taking about 27 seconds to clear completely (of course one can continue shooting at any point as space clears). I suppose for this we have UDMA compatibility (again, something that should have been on the III) to thank for pushing through the much larger 16MP RAW files. What I’m most curious about though, is how the newest 600x UDMA CF cards will perform in the camera. One online claim was that it deepened the buffer to over 30 frames… time will tell but I’m hoping this holds true.
Update 2010.01.31: I’d like to try the new 600x CF cards, but the fastest cards I currently have, the Sandisk Extreme IV, seemed to clear RAW files quite well and the impression is that it’s somewhat faster than with the III.
Update 2010.02.14: Based on feedback at one of the forums, it appears the fastest UDMA mode 6 CF cards will clear the IV’s buffer in about 7.5 seconds. It appears that these fastest cards can stretch the buffer a few frames if shooting at 10 fps. One tester noted that dialing back to 7 fps gave him 44 RAW frames before hitting the buffer. In any case, 7.5 seconds is great compared to the around 23-25 seconds the III took to clear the buffer with the Sandisk Extreme IV cards, which were the fastest in the camera back in 2007.

Update 2010.01.31: Battery life: It appears that RAW files eat through battery capacity much faster than JPEGs. Shot around 600 RAW outside in clear and sunny -15˚C conditions, with some chimping, and was down to around 60%. Shot another 1300 of the skiers, but in JPEG-L, and battery life remained above 50%. Once the camera had warmed up somewhat the battery jumped back up to around 62%. The difference might also be that of time. The 600 RAW were shot over the span of about three hours with average in-camera reviewing, while the 1300 JPEGs were done in about 20-30 minutes with very little in-camera reviewing. Did a second session outside in the evening just around sunset, about 175 frames and a lot of live view, and the battery was just under 50%. It will be interesting to see how or if this differs in warmer conditions, but so far the IV does appear to be draining the battery more rapidly. BTW, I was using two IIIs along with the IV during the three hour session, putting about 350 frames on each camera and their battery levels dropped to about 90%. The IV was used exclusively with the 600 f/4 IS (mostly with IS off), which might also account for additional battery drain. However, in spring/summer/fall conditions I can usually get 2000-3000 frames on the III and 600 before hitting 50%.

Update 2010.02.14: Video. I’ve shot a few short segments for fun, but nothing serious. Yeah, it works. I can see it being useful, though at the moment it’s not one of my main priorities with this camera.

– Serial Number: This is trivial, but it appears Canon has revamped their 1 series serial numbering system. Previously it was a six digit number and the first digit appeared to represent the camera model. For example my original 1D is number 0012xx. My long since sold Mark II was 2088xx. My Mark IINs start with 4. My IIIs start with 5… but the IV is 02201014xx. I’ve collected serial numbers from a few online reviews that had RAW file downloads available and found: 0220100567 0230100028 0230101219 0230101868 0029900178. The last one though was from a pre-release camera. There appears to possibly be a batch or group number (the first four or five digits), then a sequence number. Like past 1 series cameras, the serial number is not an indicator of when the camera was made. To find that information look inside the battery compartment, at the end near the battery connection will be an ink/paint stamp. Mine reads OX1201. O is the factory, X is the date (2009), 12 is the month and 01 is likely a batch number. In other words, mine is probably from one of the first, if not the first groups. More info about Canon’s dating scheme here.

Update 2010.04.05: Memory card performance

I haven’t yet invested in UDMA Mode 6 capable cards, though am considering the Transcend 16GB 600x as an economical option.

In the meantime I picked up an ADATA 16GB 533x card from a local computer shop in hopes of quick performance. Unfortunately by whatever method ADATA calculated 533x it wasn’t in a Mark IV. RAW buffer clear time is in the range of 35 seconds and around 15MB/s. That’s extremely slow for a UDMA card/camera combo (assuming the card is UDMA, it might only be Mode 4, though the speed rating implies at least Mode 5). For comparison, older Sandisk Extreme IV 8GB cards I have on hand clear the buffer in around 22 seconds at 22MB/s. A Sandisk Extreme III “30MB/s” 16GB card clears the buffer in around 15 seconds and 33MB/s. The ADATA was pretty cheap, only $65 CAD, so I thought I would risk it. Unfortunately this time it didn’t pay off. I’ll probably keep it as a fallback card when the others are full, or for remote cameras, etc. but it certainly hasn’t lived up to its 80MB/s read 50MB/s write speed rating. BTW, transfer to computer via a Sandisk FW card reader was a leisurely 20MB/s to my MacBook, so not sure if there was a problem, but it transfers at between 50-60MB/s with the Sandisk FW card reader via the FW800 port on my G5 tower.

Some photos:

The building photos at the end of my TS-E 17mm f/4L lens review were all done with the IV.

Update July 19, 2010

I finally got around to some outdoor sports with the Mark IV… CFL football to be specific.

The above image is a prime example of a benefit of the Mark IV’s 16MP sensor, though difficult to tell at 600 pixels. It was a punt return shot from the opposite end zone with the 600mm f/4 lens. The extra resolution of the IV allowed for greater cropping. Previously I probably would have shot this with the 1.4x, but tried it without for this game and think it worked well. With the 1.4x there would have been a bit of softness added, and one stop loss, so I would have bumped up the ISO to 800, though that’s not an issue with the IV.

Most of the photos from this beautifully lit evening game were done with the IV, though any of the near action, shorter telephoto and wide photos were on Mark IIIs.

But being a late evening/night game, I wasn’t really that concerned with the performance of the IV because the III has traditionally also worked fairly well at that time of day. The real problem with the III has typically been ‘high noon’ games such as this one a week earlier. Part of the problem is the heat that radiates off the fake turf. You just can’t shoot really long and expect to get sharp photos.

Since I’m stubborn in this regard, and like to shoot CFL with a 600mm lens, I refused to relent and tried it again. To my surprise the IV handled the heat better than the III ever did. I had a fair number of acceptably sharp images, though there was still quality loss due to heat waves and some instances where the AF hiccuped during sequences, but not as bad as the III. I did modify my technique by standing instead of kneeling. The extra three feet of height seemed to help, since the worst heat distortion was nearest to the ground. I wouldn’t say that the IV was perfect. I still had sequences with random soft frames and after the fact I noticed that my 400 needed some micro focus adjustment due to a slight back focus tendency, but I was also playing around with the custom function options.

For the evening game, with the beautiful late evening sun setting over the west end zone lighting up the whole field, the IV was very, very consistent with the 600. I have long sequences of plays where every frame that mattered was properly focused. And, the IV worked quite well in backlighting, both at the ‘high noon’ and evening games, though there were instances when it was reluctant to lock with an off-center AF point on an area of lower contrast, which was understandable, considering that at f/4, only the center sensor is cross sensitive. All the other sensors revert to single direction sensitivity which will affect low contrast acquisition (i.e. backlit) performance. My solution has and likely will continue to be to shoot as much with the center sensor as possible if there is concern about subject detail/contrast and AF acquisition.

Regarding custom function settings pertaining to AF performance, after some fiddling at the ‘high noon’ football game, I settled on the following:

C.Fn III: 2 set to slowest setting
C.Fn III: 3-0
C.Fn III: 4-1 (this is a somewhat contentious option)
C.Fn III: 5-0
C.Fn III: 8-1 left and right points

Settings 2, 4 and 8 are somewhat dependent on each other, and may not take full effect unless all or most are active. 4 and 8 are definitely related, as 4 won’t come into play, even if it’s set, unless 8 is also set to an option other than 0. It’s probable that I got burned with C.Fn III: 2 at slowest at least once where I accidentally picked up the background instead of the ball carrier, and lost a number of frames because the camera wouldn’t reacquire focus. The only way around would be to release the AF button and start again, but for action that lasts half a second or so, my reflexes simply aren’t fast enough. On the whole, I think that for football, the slowest setting works well due to the extremely high probability of players crossing through the field of view in front of the intended subject. Combined with expanded AF points and the ability for the camera to shift primary AF to one of the expansion points to continue tracking the subject (setting 4-1), there were a number of sequences where the camera held correct focus with considerable interference spanning multiple frames. You’ll have to take my word for it, since all of those shots with a blocked subject are otherwise rejects and not in the final edit posted online.

Where I noticed a potential issue was once the game was under full stadium lights, the IV coupled with the 400 2.8, would at times feel sluggish. This was possibly due to C.Fn III: 3 AI Servo 1st/2nd image priority set to the default of AF/tracking priority rather than 3-3 release/tracking priority, which was the default for 1 series cameras prior to the III. In general, I’ve found C. Fn III: 3 at 0 to work well, so it might just be that the IV is a bit slower in dark action conditions (as has been reported). Keep in mind that at Ivor Wynne Stadium, the night exposure is in the range of ISO 3200 1/400 f/2.8, or about two stops darker than most NFL or good college stadiums. While that hesitation did occur at times, on the whole the IV still worked well and produced probably the best results I’ve achieved yet at night at Ivor Wynne.

Both of the 2010 games listed in my CFL photo archive were primarily photographed with the Mark IV>

A short pitch:

If you’ve found the information on this page helpful and are considering purchasing this camera, or any other photographic equipment, clicking through with the following links to purchase at B&H Photo Video would be greatly appreciated!

Canon EOS-1D Mark IV at B&H

Or search B&H directly from this link:


B&H Photo - Video - Pro Audio








Links:

A great NFL game action review by Brad Mangin

An excellent 'real world' sports action comparison between the Mark IV and the Nikon D3s by Unique Photo. Many options here to download photos for evaluation.

Sports Illustrated photographer Peter Reid Miller gives his glowing approval of the Mark IV.

Rob Galbraith has once again started a firestorm on the photography forums with his relatively unkind review of the IV. He's the guy who essentially broke the story about the AF issues surrounding the III, and now it appears he finds that the IV still isn't out of the woods. I agree with some of his findings, though on the whole have been pretty happy with the IV's AF performance. I wouldn't call it a perfect camera, but I don't think such a camera exists. Here's his review. The thing is, the IV is still a good camera, as was the III. Yes, there were times that the III was frustrating to use, but conversely, it was frequently the best 1 series camera experience I had (until now), and I've owned all of the non-s 1D cameras.


Copyright © 2017 TechTalk with Ron Scheffler. All Rights Reserved.
No computers were harmed in the 0.186 seconds it took to produce this page.

Designed/Developed by Lloyd Armbrust & hot, fresh, coffee.