Canon 5D Mark II - Auto focus

Discussion in 'Canon EOS' started by sreegraphy, Sep 24, 2009.

  1. I bought a new 5D mark II and 16-35 L II lens. I was doing test shooting inside my room with my new toys using wide angle lens and with 24-70L which I already own. Both has the same aperture of f2.8 and AF was working fine with 16-35 L-II in available light but when I tested using 24-70 L AF was searching on. I am not surprised about 5D Mark II is not equipped for low light coz I got ample info abt 5D M-II's Auto Focus with the earlier post. Still I made my mind to go for it just bcoz it's a affordable full frame camera by Canon.
    I am just trying to understand the concept. I appreciate all, for your comments in advance Thanx..
  2. Sreehari, the 5D Mark II is one of the two or three finest SLR's currently available for low light photography. Anyone who suggets otherwise either doesn't know what s/he is talking about, or is lying.
  3. The 5D2 has a wonderful sensor, but average focus and metering systems. It's below the 1D3 series bodies which got completely new focus and metering systemss when they went to the 3 series. The 5D2 still has the focus and metering systems from the 3 year old 5D1 body. I think the new 7D is getting updated auto-focus. It's simply Canon marketing. Canon doesn't feel the need to put all their pro level features in the 5D2. For example the 5D2 auto-focus has 1 cross point and 6 assist points (which are constant in the center area of the frame) and only has auto-focus down to f5.6. The 1D3 bodies have 19 cross point sensors and 26 assist points which follow the selected focus sensor. They also auto-focus to f8. I had a 5D1 and a 20D and got rid of both for the 1D3 and the 1DS3 bodies. The 5D and 20D gave good results in "normal" situations but the ability to "get" any shot in difficult situations with the 1D series is amazing. Enjoy your 5D2 as it's a great camera but you need to use it within its designed limits.
  4. Is it that bad? worse then a 30D or 40D? I am happy with how my 5D2 focuses. I was just taking some shots with my 5D2 with the much hated 28 1.8 ( I seem to be the only person that likes this lens ) I shot it a 2.2 and the photos look very good. Focus was in an instant and right on. I don't get it, maybe I have never tried a better camera so this all seems normal but its focuses just fine for me and when it don't its most likely my fault. I wonder how much of this is just people reading reviews and then just spreading it around.
  5. Steve, after reading your comments I checked the lighting conditions at the point I was focusing with a light meter the Exposure value was f4.0 and AF getting locked perfect. I used 24-70L FYI.... set to f2.8 I am just trying to learn the technical aspects to see where it can fit in to capture quality results but besides I am least concerned abt the brand and model I use. I'd even like to take the pain of getting use with MF as long as I get a clean image result at the end of the day.
  6. I use the 5DII for travel, landscapes and street photography, much of which is at twilight, night and in dark churches, bars and casinos. Sure I try to focus on points of contrast when possible, but it is extremely rare to miss focus under any conditions I can hand hold the camera, including F2.8@ ISO 1600 or 3200.
  7. "ocus and metering systems from the 3 year old 5D1 body" If it ain't broke, don't fix it. My 5D1 rarely (very rarely) missed giving correct focus or proper exposure. In fact I would say its metering is among the best available when it comes to accuracy. And while the AF is not the fastest available, it is certainly fast enough - I never have any issues even when shooting sports.
  8. Sreehari - For consistently best auro-focus results with you 5D2 try to use the center focus spot. This has the 6 assist points to aid focus. The 6 assist points are hidden from view and are located within the center circle in the viewfinder. Canon pros realize the outer focus spots are the poorest under difficult shooting conditions as there are no assist points and the focus spots aren't the high quality cross type. There is an active thread on this very topic (see below) on the Wedding site. Again, as I previously stated, the 5D1 and 5D2 are fine cameras but need to be used within their design limits. I used one for a couple years with great results, and learned to "work around" any issues. My 1D's simply allow me to work faster and "get" more shots as auto-focus locks on much faster (with virtually no hunting) than the 5D. The 5D will focus under difficult conditions but you need to compare side by side to realize the difference in AF.
    From Wedding forum (Party shots in dark rooms thread):
    "On another topic raised in this thread: I continued, last night - (at an award function, by coincidence) - to play with the outer AF points in my 5D, (in low light). For most of the night I had a 35L on the 5D - basically I found the outer AF points useless - I found a similar experience on a 30D which had the 16-35L."
  9. Sreehari - Here is the info I referred to in my prior post. From Wedding forum. Party shots in dark rooms thread.
    "On another topic raised in this thread: I continued, last night - (at an award function, by coincidence) - to play with the outer AF points in my 5D, (in low light). For most of the night I had a 35L on the 5D - basically I found the outer AF points useless - I found a similar experience on a 30D which had the 16-35L"
  10. I am surprised that you do not think the 5DII focuses in low light. Mine will focus in low light with either of the two lenses that you mention. Indeed it will focus down to the same light levels as my pair of EOS 1Vs with the full professional AF system. The issue with the 5DII AF is not it's ability to achieve focus in low light but the time it takes to do it. In essence the 5DII AF is slower than the 1 series but works at about the same light levels. The canon AF spec is that the 1DIII works from -1 Ev to 18EV while the 5DII is -.5EV to 18 EV - according to Cannon the 1DIII works in light 0.5EV less than the 5DII.
  11. I use mine for shooting night football games. No, it's not as good as a 1-series, but I get a surprising number of good shots.
  12. My 40d focuses in low light much better than my 5D2.. MUCH better. The 5d reminds me more of my previous Rebel XT. I still love it though
  13. I think the OP is refering to low AF peformance....not ISO performance. When someone downs the 5D2 for poor low light capablity, it doesnt always mean its high ISO performance. Its well known its pretty good. Its AF that causes headache. Yes it works for some, but many people have had complaints about it being slow in low light and inaccurate....just as the OP did....just as Shawn Davie stated.
    I shot a high school football game last Friday night and there was a guy there from the opposing team next to me shooting with 2 cameras. One was his new 5D2 and 3002.8L and 40D with 70-200 f2.8. He was cussing the whole time. Almost every shot he made with the 5D2 was OOF. Sure, he got a couple here and there that was good, but most were OOF. Which led me to believe that the good shots were accidental. The static shots were mostly good, but he said anytime movement was involved it couldnt keep up. He was very upset because he bought it for the low light performance(ISO) for these games. Anyone who has every done them know that ISO 6400 is almost always a given. He figured the AF was at least as good as the 40D was using. But its not. Its fine for weddings and such, but low light levels are its problematic areas. He is just one of the few that complain about it. No its not a sports camera...and no Canon doesnt market as such, but at least they could have given it what the 40D the very least.
    I have to wonder about the posts where the the 5d2 users always say they never have an issue or never miss shots because of its AF. For one, I find it hard to believe that if someone shot a wedding and clicked 500-600 images...not 1 would be OOF due to the camera. I find that very hard to believe. Which also leads me to believe that either they dont know the shot was slightly oof or dont care. No camera will deliver 100% so I know thats its BS when someone says they never have missed shots with "their" 5D2.
    I think a lot of wedding photogs that use it usually have ideal settings to use it in. They set up additional lights and all this helps give something for the AF to catch, so yes it performs quite well. But, take those lights away and use it in real world lighting and where lots of these owners use theirs, and it misses quite a bit, and just about anyone will complain about it being slow. Owners who use it in available light indoors where its low and lacks lots of contrast just as the OP posted and these are the statements we get. But I guess everyone will chaulk it up to "user error" I'm sure. Shawn stated his 40D was much better than his 5D2. So I guess he does know how to use his 40D, but doesnt know how to use his 5D2:) No body says the 5D2 is just terrible. I like its IQ and the assets it provides, but if someone complains about its AF, the post loads up with users defending it like a disabled child being picked on. The Af was the major downfall to an otherwise very good camera. Add the 40/50D Af to it and its becomes a much better option. Just my observation.
    To the OP, I'm not real sure why the camera would do fine with with the 16-35 and not the 24-70 being they have compairable AF and same aperture. It is strange. But, I'll actually give you you the benifit of the doubt in that you do know how to use it, so I'll assume its NOT user error, but more so an inconsistent test. My question would be, did you use it at the same focal length. Which is a very hard thing to do because the 24-70 could only match at the 24-35 focal range. So did you change your position to match the FOV of the 16-35.
    I say this because with the 16-35, the camer sees more broad FOV and could maybe allow the AF to see more contrast. Example: Sitting in low light like ISO 3200 f2.8 and 1/80. I see this alot indoors. Lets say the 16-35 was used at the 35mm end. And this gives you a full body shot of someone in the room and leaves enough area on top and bottom of image to crop an 8x10 easily. The AF points size relative to the subjects in the viewfinder is large. It would cover a large portion of the subject, so it sees more of the subjects contrast. Now, without moving, using the same subject but now setting of 50mm on the 24-70, the AF point is smaller in size relative to the subject. It may actually be half the size of the face for instance where before, it covered the whole head or even more. But the face lacks the contrast it needs. Before, it found contrast in the hair, eyes, lips, ears etc. Now at longer Focal length, th AF points size relative to the subject is smaller and only sees a cheak. So unless the subject has a beard, then the contrast is quite low. Just be sure you compare at same FL with these lenses, it would really bias your tests....they are very different lenses. That said, the 24-70 can be quite random in AF calibration consistency. Mine was sent in twice and still backfocused even at f4. I used the AF Micro adjustment and Lesn Align Pro recently and set -11 to center it up. Not its pretty good. But I'm sure in 6 months, I'll need to re evaluate it. This has beein the MO with this lens since it was new. Maybe its my camera, I use a 1D3 and its known for AF issues. Good luck.
  14. I thank every one for sharing their views
    David, I don't really think I matched 24-70 with 16-35 in terms of FOV But I was trying in all focal length. Like you said it could be inconsistent coz with the same ambient light and setting it focus some time when I tested later.
    Whatever technical set backs 5DII has got is offset with it's Image quality. IT'S SIMPLY INCREDIBLE". More realistic....
    Rather than crunching abt what it has not got... It's better to explore what it has got and keep going.. No doubt it's a wonderful piece of machine. I was in high dilemma to go for 5DII. Before, I was even eying for 1Ds III though I was not financially equipped to own one. I prefer myself to excel with limited resources. But I do have more than limited. There are pro's who has top profile in P.N and doesn't even own more than Rebel XT.

    Once again Thanx to All for sharing your comments.
  15. David
    I will try and respond to your points on the 5DII. I have not shot much with the 1D bodies (I have use the 1DIIN and 1DIII) but have shot a lot with 1Vs (I still own a pair) and earlier Af bodies such as the 1NRS. My impression is that the 1D bodies are slightly better in low light than the 1V. The 5DII is not in the same league as any of the 1 series bodies for AF but is actually not bad. It is very critical of how it is set up - much more so than the 1 series bodies and I think a lot of issues stem from that. I shoot quite a lot of indoor ice hockey and ski racing - both of which need a good AF system. Since I do not do this for a living anymore I cannot justify a 1DIII (I may check out the 7D if it's AF is as good as claimed). Here are my conclusions from using the 5DII for sports with the 300 F2.8 or 70-200 f2.8 lenses.
    1 You must use an F2.8 lens or faster on the 5DII to get decent AF - adding a 1.4x TC will results in big problems with missed focus
    2 For sports it is best to track and use AI Servo - I find that having just the center point and invisible AF assist points is the best approach - using more AF points slow the AF system down and you miss a lot more shots.
    3 One Shot AF on the 5DII is generally too slow for sports - especially in low light.
    If you follow these rules I get a very good hit rate on ski racing (Helped by the fact that I still do it so I am very familiar with the athlete motion) probably losing only 10 or so photos out of 300 -600 that I shoot in a race. With kids hockey it is a lower sucess rate and is sensitive to the light in the arena (I have been in arenas at LV8!). In a reasonable arena i will get about 90% in focus. If I deviate from the above rules then the hit rate is terrible. The first time i shot kids ice hockey with the 5DII I hardly got any shots in focus as I was switching between one shot (only center active) and AI servo but with the whole Af pattern active.
    One other issue I have noticed with the 5DII is that the invisible AF assist points should be switched off for critical use - e.g. portraits as they can actually move the point of focus without you being aware - it took me a while to figure this out.
    As I said above the difference in absolute AF performance (i.e. can I get focus) between the 5DII and the 1V is negligable. I just played about with a white (off white) wall and the dimmer on the lights and I was unable to find a light setting where the 1V would achieve focus lock and the 5DII would not. This is not to say the AF systems are the same (the 1V was always faster to focus) merely to say that I was unable to set a light level where the 1V would focus and the 5DII would not. As I stated ealier Canon specs the 5DII AF at -.5 to 18 EV while the 1V is 0-18EV. The test would seem to vindicate Canon that the 5DII does indeed achieve focus in very low light. The 1DIII has a 0.5 EV better working range (-1 to 18EV) but it is unlikely that it will achieve focus in a much darker room that the 5DII does. It is possible that Sreehari has unrealistic expectations of the light level in which a camera can achieve focus.
    I am interested that your freind found the 40D to be much better - on paper the 40D and 5DII have essentially the same AF systems. Perhaps it is just that the 5DII has less procvessing capacity for the AF system. I suggest that he tries my approach and sees if this helps. If not he may have a camera issue. Ice hockey is typically ISO 3200 1/400 to 1/800 and F2.8 (i.e. about LV 8 ) from memory of when I shot outdoor sports at night this is about the same level - I seem to recall shooting in LV 7-8. Of course with hockey you tend to have lots of contrast so this may be why I have been more impressed that i expected with the 5DII AF.
  16. As stated above, he was using a 300 f2.8 and no converter. He was using only center point focus and he was using AI Servo for tracking.
    The 40D has a better AF system. The ceneter point is more sensitive and is a diagonal cross type, where the 5D2 is not. He could switch to the 40D with 70-200 f2.8L and start getting a lot more keepers right off the bat.
    And the 1D3 has a much more sensitive Af system than any of the camera mentions. I have found many situations where my 1D3 would lock focus without hesitaion and the 40d will just hunt. So this also lends belief that the 5D2 is even further behind.
    Not trying to argue the point of the 5D2 AF being the worst of the Canon prosumer bodies. Everyone knows this already.
    The OP clearly stated his issue. The 16-35 would lock, the 24-70 would not in many cases. He also responded stating he didnt match focal lengths like I asked and that he probably messed up the results by not testing both lenses the same way.

Share This Page