1D X AF in Low Light action

[bdunbar79]

The 1D X finally met its match Tuesday night.  I filled in for a reporter for the News Journal at a game in Mansfield, in a terribly lit gym.  I used a 135L with a 1D X at 1/500s, f/2, and ISO 3200 to 6400 depending on location on the court.  The majority of my shots were slightly OOF, and I attribute this to the AF system not being able to lock focus in the dim lighting.  Looking at my photos, it didn't apparently lock on anything in the frame.  In a few burst shots, the first one was in focus (while athlete still on the floor) but the camera lost focus as the player went up for the layup, even with face-recognition on.  Needless to say, as I'm going through my RAW files right now, I'm highly disappointed.  I will admit though, that the 135L is not as good at locking on focus as the 70-200 f/2.8L II IS zoom lens, but I couldn't afford the loss of a whole stop of light in this gym.  I know I've been spoiled with a lot of light lately, but I just wanted to point out that as good as the 1D X is, and it is, it is certainly far from perfect.  My suggestion is that in these cases maybe don't use high burst rates, but wait for a moment and try to lock on, then move the camera with the player, sort of like panning, and then fire a shot, maybe even just one shot.  Bursting did NOT work at all here.  Ok, I'm done venting.

[Razor2012]

The 1D X finally met its match Tuesday night.  I filled in for a reporter for the News Journal at a game in Mansfield, in a terribly lit gym.  I used a 135L with a 1D X at 1/500s, f/2, and ISO 3200 to 6400 depending on location on the court.  The majority of my shots were slightly OOF, and I attribute this to the AF system not being able to lock focus in the dim lighting.  Looking at my photos, it didn't apparently lock on anything in the frame.  In a few burst shots, the first one was in focus (while athlete still on the floor) but the camera lost focus as the player went up for the layup, even with face-recognition on.  Needless to say, as I'm going through my RAW files right now, I'm highly disappointed.  I will admit though, that the 135L is not as good at locking on focus as the 70-200 f/2.8L II IS zoom lens, but I couldn't afford the loss of a whole stop of light in this gym.  I know I've been spoiled with a lot of light lately, but I just wanted to point out that as good as the 1D X is, and it is, it is certainly far from perfect.  My suggestion is that in these cases maybe don't use high burst rates, but wait for a moment and try to lock on, then move the camera with the player, sort of like panning, and then fire a shot, maybe even just one shot.  Bursting did NOT work at all here.  Ok, I'm done venting.

You can't beat real world, in the field testing.  Even with a top-of-the-line camera like the 1DX, nothing is perfect.  The good thing is, you could of been shooting with something else. 

[Dylan777]

I'll trade you my 5D III

[nightbreath]

Try to lower your framerate as 12 fps makes a lot of light lost.

[bdunbar79]

I get your point guys.  I was just frustrated because I had some potentially good shots that were OOF.  I guess you have to make do with the situation you are in and to just try to do your best.  The nice thing about this website is that everyone seems to be ambitious and motivated to improve his or her photography. 

[bdunbar79]

Try to lower your framerate as 12 fps makes a lot of light lost.

Thanks.  Yes, that is what I was considering after I left the game .

[JaxPhotoBuff]

The 1D X finally met its match Tuesday night.  I filled in for a reporter for the News Journal at a game in Mansfield, in a terribly lit gym.  I used a 135L with a 1D X at 1/500s, f/2, and ISO 3200 to 6400 depending on location on the court.  The majority of my shots were slightly OOF, and I attribute this to the AF system not being able to lock focus in the dim lighting.  Looking at my photos, it didn't apparently lock on anything in the frame.  In a few burst shots, the first one was in focus (while athlete still on the floor) but the camera lost focus as the player went up for the layup, even with face-recognition on.  Needless to say, as I'm going through my RAW files right now, I'm highly disappointed.  I will admit though, that the 135L is not as good at locking on focus as the 70-200 f/2.8L II IS zoom lens, but I couldn't afford the loss of a whole stop of light in this gym.  I know I've been spoiled with a lot of light lately, but I just wanted to point out that as good as the 1D X is, and it is, it is certainly far from perfect.  My suggestion is that in these cases maybe don't use high burst rates, but wait for a moment and try to lock on, then move the camera with the player, sort of like panning, and then fire a shot, maybe even just one shot.  Bursting did NOT work at all here.  Ok, I'm done venting.

Your story is interesting to me because I have been shooting primarily (personally not professionally) high school basketball for the last several weeks in a variety of gym lighting conditions using the 1DX, and I have not experienced the problem you ran into.  I know from past years that some gyms have truly horrible lighting conditions, worse than what I've been shooting in lately, so I imagine that the symptom you describe may be both setting and lens dependent.  I shoot my 1DX with either the Canon 24-70 f/2.8 L or the 70-200 f/2.8 L IS.  The latter seems to track focus better and allow faster burst rates, but both have reliably achieved focus quickly. I shoot in AI servo, with the high speed burst rate set to 10 fps (not that it always achieves that). I also have the first and second frame focus priority settings set to focus priority (i.e. "don't fire unless focus is locked"). Sometimes that means the camera takes a fraction of a second before firing, if it hasn't locked on, but it has usually meant that the shots are in focus (unless I screwed up and put the focus point on the crowd).  I also set the camera in aperture priority mode, with an aperture of 4 to 5.6 and a minimum shutter speed of 1/500th and the ISO set to Auto. This has resulted in ISO settings of as high as 20000 in some cases, and I usually need to apply 50-60% luminance noise reduction in Lightroom, but the pictures end up looking great. Don't know if any of that helps, but thought I'd offer my recent experience.

[Drizzt321]

I'm assuming you performed AFMA on that lens + body, correct? If not, perhaps that contributed to the OOF. f/2 on the 135 is a pretty shallow DoF.

[bdunbar79]

Thanks for the responses.  They are much appreciated.  To answer some questions, this was perhaps the worst gym ever in which I've had to shoot, so to answer your question, yes it was pretty bad.  I did a game at Ontario with primarily the 135L and didn't have any problems.  If I have to return to the gym, I will likely shoot much less fps if I go with the 135L again.  I agree that the 70-200L focuses much better and faster, but shooting at f/4 to f/5.6 was out of the question and I didn't like the metering with f/2.8 either. 

My question is, if you do shoot at say, ISO 8000 and above, how do you post-process so that when you convert to jpeg it doesn't look grainy and terrible, and/or blurry?  Typically when I print 8 x 10's for the university and had to shoot at 8000 and above, the faces begin to seem blurry, for obvious reasons (noise reduction).  I try to avoid this at all costs and not go above 5000 if I can help it with aperture.  If you do need to, do you just add more clarity in post?

When the 135L does get it right, it gets it right.  This is from the game I was having problems:

[Sporgon]

I thought the 135L had an acknowledge problem with the design - focusing behind. As someone has pointed out DoF at f2 is razor thin and I certainly gave up trying to get 100% focus on a moving subject at that aperture.

I don't think you had a chance.

[bdunbar79]

I thought the 135L had an acknowledge problem with the design - focusing behind. As someone has pointed out DoF at f2 is razor thin and I certainly gave up trying to get 100% focus on a moving subject at that aperture.

I don't think you had a chance.

I don't have this problem in well-lit gyms.  I use it all the time for basketball and volleyball at f/2.  I think given my hit rate in the past, I had a great chance.  If there is a "focusing behind" problem, I was not aware of it, but that would certainly be a problem.  Either way, getting good shots at this high school isn't going to be easy if I have to return again.  The ones from the past in the newspaper I've seen are very dark.

[bdunbar79]

I thought the 135L had an acknowledge problem with the design - focusing behind. As someone has pointed out DoF at f2 is razor thin and I certainly gave up trying to get 100% focus on a moving subject at that aperture.

I don't think you had a chance.

I forgot to add, I think considering this, I will go back to the 70-200 f/2.8L II IS lens as my MAIN lens and on my 2nd body shoot some with the 135L and compare.  If I have a higher hit rate with the zoom, despite ISO and noise, I'll try to work around the other problems.

[Sporgon]

Colleagues who are Canon users in the States have told me
the focus issue with the 135L is one that Canon are aware of. In using it on
By using the 1DX no doubt you gave it the best chance !
But I doubt if the 1DX was to blame.

[bdunbar79]

Colleagues who are Canon users in the States have told me
the focus issue with the 135L is one that Canon are aware of. In using it on
By using the 1DX no doubt you gave it the best chance !
But I doubt if the 1DX was to blame.

Understand.  Thank you!

[helpful]

Yes, as soon as I had the 1D X, I was already wishing for more. Sometimes I think the AF system is fine-tuned for decent lighting, and I have felt that it can miss some photos in worse lighting by being too good at trying to track the object, whereas a simpler camera might actually have gotten the shot in focus, not because it was smarter, but because it was dumber. The super advanced tracking algorithms don't seem to work as well in lower light. Yes, I have tried reducing the frame rate from 12 fps down to 8-10. It doesn't seem to make much of a difference. Sometimes I subjectively think it even hurts to change the frame rate below 12 fps (no scientific way I know of to measure it, though). It may be the math formulas that calculate focus are pre-optimized for the exact mirror bounce, delays, resonance and counter-resonance, and other physical issues that happen while shooting at 12 fps. I'm probably way off, though. I was just expecting much better results at a lot lower frame speed (since the AF sensor would be exposed to the image for a much greater percent of the time), and my results weren't really any better.

Another problem I have with my 1D X is that initial focus point acquisition in AI Servo seems to be imprecise. In 61 point mode, I was hoping to be able to acquire focus even on a subject included in only the center point (like the manual says that one can do) and then have that subject tracked using all 61 points.

Instead, the initial focus shifts to nearer objects if the desired subject is covered by only the center focus point. If the desired subject was already locked on when it was more isolated, then it's fine.

So basically it's almost impossible to shoot someone whose head is above a group of players, for instance, because it will lock on the other players all the time, unless I was tracking that player to start with.

If I use single point mode or four-point expansion, then I don't have problems, except I lose the benefit of the 61 point AF tracking system.

I hope that I am being stupid and overlooking something obvious, but I have tried every single possible AF combination, especially the obvious ones like disabling automatic AF point switching, and even the non-obvious ones like the advanced color tracking system, etc.

My 7D cameras seem to have absolutely no problem with acquiring initial focus on whatever is in the center point, even when they are in 19-point AF mode and lots of other objects are in the surrounding AF points.