Field test: Shooting Roller Derby with the Olympus OM-D E-M5 II

Oly (and fuji) inflating ISO values is a fairly well known and consistently observed by many.

Ah, yes, Maybe that's it.

For any sports photography, you just need sharp, well isolated from background, clean image of something interesting. Another option is pan the subject to make the image more dynamic.

Those pics are technically not satisfying. Compositions are confusing (not a full close up nor a full shot), not quite dynamic, and there are tons of distracting backgrounds around.

And most of those issues came from equipments - lack of reach, lack of background blur, lack of fast fps.

If you'd like to give me 10-15 grand or so, I can put together a rig that would satisfy your requirements, nerd2. Something like a D4 and a 200f2 would be the place to start.

Armchair criticism is easy. Roller Derby is easy to access. Go shoot some and report back.

A few people here seem to be under the impression that the goal behind shooting derby is to grab a handful of shots for your blog. That is incorrect. The photos are usually for the players so they can see the bout from a more relaxed viewpoint (as opposed to being in the thick of it). Finding some nice artsy shots is a bonus, but that's the sort of thing you play around with when there are enough other photogs providing decent coverage of the bout.

And, as I mentioned before, if your plan is to use DoF to isolate players, then, IMO, you're doing it wrong. It's not the best tool for the job. A much better approach is to isolate with lighting.

Precious few of the greatest sports photographs fit your list of requirements.

Another fun video @Barney, but yeah expect tough crowds when you do mirror-less for sports. But let's get you to a major American Sporting event @Barney - eg college or pro - that pays for the shot(s). Then mirror-less will go to the wayside and 7d2+ canon glass will get gold at dpreview for the $ invested in gear as it generates many more returns, and is a sports artist's brush, that your site thought was a general camera to be rated by a general checklist.

Problem is that those pics were taken with $4K amount of gear, not a $500 entry DSLR with kit lenses.

For that kind of event and $4K budget, I'd rather use D700 (now dirt cheap, still have great high ISO capability, 8fps, pro ergonomics) paired with f4 lenses (1 stop faster than oly lenses in equivalence)

And you'll end up with similarly noisy photos with such an ancient FF camera, not helped with the F4 limitations of your zooms for exposure.

With 12MP, you will also be limited in reach not be able to crop as much. If you try to compensate with longer lenses, be prepare to pay with money and extra weight to carry. No crop factor to help you here.

But you still get your extra stop of DoF control. Woo-hoo! Take that, m43!

Fortunately, it's 2015 and not 2008 so there are more choices around.

"A few people here seem to be under the impression that the goal behind shooting derby is to grab a handful of shots for your blog. That is incorrect. The photos are usually for the players so they can see the bout from a more relaxed viewpoint "

The goal behind every shooting could be whatever. The goal behind this field report was to show readers that 4000$ mirrorless gear is capable of shooting roller derby. Was the goal succesfully filled? I don't think so. Images are really bad, both from technical and aestethical aspect. And it's definitely not just camera's fault, but also photographer's.

And even if all his f4 zooms have optical stabilization, it's less effective than the EM5 II IBIS so nerd2 will have to shoot at higher ISOs in such a dark environment to compensate. I don't think the results would be particularly pretty.

He could always shoot with a tripod but then he'd be conceding m43 is better hand-held. Poor nerd2. So close yet so far.

bluevellet: 5axis stabilization for shooting roller derby like in the video? Tripod? Please, do not post about things you know nothing about.

I wasn't talking to you actually.

And yes, I knew what I was talking about.

Stick to your initial criticism about the video vs the photos otherwise.

D3/d700 still have vastly better high iso performance than any m43 to date. And Oh yeah, 5 axis ibis is crucial for action photos with 1/500sec shutter speeds or faster.

After all its your money, and its up to you which gear to invest.

The Nikkor 150 f2 is a spectacular lens indeed, but "starting there" would mean breaking the 4k limit just with this lens. I think that was the point of bringing up this lens. :p

Well, what would you use to isolate using DoF? The 70-200 doesn't cut it (I shoot alongside FF users using that lens. We all have problems with cluttered backgrounds - not surprising as there's less than a stop in DoF, between the 70-200f2.8 and the 75mm f1.8).
PS. I don't know any Nikon 150f2. I was thinking the 200f2.

My father uses em1 with 2.8 lenses, and theres no way m43 body has only 1 stop disadvantage from d700. I even think the ancient 5d is still a much better performer at iso 3200 than em1.

Also portrait grip is a must for any serious shooting imo, which oly charges as much as nikon too.

So you don't believe DXOmark?

Oh, but look at your other comment. Looks like you do believe DXOmark for ISO (You just don't know what it means).

What other fallacies will we see from you today?

@Looleylaylow
Huh? You actually dislike DPR staff interacting with the readership in the comments?

Seems like you're baiting me, @ Looleylaylow.

You've never needed baiting, Barney. And I'm not trying to draw you into an argument. You do good work, and I take it seriously. It's harder to do that when you undermine it with smug comments in these forums, that's all.

that's like saying people are foolish for paying $1000 on a 12-40 pro zoom when they can get a $250 kit zoom... disregarding that the pro zoom is heads and tails better than a plastic janky kit zoom.

but why stop there when you can get a point and shoot with a 10x zoom clearly written on the barrel..

next time you might want yo try comparing like for like.

Huh?

28-300 is reasonably sharp (to resolve 16MP at least), and all of those pics are taken at high ISO where noise kills most of the sharpness anyway. Just download originals and see. None of those pics are pixel sharp.

Of course more sensible choice for indoor sports will be f4 combo (24-120VR + 70-200VR) or even third party f2.8 lenses but then the results will be totally in different league and not really comparable. What I am saying is this kind of shooting situation is the worst case scenario for m43, and is really not any better than pairing f5.6 superzoom to D750

As far as testing lenses like the upcoming 300mm, I think it would have higher rate of keepers because a lens like that has to be shot off of a high-level tripod with a Wimberly gimbal mount, or at the very least a hell of a monopod. That would help focus accuracy somewhat I think because you've eliminated camera movement from the focusing equation and now the camera only has to keep track of the subject.

@SmilerGrogan, 'As far as testing lenses like the upcoming 300mm, I think it would have higher rate of keepers because a lens like that has to be shot off of a high-level tripod with a Wimberly gimbal mount, or at the very least a hell of a monopod.'

But but but... Olympus Head Honcho said that lens is for handholding unlike other equivalent FF lenses. So you must be wrong... :p

I'm certainly going to be using the 300 handheld, and waving it around like a maniac. :D

20-25% AF failure is NEVER normal for pro cameras.

Wow. That sounds like an awesome challenge. The birds in flight photographers are going to be very interested in that lens. If you hustle you still have time to test it out on the birds at Bosque del Apache.

nerd2: I'd like to clarify that by AF failure I'm talking about AF not being spot on accurate.

See this:
'Still, that gives us a "missed focus" rate of 20 percent in this little test, and that's using center point AF with a still target. That's not a big sample size obviously, and we repeated it with several other 50mm f/1.4 lenses and got "missed focus" rates of between 10 and 20 percent for all of them.'
from - http://www.lensrentals.com/blog/2012/07/autofocus-reality-part-1-center-point-single-shot-accuracy

Pro cameras and lenses usually have higher than 90 percent af hit rate with 2.8 lenses. And those oly lenses are f5.6 equivalent, and having huge DOF to boot. They are totally in different league.

@Macx
You can watch the pictures and guess the DoF. I didn't do that.

What I did is:
1) I checked the focal length in EXIF
2) I *assumed* images weren't cropped.
From 1 and 2 you can:
3) Estimate the distance to the subject.
From 1 and 3 you can:
4) Estimate the DoF.

You can easily find all needed equations on wiki OR use some online calculators.

@Barney Britton
I'm not saying it isn't a good CDAF. It is. But most Olympus bodies have great CDAF, so we shouldn't be surprised.

But the % of failed shots is still quite high (IMHO).
As said before: DoF was covering the players, you had a large distance to the background and not many objects that could distract the AF.
In such conditions I would expect better results. If the camera didn't focus on skaters, it must have chosen the background or just surrender.

By contrast, if this was shot with a FF + 85/1.8 DoF would be under 1m. At this point a failed shot could mean locking on a leg or a player behind - you could already miss the face.

E-M5II is clearly a very good camera, but IMO you're getting soft with mirrorless. :) We want them to get better than DSLRs in every way. Because they can.
The "almost there, but at least smaller" approach is worrying. :D

@Barney Britton
And just a small suggestion.
This is another (after - at least - the 7Dm2 and a6000) review, when you shoot with someone using the same camera.
I totally get the idea: comparing, tutoring, discussing the results etc.

But wouldn't it be good to have another person with a different camera on set? Something older and well known - to give us a reference point? :)

is this e-m3-ii c-af with old 4/3 lenses or new m4/3 lenses???

inorogNL, all of the C-AF testing - that Barney and I did using the E-M5 mark 2 for this video, and that I did myself with the E-M1 - was done with native micro-4/3 lenses. (I personally don't own any of the old Four Thirds DSLR lenses.)

I find this very fascinating. It is my understanding that sensor based PDAF does not do as well indoors or as the lighting is reduced. I saw Panasonic explain this is one reason they think their DFD is a smarter technology.
With the E-M5ii also possibly outperforming the PDAF based E-M1 indoors, it adds to my impression that CDAF which can use all the pixels instead of a couple hundred might be better than sensor based PDAF.

@Mr. McDaniel
So you shot H mode set to 5fps? I have found the CDAF+PDAF hybrid to be great in C-AF before 3.0 firmware but now the PDAF only on H+C-AF is awesome.
Have you noticed that the flicker from lights would affect the focusing accuracy?

@Fri13, for both tests (E-M5 mark 2 and E-M1 v3 firmware) I shot in Sequential L mode (not H) with L fps = 5 and Rls Priority C set to Off, 1/500 shutter and ISO 6400 with the f/2.8 zooms and around 3200 for the f/1.8 primes. My understanding is that L (low speed) mode does frame-to-frame AF while H (high speed) mode does not. So I didn't test H mode with it's minimum 5 fps.

It's a question of light. The PDAF of the E-M1 needs lots of light to function well, because the PDAF sensor area is small. Once light levels drop down to about "indoors with a big daylight window" it first becomes jerky (drop of Live View / focus frame-rate) and then automatically switches to CDAF. And the user mostly has no control over its behavior (switching).

@Timur, thank you! I remember now - I *was* getting that stuttering, low EVF frame-rate when testing the E-M1 at the Rat's Nest. I had no idea why it was happening, and no idea that it was due to the PDAF usage switching on or off. I also did not get that stuttering when shooting with the E-M5 mark 2, and now I know why. Do you have a URL where this behavior is discussed in more detail?

@James McDaniel

There is your cause for difference.

The firmware 3.0 brought update to C-AF is only available in H mode with C-AF. Otherwise you use the same old PDAF+CDAF on L + C-AF and that is the limitation.

The Firmware 3.0 change was that now the High burst mode has C-AF capability, so no more limited to Low burst mode to get continues autofocus, and then the framerate to burst mode was raised from 6.5 to 9 with C-AF.

So try some time by setting E-M1 to H burst mode, FPS to 10 and C-AF. The camera will automatically drop to 9 FPS when using C-AF and when using S-AF it use 10 FPS.

And in H + C-AF mode you get the PDAF only being used. And there is the difference to previous firmwares and to L burst mode. Sure you can use 5fps as well in H mode but seems to be a limitation if shutter speed is capable to higher.

This is the thing in firmware that really changed the E-M1 to totally new level as high burst speed continues autofocus is possible with PDAF only.

It wasn't more or less hybrid before v3, I see no difference in behavior there, at least with non sequential modes. On a side note, the E-M1 even uses limited PDAF in AF-S, it's just not advertised (or even denied).

As I wrote, only in H + C-AF the new mode, only PDAF, is used. Otherwise is it exactly the old one.
And PDAF is not used in S-AF, unless you shoot with a 4/3 lenses. The hybrid mode is used only in C-AF L mode.

Nope, PDAF *is* used with MFT lenses. It can and has been reproduced. ;)

As I wrote, PDAF is used with m4/3 lenses, but not in S-AF mode (only CDAF in S-AF) and only in C-AF and C-AF+Trk modes. Now since firmware 3.0 it is used as only focus method in H + C-AF combination. The PDAF is only thing used when used 4/3 lenses, not CDAF at all as Hybrid.
Olympus documentation covers all this.

Olympus documentation does and does not cover a lot of things. Albeit it is irrelevant for this video it still stands, the E-M1 uses PDAF for AF-S with M43 lenses. It is used for initial validation of whether the target distance has changed at all. If it didn't change a bit then you get an instant in-focus and no further (CDAF) focusing procedure is initiated at all (there isn't even a screen contrast change like what happens with CDAF). It doesn't seem to work in all instances, but again, it has been reproduced and is only reproducible within the PDAF area. Search the MFT forum here on DPR.