Rico just posted his X-T10 AF Tracking Samples on Flickr. The samples are taken wide open (F2.8) with the XF 50-140 zoomed out all the way at 140mm.
Read Rico’s Full First Look at the X-T10. And in order to avoid GAS, please do NOT read & see his XF90mm First Look.
_ _ _
X-T10 (body + kit options): USA: AmazonUS / BHphoto / Adorama EU: AmazonDE / AmazonUK / AmazonFR / WexUK / PCHstore / AUS: Camerapro
XF 90mmF2 WR: USA: AmazonUS / BHphoto / Adorama EUROPE: AmazonDE / AmazonUK / AmazonFR / WexUK / PCHstore / AUS: Camerapro
67 Responses
I’d say it’s quite impressive.
Are these the only successful shots is a series? If so, how many of them are not?
Please Rico, tell us all the story. Perfect-focus rate / failure rate, causes of failure (busy background, user error, non-predictable movement…)
Since there only 3 or 4 shot numbers missing in the shown sequences, there isn’t much of a story to tell, I’m afraid. I always discard shots that aren’t okay at 100% on my Retina.
Good to hear !
I wondered if you used LR numeration or kept the Fuji customized file names.
Nah, all I did was changing the EXIFs to “X-T1” and then back to “X-T10” again to force LR to deal with the RAWs. That’s the beauty of using the same sensor in different models: We have LR support even when we don’t have it.
Having just tried shooting my son running around in the park with the X-T1 and the 50-140mm lens, I have to admit these results are really amazing. I had a tough time getting one shot in focus, let alone multiple shots…
What a hideous bokeh! I can’t stop wondering how such experienced manufacturer like Fuji could make the flagship of the lens lineup with such bad bokeh. Canikons are way better here.
PS. AF tracking is great).
All lenses give hideous bokeh with cluttered backgrounds (such as trees), backgrounds too close to the subject, etc. I doubt whether – all these factors being equal – you’d see the difference between Fuji and Nikon/Canon lenses.
I would have tested it with a 90mm at f/2, but I didn’t have one, so 140mm at f/2.8 was the best choice. Obviously, the goal of the test was to create the worst conditions for the camera: contrasty background, backlighting, maximum aperture, minimum DOF, the model going zigzag and CDAF tracking over the entire frame.
so how do you summarize the performance? is there an improvement over the previous firmware of the XT1 ?
Of course there is. Simply because the zigzag shots aren’t possible with the older firmware.
yeh, like the 3d tracking in nikon DSLRS or the way sony samsung and panasonic are tracking.
i wonder if this functionality is not a byproduct of the future autofocus system of the next generation fuji sensor because i think it works better with more phase detection pixels.
It certainly would work nicely with PDAF, but this requires a larger PDAF sensor area. Of course, the next sensor will feature a larger PDAF area.
The surprising thing for me is how well it works with CDAF.
the DFD (depth from defocus or whatever this means) used by panasonic tracks really good without PDAF pixels at all. are you familiar with it?
so i am not sure they necessarily need to go to the larger PDAF area…
Of course I am familiar with DFD. Great stuff. Fuji doesn’t have it, though.
yet… maybe it takes some more processing power, maybe they will have it for next-gen of processors… it has great potential. i think most manufacturers will adopt it (given patent limitations).
I doubt that Panasonic will grant an affordable license, but I certainly suggested to Fujifilm to try getting the technology.
I agree that – bokeh-wise – the 90mm would probably have given you somewhat better results. Still I think that other factors (background choice, subject-background distance, camera-subject distance, etc.) are more important in that respect.
Of course, and these factors have to be maximized to create a difficult situation for the camera and the photographer. Obviously, I have chosen those contrasty picket fences in the background deliberately, as they tend to attract both PDAF and CDAF. The fact that the camera wasn’t distracted illustrates that the new firmware works as promised. In order pass or fail, one has to put up a real challenge. I also repeated the test with the picket fences in landscape orientation with PDAF (#4), knowing that PDAF loves vertical edges. The backlight scenes two days later also took place at a proven location, because that’s where we do AF practice sessions for my Fuji X Secrets workshops. I simply used an X-T10 at 2 of those practice sessions, which are designed to confront delegates and their cameras with practical challenges.
You did a great job testing the X-T10’s tracking AF. I realize that you were not after optimal bokeh, so let’s talk about AF again… 😉
I never had the impression that most folks in the comment section are interested in stuff that works. Instead, they tend to focus on things that don’t work, or (as a last resort) making up things that don’t work. It’s similar in the forum(s). But keep one thing in mind: only a few people comment, but thousands(!) have already viewed the album in the past hour alone.
Yeah, you’re right. Thanks for the test! It’s extremely helpful.
When something works there is nothing to talk about).
Rico, I think that there is another reason for the negative comments. I think most of them come from those who own cameras by another maker, or are their covert employees, and since the new firmware is so good, are trying to seize on anything negative.
Have you seen any difference between vertical and horizontal orientations in this scene?
Not really. But only PDAF is partial zu line orientation, anyway.
I totally disagree. There are lot’s of lenses that deal with such complicated background even in transition zone. And XF 50-140/2.8 is definitely not one of them. Even non-APD XF 56 is a great performer.
I’ve already made the side by side comparison against 70-200/2.8 VR II and Fuji was way worse. It wasn’t a surprise for me as the neutral nature of this lens’ bokeh became obvious to me from the fist samples. Fuji probably were after maximum sharpness making bokeh neutral, not positive. But IMO this kind of lens should feature great bokeh.
I don’t doubt that some lenses render better bokeh than other ones. My point is that other (controllable) factors are way more important.
Well yes. But when you have no control the better lens can save your ass. The problem with professional photography is to be a tad better than the others. So even a small advantage can make one a winner. I’ve seen side by side comparison of 90 and 50-140. Unfortunately it wasn’t complete enough. 90 was also obviously only pre-production lens. But I’ll be dare enough to say that 90’s advantage is not small by any means.
Possibly, that’s why I’d like to test it. In my worthless opinion, the 90mm is phenomenal.
Well, I guess I’m overestimating the “other factors” then. As always, truth is somewhere in the middle… 😉
Bokeh on this pictures isn’t so bad. Probably you have never seen bokeh from 18-55 / 2.8-4 zoom. In the past I had plastic-fantastic Nikkor 18-105. Dark, built badly but bokeh was very nice (that time I thought that it is standard for all lesnes, now I now that it isn’t true). To the bokeh from 18-55 I just can’t used to. Helpfully Fuji’s primes are very good. I own 12, 23, 35 & 56 mm primes and use them exlusively all the time.
18-105 has positive SA on the long end. Thus it looses sharpness but gains great bokeh. This is typical for cheap standard and tele- zooms. I’ve not tested XF 18-55 but there is no point of discussing what’s worse. Let’s discuss what’s better. XF 56 non-APD definitely is.
That’s true, but what about nasty colours of theses shots? This is what X-Trans2 makes with colours – brings artificial greens, ugly magneta combined with yellows… X-Trans2 creates such pastels that look unnatural. I really hate X-Trans2 and can’t understand why Fuji spoiled exellent colours that first generation of cams (X-Pro1/X-E1/X-100) have. I am worried what sensor X-Pro2 will get – if it’s related with X-Trans2, than it will be disaster for me because I have been waiting for Fuji top cam without this contrived X-Trans2.
Because AF (tracking) speed was the only thing withholding me from buying into mirrorless (I’m spoiled by my Nikon D300), firmware v4 may very well give me the final push. I just hope it works as well on the X-T1 as on the X-T10…
Was she fast ?
i think Rico was used the 7fps burst bracket? you can notice the girl is far from start and very near at end
The only problem with the 50-140 lens is the bokeh that is in certain situations pretty bad… One needs to carefully choose the background if one wants to use it for portraits…. Really a shame for such an expensive lens…. Fuji needs to pay more attention to all characteristics of the lens if they are going to sell it for 1499€…
The thing is that you need to carefully choose your background with any lens if you want optimal bokeh. I think that you overestimate the role of the lens here. Look at the difference between the APD and non-APD versions of the 56mm. There’s only so much you can do with just the lens to get great bokeh .
The 90mm is pretty nice stopped down (f/16), even with a busy background, as my quite popular bokeh test has illustrated: https://www.flickr.com/gp/25805910@N05/b35T16
I can compare it with the 140mm zoom at 90mm when I get my 90mm lens back. Could be interesting.
Rico goes to the bother of doing an AF tracking test, and everyone moans about the bokeh!
Of course they do. What did you expect?
Silly me….I didn’t even look at the background. I assumed the point of the photos was to test the focus, so I was looking at the target subject rather than the background 😉
Well, the drill is to look for image parts that aren’t in focus, then complain about them. 😉
That said, I can’t remember showing a single image containing out of focus areas where someone didn’t complain about the bokeh. And why shouldn’t they? This bokeh thing is a matter of personal taste and preference. However, nobody says that it’s a matter of opinion, it’s rarely “I don’t like the bokeh in this image”, it’s always “the bokeh is bad”, “the lens is terrible”, like their personal opinion was some objective fact that everybody has to acknowledge and agree on. It’s like stating for everybody that “blue is an ugly color”, and if you don’t agree you are an idiot or paid by the blue color industry.
It’s too bad the new firmware for the X-T1 wasn’t available yet because I had to shoot a wedding this weekend (wife volunteered me to do it even though I didn’t want to)…and I had some missed shots due to the X-T1 not being able to lock focus on a moving subject in low, but not very low light. My Nikon Df didn’t do any better really. I didn’t have the right tools for the job, but maybe I would have had the right tools if the new firmware helps focusing in lower light and in general, as much as they say it does. At least it’ll be available before I go on vacation next month, unless they delay it.
Here are 175 shots from a wedding I shot with the X-T10 and beta firmware a few weeks ago: https://www.flickr.com/photos/25805910@N05/sets/72157651034658853
Firmware 4 looks okay so far, as long as there aren’t any major bugs (I have test shots scheduled for Thursday), I don’t expect any delay. It should be out on 25JUL, maybe even earlier.
Thanks, I’ll check it out.
JulY 25th? Wasn’t the firmware supposed to be out end of June?
hi Rico. still haven’t tried a XT10 or a FW4.0 on a XT1. please, is the “half release” shutter button problem (keeping the focus locked) solved? is now possible to freely move the enlarged parts of the frame, when checking the focus in 5x and 10x mode?
No, this also wasn’t part of the firmware 4 feature spec. I’d like to have a single point focus frame shift when zoomed it, too, it would be a time saver especially in MF mode.
yes i am mailing fuji since the XPRO1 vers 1… :/ in studio is really a pain to manual focus now… :/ was hoping with the 3.1 “direct focus point access” to see this, but nothing..
We have been mailing Fuji for almost 10 years to add a 24h clock format, to no avail so far.
25JUL ? I thought 22nd June
BTW thanks for the test, quite convincing!
25JUN, sorry.
GREAT wedding shots!!!!
You can get paid by the color blue? Because I love blue.
Wow, the close ups look airbrushed, just terrible. Skin texture is all but gone no acuity the Xtran 2 just ruins it. Why not go back to version 1 or Bayer (which plays great with regular converters).
Looks good!
Rico, are these shots taken in AF-S + Zone mode or in AF-C + Zone mode ?
In AF- C + Zone mode the exposure/WB measurement is only taken at the first shot within the burst. Is this correct ?
Yep, WB should remain constant during a burst. Obviously, series are AF-C.
Great shots Rico thats for taking the time to test the new AF features.
You are welcome. I’m sure the new modes will help many users achieve a decent keeper rate with predictable und unpredictable moving subjects.
Is the AF x-t10 faster than the x-t1 and x100t?
Thanks for the elaborate and challenging test: an erratically moving subject, a highly distracting background, but a nicely obedient AF tracking. I sincerely hope fuji nails this one, it would be really great to see it work for myself.
Rico, this is terrific. Thanks for doing this testing under what seem to be some pretty challenging conditions in terms of motion, background, etc. One question: With my X-T1, I’ve often found tracking a moving subject successfully in the EVF to be at least as challenging as getting the focus system to be responsive enough. Small delays in response can make it difficult to keep your subject centered while tracking. Given the fact that you’re both maintaining accurate focus and the subject is remaining centered in the frame, I’m wondering whether you’re seeing any improvements in that particular area? Also, did you use zone or full frame tracking (forgive me if I got the terminology wrong)?
The used modes are described in the respective overview file names.
Zones and Wide tracking make it easier to follow subjects. So less experienced action shooters should have a better success rate. Of course, I shoot with 3 fps and Focus Priority to maintain a good keeper ratio and keep screen blackout times as low as possible. But that’s no change to the previous firmware. A smart configuration is always useful, so is using high-performance mode.
OK, that makes sense, Rico, thanks much. Totally missed the tracking mode info when I looked thru the shots. Far as tracking delay goes, I should probably consider a somewhat slower frame rate for most times I’m tracking action — have been shooting at the higher rate more out of habit than necessity and it stands to reason that less time between frames would make it easier for the camera to keep up with the action. I always keep high performance mode enabled with the understanding that this impacts battery life. Really looking forward to the firmware update.
8 fps doesn’t allow any tracking, as the live view is lost. You only see the past, a frame showing what happened 1/8s ago. I can’t track anything in this setting, except stuff that moves perfectly towards the camera. Not to mention that 8 fps is only available to the center 9 AF frames.