The 4K Fuji X-T2 is here

[BenEricson]

17 minutes ago, Marco Tecno said:

I don't see anything, in those video, to write home about. Nothing that NX1 or GH4 could not already do (possibly better).

The baked in color is just amazing though, GH4 can't touch that.

[Marco Tecno]

Mattias, of course. But when a camera fails to reach the level of another released two years earlier, then I cannot understan all the "wow" factor around it.

I'd be the first in line if a new APSC camera comes with full readout, 4:2:2 10bit internal recording, 60fps at 4k and 240fps at 2K, at the price of NX1 two years ago. That's probably where samsung would have been today, with a mk2, had it continued the NX line.

Without samsung in the game, I only see average and old hardware at a big price.

[Lothar]

3 minutes ago, Marco Tecno said:

Mattias, of course. But when a camera fails to reach the level of another released two years earlier, then I cannot understan all the "wow" factor around it.

I'd be the first in line if a new APSC camera comes with full readout, 4:2:2 10bit internal recording, 60fps at 4k and 240fps at 2K, at the price of NX1 two years ago. That's probably where samsung would have been today, with a mk2, had it continued the NX line.

Without samsung in the game, I only see average and old hardware at a big price.

Maybe Samsung discontinued the camera business, because everybody would have expected such a camera. Meeting high expectations means, that if you can't deliver, you're screwed.

[Marco Tecno]

2 minutes ago, Lothar said:

Maybe Samsung discontinued the camera business, because everybody would have expected such a camera. Meeting high expectations means, that if you can't deliver, you're screwed.

Considering where they were 2 years ago, and where they are now with the recent exynos chips, I think it'd be very likely they could have that technology. Unfortunately, we'll never know, at least not in the form of a new NX camera.

[Mattias Burling]

28 minutes ago, Marco Tecno said:

Mattias, of course. But when a camera fails to reach the level of another released two years earlier, then I cannot understan all the "wow" factor around it.

According to your needs.

According to mine the Samsung NX1 couldn't reach the level of the much older and cheaper bmpcc.

According to deep sea divers none of them reached the level of the yet even older Gopro.

You dont see the "wow" of the Fuji for the same reason I dont see the wow of the GH4. Its simply not for us.

[Lothar]

27 minutes ago, Marco Tecno said:

Considering where they were 2 years ago, and where they are now with the recent exynos chips, I think it'd be very likely they could have that technology. Unfortunately, we'll never know, at least not in the form of a new NX camera.

Exactly. But coming back to the Fuji topic. Fuji and Olympus have something in common. Olympus has such a great stabilisation system. Everybody said, "If they could do proper video."
Fuji has such great colours, "If they could do proper video".

Now, Fuji did a huge step into the right direction. To me, this is good enough, because I also need very good stills. Of course, I would also be interested in slow motion, but every camera has it's downsides. For instance, I wouldn't shoot stills with an URSA mini.

[Francesco Spiezia]

Just got my hands on this camera and excited to see what it can do. I mean aside from a stabilization system and Flog it feels like a great 4K camera for the price along with an excellent still camera. Flog is always possible down the road through a firmware upgrade if Fuji is kind to us all. I can post something next week once I get the time to shoot some 4K footage. 

[Lothar]

Isn't it remarkable that this camera can flog?

[squig]

On 14/09/2016 at 0:12 PM, squig said:

We can get some accurate rolling shutter figures if somebody can do a quick test and upload the footage for Samuel http://www.dvxuser.com/V6/showthread.php?303559-Measuring-rolling-shutter-put-a-number-on-this-issue!

You can send me a link and I can pass it on to Samuel. A 1080p test would be good too.

I've done these tests myself, only takes a couple of minutes, you just need a perfectly levelled tripod. I shot the edge of a tall black bookshelf beside a white wall. Thin things like poles don't work, they turn to mush when you do the whip pan.

Surely somebody can spare a couple of minutes to test the rolling shutter. I want to sell my A7s and get an X-T2 for stills and BTS video, but the rolling shutter is a concern. I know the 4k rolling shutter is high, but without an accurate test it could be anywhere between 25 and 40ms, and I have no idea how high the 1080p rolling shutter is.

The OIS works well on the 18-55mm, no rolling shutter micro jitters like I've seen with longer lenses. Autofocus is nowhere near as smooth as Canon's DPAF, but maybe there's a control to slow it down a bit.

 

[Inazuma]

9 hours ago, squig said:

Surely somebody can spare a couple of minutes to test the rolling shutter. I want to sell my A7s and get an X-T2 for stills and BTS video, but the rolling shutter is a concern. I know the 4k rolling shutter is high, but without an accurate test it could be anywhere between 25 and 40ms, and I have no idea how high the 1080p rolling shutter is.

The OIS works well on the 18-55mm, no rolling shutter micro jitters like I've seen with longer lenses. Autofocus is nowhere near as smooth as Canon's DPAF, but maybe there's a control to slow it down a bit.

 

I just did a series of tests for rolling shutter and noise. Uploading now. Will take a good couple hours. I will say the results look very good IMO. Cinema5d said in their preliminary review that the rolling shutter in 4k looks about average, which I agree with. So probably between 22 and 27ms.  In 1080p the rolling shutter is among the best. And the image is very detailed. Not as good as downscaled 4k but still very good.

As for the usability of this camera. There are a few problems beyond what I mentioned in my last post. First is that autoexposure is actually dogshit. The exposure changes very slowly and in steps rather than smoothly like Sony's and Nikon's. Second is that the colour profiles like to crush the blacks very quickly, so DR is limited. Although you can bring up shadows very cleanly in post. Third is that you can't change ISO while recording. All these things basically mean that you have to be quite mindful when recording.

If your style is run n gun you might have a bit of trouble with this camera but if you can deal with some of the quirks youll have one of the best images in a DSLR form factor for the price and size. 

[squig]

1 hour ago, Inazuma said:

I just did a series of tests for rolling shutter and noise. Uploading now. Will take a good couple hours. I will say the results look very good IMO. Cinema5d said in their preliminary review that the rolling shutter in 4k looks about average, which I agree with. So probably between 22 and 27ms.  In 1080p the rolling shutter is among the best. And the image is very detailed. Not as good as downscaled 4k but still very good.

Yeah 1080p looks about as detailed as the A7s, except there's some aliasing. Not as detailed as the 5D Mk3 raw, but good enough for behind the scenes footage, the kind of stuff I don't have time to grade.

I noticed there's some control over the shadow level.

[Inazuma]

2 hours ago, squig said:

I noticed there's some control over the shadow level.

I take back what I said in my last post. I just did some side by sides with my Nikon D5500 on the Flat profile (which i've previously described as having as much DR as the C100 II) and the Fuji is actually as good or better. 

Attaching screencaps. Some extreme dynamic range tests.

Here's the D5500 with Flat profile

The Fuji at 1080p with ProNeg Standard profile, -2 shadows and highlights, -4 sharpness.

+1 stop of exposure compared to the Nikon (because the Nikon is brighter at the same exposure settings)

The Fuji at the same exposure settings as the Nikon:

Now with contrast at -100 using the Lumetri panel in Premiere Pro:

These were all at 1080p. At 4k there is even more headroom and the colours tend to be noticeably nicer.

 

[squig]

13 minutes ago, Inazuma said:

I take back what I said in my last post. I just did some side by sides with my Nikon D5500 on the Flat profile (which i've previously described as having as much DR as the C100 II) and the Fuji is actually as good or better.

I've been doing some grading tests with X-T2 stuff I've found online; the DR is very good, with Flog it may even equal the 5D Mk3 raw, ie more graded DR than the BMMCC and the A7s. It's looking like the X-T2 is everything I was hoping the A7s would be. Have you done any black dot tests?

[Lothar]

Isn't this a gorgeous camera on that gimbal? I find it very important that the tool suppliers recognize the Fuji X-T2 as a video camera.

And he's even trying to flog Johnnie Behiri. :-)

 

[Hanriverprod]

 

[Inazuma]

I spent a lot of time with this camera over the weekend, including shooting one small awards event with it. On Saturday I shot some tests and last night I did a load more in a controlled environment, alongside the Nikon D5500

I've uploaded all the raw videos to YouTube. Enjoy https://www.youtube.com/user/sebcastilho/playlists

(Please excuse my voice)

I'll be using this channel now to upload more tests I've done with this camera and others I've had in the past.

[squig]

2 hours ago, Inazuma said:

 

I spent a lot of time with this camera over the weekend, including shooting one small awards event with it. On Saturday I shot some tests and last night I did a load more in a controlled environment, alongside the Nikon D5500

I've uploaded all the raw videos to YouTube. Enjoy https://www.youtube.com/user/sebcastilho/playlists

(Please excuse my voice)

I'll be using this channel now to upload more tests I've done with this camera and others I've had in the past.

The noise difference against the D5500 at 6400 ISO is huge.

[Walter H]

@Inazuma Interesting tests, thanks. Notable that you had to raise the exposure of the Nikon 1 2/3 stops to match exposure at each ISO value, is that right? 

On another note, my XT-2 is still firmware 1.0 and beta testers were using up to 1.42, which we don't have access to on the Fuji support website. I imagine they are waiting for a complete 2.0 firmware before releasing an update to us? 

[AaronChicago]

Fuji colors are just legendary. The XT-1 is the first stills camera I've owned where I don't do any editing on them. I just use the Provia color profile. The XT-2 is very tempting with video now.

 

[Inazuma]

2 hours ago, Walter H said:

@Inazuma Interesting tests, thanks. Notable that you had to raise the exposure of the Nikon 1 2/3 stops to match exposure at each ISO value, is that right? 

On another note, my XT-2 is still firmware 1.0 and beta testers were using up to 1.42, which we don't have access to on the Fuji support website. I imagine they are waiting for a complete 2.0 firmware before releasing an update to us? 

Not quite right. I didn't use the same shutter speed and aperture values between the cameras because I let the Fuji them decide for itself. It just so happens that the Fuji seemed to expose much more brightly than the Nikon, so I in turn had to use manual exposure values on the Nikon that read +1 2/3 stops on its meter.  

The 1.43 firmware was an update for the preproduction model and as far I'm aware that firmware is 1.0 in the release model