androidlad

Would wireless follow focus work well on fly-by-wire lenses such as Fujinon XF ones?

DR400/F-log mode has plenty of fine grain with NR -4, even at base ISO, no need to add in post

The 4K interframe NR is a handy feature for special occasions, when enabled the images are clean and completely usable up to ISO12800, with a small amount of detail loss.

41 minutes ago, sanveer said:

CMOSIS is now AMS, but you got the sensor Spot On. It is super capable, thought not necessarily from a ILC or a Large Fixed lens sensor viewpoint. The sensor can do Full Res 300fps at 10-bit and 132fps at 12-bit full res. It has HDR mode where it has the option of doing it through different ways. The Problem is that the Photo Resolution is limited to 12MP. And the Even Bigger problem is that the sensor is an APS-C sensor. It should have been atleast 20MP if not 24MP. 

 

https://ams.com/cmv12000

At 5.5um its native DR is only 60dB, while A7R III offers 80dB at 4um...

I can't imagine the further DR penalty if they increase the pixel density.

Looks like a CMOSIS CMV series sensor:

5.5um pixel size, global shutter and 12bit ADC match several CMOSIS SKUs

And the funny thing is the 60-90dB dynamic range swing is exactly the trick from CMOSIS, native DR is 60dB, extended DR is 90dB with Full Well Capacity Temporal Limiting.

I doubt CMOSIS would custom produce a FF version for Pixii, so the most likely SKU Pixii is going to use is CMV12000, an APS-C sized sensor that matches all announced specs.

48 minutes ago, eyesuncloudedphoto said:

I wouldn't say the 16-55 is sharper than the Sigma in "absolute" terms; have used both the Nikon and Canon versions on Fuji cameras, as well as the GH5. both with speedboosters and dummy adapters. The point is, the 16-55 is perfectly matched to the X-Trans sensor, can't put my finger on it, could be micro contrast, color rendering or whatever. The end result is, it produces a magnificent image, both for video and stills.

Speaking of the AF, hell yes, it's bloody instant. This is true of other linear motor Fujinons also. I tested the X-T3 with the lowly 15-45 yesterday, a lens you can pick-up for 150 euros new, in some places, and it's tiny and feather-weight. It has an electronic focus ring that provides a smooth zooming action, and AF is so fast that it fools you into thinking it's parfocal. 

I'm preparing a review of the X-T3 for (probably) the next 10 days or so; although I fried the X-H1 in a previous review, I think the X-T3 is the bomb. Shame Fujifilm couldn't hold on and release the X-H1 with these exact same specs, instead of hurrying it to the market and having to face the marketing onslaught of the a7iii.

 

Thank you for that! Great info!

Speaking of batteries, I did a quick test of the 126S on the X-T3, and findings were surprising. 

In the X-T2 and X-H1, the battery lasted somewhere between 40 and 50 minutes in video. The X-T3 managed one hour, 20 minutes continuous recording (here is a screenshot of the last frame of my test video). This was 4K 100mbps recording, completely static (so no AF), and, although it's far from the ~2.5 hours on the GH5, it's still much better than the previous situation. They probably did some magic with the power circuit in this camera. Worthy of note that the battery indicator was very accurate; in fact it remained on red for more than 20 minutes (whereas in previous cameras it would turn red and after a couple of minutes die suddenly). Have also to note that the camera started overheating (not to an alarming degree though, and no protection circuit kicking in).

X-T3 has many power management optimisations. One of the key aspect is the IMX571 sensor has significantly reduced power consumption, compared to the 24MP IMX271.

7 minutes ago, Attila Bakos said:

Just did a very non scientific power bank test with the X-T3. I fully charged the battery inside the camera with the factory charger, and I wanted to see how much time it takes until 1 unit on the camera display's battery level indicator is lost when using 4K60 recording (and boost mode turned on) with a power bank attached:

Anker 20000 PD: 25 mins until heat warning, I decided to stop the test there, indicator still showed full battery
Asus ZenPower 10050 (no PD on this one, just a regular power bank): 17 mins
No power bank attached: 13 mins

It seems it's a good idea to get a power bank with PD support. Actually this is what Fuji recommends as well.

Scientific tests from Fujifilm website:

5 minutes ago, DBounce said:

I mean compared to cinema lenses. They generally don’t bother with 35mm equivalents.

Really? I thought all lenses, including medium format and 65MM/IMAX lenses are marked with 35mm focal length?

1 minute ago, DBounce said:

From a quick play... salivating makes total sense. This lens is ever bit as good as you have heard. Important note for anyone thinking of picking one up. Remember that the focal range is as marked... no need to multiply by 1.5. So 18-55mm is really 18-55mm.

What do you mean no need to multiply by 1.5? I thought all lenses are marked by 35mm focal length. So 18-55 would have FF equivalent FOV of 28-85?

2 minutes ago, Attila Bakos said:

If you don't like the 4GB file splitting and don't mind a bit of work, download ffmpeg, put ffmpeg.exe and your mov files you want to concatenate into a folder, and create a bat file (in windows) in that same folder with the following two lines:

(for %%i in (*.mov) do @echo file '%%i') > mylist.txt
ffmpeg -f concat -i mylist.txt -c copy output.mov

Running this will concatenate all the mov files in that folder without re-encoding into a file named output.mov. Keep in mind that on an average HDD this can take at least 1 minute/4GB.

The 4GB file splitting will be fixed in a firmware update in November.

20 minutes ago, Danyyyel said:

That's quite good for both slow-motion mode at 60p (56mbps) and 144mbps 120p (I would have like hdmi output). Somehow they could have upped the normal 24/25/30 to 50 mbs for example. An internal 50 mbps 422 even in 8 bit (10 bit would have been better) would be broadcast ready. I am forced to use a Ninja now.

Well X-T3 can do 200Mbps for 1080p at all framerates.

More user reported problems with Ninja V:

Quote

The Ninja-V was NOT behaving as expected; it still needs some special treatment when used with the Fuji X-T3. E.g: if you end a recording via the camera, the Ninja will flash black and the kangaroo icon appears; after that, most of the time, the Ninja will stutter and you will have to remove and plugin the HDMI. Nothing that a firmware update can't fix i guess.

How do you monitor in 2.39:1 or other aspect ratios? There's no framing guidelines in DJI Go app.

23 minutes ago, Andrew Reid said:

Rumours sites aren't an authority on technical matters, they just identify second hand info and label it. That chipmod source just tells us the manufacturing process is Sony. Doesn't say the design is.

And made by Sony does not mean it's off the shelf Sony.

Nikon designed the sensor in the Z7, Sony fabricated it for mass production.

The sensor has Nikon IP and circuit layout, Nikon specification, Nikon micro lenses and other patented Nikon design aspects.

It has a Sony BSI manufacturing technique, on Sony mass production wafer, Sony quality control and made in numbers determined by Sony's manufacturing capacity at their factory, which isn't as high as Nikon would like because they have so many other customers. Nikon should diversify their manufacturers, but making a sensor like this is hard. You won't see a BSI sensor from Canon. It doesn't exist.

It's a teardown, the Sony SKU is clearly marked on the sensor assembly ?

Just putting this out there - Nikon does not have the capacity to design image sensors from scratch, like many big companies such as Apple, Fuji, Phase One, Hasselblad, they buy off-the-shelf products from Sony Semicon, the key difference here is Nikon secured exclusive supply from Sony, the IMX309 in D850 is an example, and possibly Z7 as well.

5 minutes ago, Mattias Burling said:

Link?

Everything Ive read prior says it aint so. Im just curious.

https://nikonrumors.com/2018/06/15/new-reports-indicate-that-the-nikon-d850-sensor-imx309aqj-is-made-by-sony.aspx/

16 minutes ago, Castorp said:

I was under the impression that Fujifilm, Nikon, et c, design and specify their own sensors? Sony manufactures some components?

To say that a component manufacturer “made the sensor” seems a bit rich? 

Also, there seems to be more to the final image than just the sensor. There’s a whole chain of things that make the image such as processor and so on. 

With film it made more sense to ascribe importance to the emulsion and who made it, but even with film, things like lenses and control of film flatness, the behaviour and accuracy of the shutter and aperture mechanisms have great impact on image quality.

Fuji designs the CFA on top of the Sony sensor. While Nikon just use the Sony sensors as is.

Atomos has a normal priced one too: http://amzn.eu/d/dPjvpYa

4 hours ago, Andrew Reid said:

NX1 was down at 2ms like an Alexa in 1080p for rolling shutter, I once saw, can't remember where.

I am surprised no camera has a distortion correction mode in firmware. Why don't they correct rolling shutter in-camera, just crop-in slightly and re-align the image.

Alexa's rolling shutter is about 4.5ms, the minimum required to reach 200fps full readout.

How would a camera correct rolling shutter when a train flies through the frame while the camera is static?

11 minutes ago, deezid said:

Seems like the X-T3 even captures more detail than the NX1.
Wish they would have dialed down sharpening in the comparison to make it smoother.

Hopefully Slashcam will do a test as well, they usually include standard vs. all the way down sharpening samples.


Edit:
Even the 4k 60p looks great!

Yeah I wonder how the 1080p is derived from 6K, the image is sharp but has heavy aliasing. Looks like a combo of pixelbinning and lineskipping.

21 minutes ago, Andrew Reid said:

Nah, it's not.

The Fuji sensor has an X-trans CFA which is Fuji's property, so the chip cannot be listed for sale off the shelf by Sony without violating Fuji's IP.

Sony 26MP APS-C sensor readout speed is listed at 19fps at full resolution.

The Fuji X-T3 sensor readout is much faster.

Full pixel readout at up to 30fps (over 6K res) and 60fps for 4K with that slight crop of the sensor readout. The Sony chip at 19fps full res, just isn't fast enough to enable  4K/60p from that 'almost' full sensor resolution readout.

Come on Andrew, the NDA lift was yesterday. Fuji staff has confirmed X-T3 uses IMX571 on Weibo.

Sony sold the sensor to Fuji and made the CFA on top of the sensor to Fuji's specification.

The difference in readout speed is simply caused by bitdepth. The quoted 19fps resolution is at full 14bit. The sensor can also operate at 12bit and 10bit.

9 minutes ago, wolf33d said:

So the 16mm 1.4 is no good for video AF? 
Might get a 16-55mm then. No IS and no IBIS though

All the 1.4 primes work ok and actually improved with X-T3. But they're highly unpredictable.

You could wait for the new 16-80 f/4 zoom with OIS.

12 minutes ago, wolf33d said:

Does someone has a link that says which Fuji lens have good AF tracking performance in video, and which are to be avoided? 

Edit: answering myself: https://alikgriffin.com/a-complete-list-of-fujifilm-x-mount-lenses/

Edit2: this gives info for contrast AF only. Not relevant. My question still stands

All of the new f/2 primes and f/2.8 zooms.

DPReview just published X-T3 lab test for video, the 120P quality is significantly better than A7 III and GH5.

https://***URL removed***/reviews/fujifilm-x-t3-first-impressions-review/7

12 minutes ago, thebrothersthre3 said:

wait the XT1 can't possibly look that bad

Sadly it's true

https://***URL removed***/reviews/fujifilm-x-t3-first-impressions-review/7

Select X-T1

NDA lift on X-T3 sensor, it's now listed on Sony Semicon website, as IMX571

https://www.sony-semicon.co.jp/products_en/IS/sensor2/products/index.html