Taranis

8 minutes ago, jonpais said:

I read the thread. The last example with the 10-24 contains no information whatsoever about anything - how it was shot, which settings were used, post processing... In fact, it tells us nothing at all, either about the camera or the lens. All of the YouTuber's clips are ten seconds long and give no information. Why bother posting it?

Because I thought it shows the exact same issue. You can easily ignore it if you don't like it.

8 minutes ago, KitaCam said:

I don't believe I'm witnessing this with my footage using the 18-55mm. I could be wrong, but in those examples it seems very obvious to spot. Possibly an issue linked to continuous focus? Application of stabilisation via software etc.?

Yeah I haven't seen this in my footage either, it seems to be a rare problem. It's not a post processing error. 

I posted two examples of in camera lens correction problems, to prove that in camera correction works for video too. Please read the thread before you start to criticise.

6 hours ago, Luke Mason said:

Thanks, that looks very concerning...

By accident I just found another one, this time with the 10-24 lens. I wonder if this is some kind of lens problem, or body.

9 minutes ago, Luke Mason said:

Quick question for those XT-2 video shooters, does the camera apply chromatic abberation, shading and distortion correction to the images in video mode?

From what I've seen so far I'm pretty sure it's applied in video mode too. An evidence might be this video where the correction is faulty. It's been shot with the 18-55 btw.

I uploaded a short video for reference. It contains 4 of my sharpest clips from my last trip, using the 18-55: https://www.dropbox.com/s/y5tselgh7a3yol6/tihany_spring.mov?dl=1
These are all shot with sharpness set to -3, at high shutter speeds with OIS left on.
My other clips from this trip are all blurrier than these 4, a lot of them are completely unusable. I'll do some more tests this week at the same location with the same settings but with OIS turned off. I'd like to know if I have a focusing issue and/or an OIS issue.

Another screengrab from a recent footage with the 18-55: https://www.dropbox.com/s/qyesr6rvh25g6el/DSCF1519.jpg?dl=1
This is with -3 sharpness, looks okay.

1 hour ago, BTM_Pix said:

That log stuff I put up was with the 18-55 

Okay, this is a screengrab from my movie:

https://www.dropbox.com/s/weec4k8c4jdmpot/compare1.jpg?dl=1

And this is one from BTM_Pix's ProNegStd.mov:

https://www.dropbox.com/s/tni42dns4ibr9li/compare2.jpg?dl=1

This one is even more blurry than mine, at least to my eyes, but I let you guys make the judgement.

1 hour ago, jonpais said:

No, it is definitely the 18-55mm. If you can't tell this is sharp at 1080p, there's a problem with your monitor.

It's sharp when I compare it to other 1080p footages, but I watch it on a UHD monitor, so it can't be as sharp as a properly shot UHD footage, naturally.

2 minutes ago, jonpais said:

I believe this was shot with the 18-55mm, but there's no documentation of that.

 

Yeah we don't know that for sure, and this is a 1080p footage, hard to make a judgement.

8 minutes ago, jonpais said:

These were all shot with the X-T2, screen grabs from 4K footage, they're not examples of perfection or anything, but there should be more brilliant highlights and crispness than what I'm seeing.

 

Yes I see the difference, but I don't know how sharp those scenes would look with the 18-55.

Just now, jonpais said:

May I ask if you are shooting with the lens closed down to f/22 or something? Because the X-T2 is capable of so much more detail than what I'm seeing here.

No, I rarely use smaller than f/11. It would be nice if another 18-55 user could comment on the sharpness of this lens. Should I be worried?

1 minute ago, jonpais said:

Moire is going to crop up from time to time with any camera, and I find the X-T2 does a fine job, even with patterned fabric. Your footage looks soft and underexposed to me. It doesn't have any sparkle in the highlights, and the shadows are muddy. 

Yes it was one of my first, it's mostly underexposed, I only posted it to show moiré. You might remember this one, is this sharp to you? This is with the same lens, but with -3 sharpness. If this is sharp to you, then -4 might be too low for this lens after all:

13 minutes ago, jonpais said:

These images are all soft, they look like they were shot through a filter with vaseline. Something's not kosher here.

Are you sure you are not talking about the D5200 footage? The X-T2 footage is fine, but from 0:33 you see the moiré at -4 sharpness.

Just now, Luke Mason said:

Try all the way down if -4 is the lowest you can go, I don't own XT-2 but the clip you posted has a bit too much digital sharpening.

Check this roof in 4K, I shot it with -4 sharpness. It doesn't really help.

6 minutes ago, Luke Mason said:

Judging from the clip It's not the common type of moire which should be mostly chromatic (jumping with colour), was the sharpness setting turned all the way down? It might be also due to the in-camera debayer algorithm for video, considering the camera is debayering from 5K area of the sensor, downsampling to 4K and compressing it in real-time, the debayer quality could be different from what you get from RAW stills in Capture One or Lightroom.

The sharpness was at -3. I used to have it at -4, but then I read somewhere that at -4 some extra blur is applied. I did not test this at all, just used -3 from that point on. Does anyone have info about this? I can't test it for a while.

55 minutes ago, BTM_Pix said:

Its difficult to say one way or another but of its constant then that would suggest a problem but when you do have OIS on for low shutter speed do you see the same problem?

The only way to be sure is to put it on a tripod and use a shutter release cable (or the app or self timer) to take away that variable and shoot a page of text that fills the frame. Repeat that at shutter speeds from below, above and way above the reciprocal (use ND to keep aperture and ISO constant) with OIS on and off at each speed and also using both OIS modes.

That way, you should end up with 9 samples and at 35mm for example I would go with the following 

1-3 - 1/15th with OIS Off, On continuous and On Shooting only

4-6 - As above using 1/125th 

7-9 - As above using 1/800th 

Because of the suspected cause of the issue perhaps being the timing of where the OIS is in terms of its measurement cycle when you press the shutter, I would say that you should probably run multiple passes of 8 and 9 as these are the most likely to fail.

If you test it in this way then it should eliminate the lens itself having an issue as you will be running it in perfect conditions off a tripod and with a remote release. 

It will also aid your case in getting a straight replacement if there is a fault as you can demonstrate it easily.

If there is no fault when running the test then for your own piece of mind, I'd remove it off the tripod and try the same test handheld as if doing that brings the fault back then you should be able to see the point at which it's happening.

My guess is that that point will be test 8... 

Thanks a lot, I'll test this out as soon as I can!

8 hours ago, BTM_Pix said:

The usual rule of thumb is not to use it unless your shutter speed is below the reciprocal so, yeah, with a crop sensor that would be 1/80 at the long end of the 18-55.

At 640/f8/200 iso that you shot that then you are obviously well above that threshold where OIS would be required and it can certainly have an effect on sharpness.

But its not always the case as it seems to be timing based on where the OIS system is in its measuring cycle when you press the shutter, which is why there seems to be conflicting reports about it. One thing that does seem to have an impact though is if you set the IS mode to 'SHOOTING ONLY' then it will only engage it on a half press or shutter release as opposed to it being continuous. So if you are going to leave it set to ON on the lens then setting the IS mode to 'SHOOTING ONLY' in the 'SHOOTING SETTING' menu should give you a bit more of a chance.

My problem is the other way in round in that I have to remember to switch the OIS to 'ON' on the lenses when I'm doing video as I very rarely use it when shooting stills as its always fast shutter speed stuff.

 

Thanks, I'll switch OIS off and try to remember to switch it on then. But do you think this something that should be serviced?

9 hours ago, Inazuma said:

@Taranis That photos looks blurry all round to me :/ Probably a faulty lens then? Sometimes I will get blurry shots with the 18-55 if AF-C is on though.

Yeh I think the X-Trans benefits are a myth or marketing hype. Although I've never seen anything as bad as what you posted. Strange!

Yeah it's not the sharpest in general but this one shows what my problem is. The edges are clearly sharper.
About moiré: If you never seen anything this bad, then you are lucky. This is not my first rooftop that makes a footage unusable. For me it's been black or white so far, either I don't see any moiré, or I see it to this extent.

I bought the X-T2 for video, but every now and then I take photos too. I use the 18-55 and I noticed some kind of flaw and I'm not sure if I should send it back for repair. Some of my photos are only sharp at the edges and blurry in the center. I've been told that the OIS is the cause, and I should switch it off when the shutter speed is faster than 1/80. I did not test it yet but I don't really want to focus on turning that thing on and off. Here is an example, check it at full res:

On another topic, I thought the X-Trans sensor was developed to fight moiré, well, this roof is not looking good, again, check it at full res:
https://www.dropbox.com/s/nri30p77o3nqnv0/xt2_moire.mov?dl=1

Uh, that's massive...

32 minutes ago, BTM_Pix said:

For completeness, here is a (very) short compilation extract of a couple of the shots as they were recorded to the Ninja Assassin using the ProNegStd -2/-2  profile in 4K ProRes 422LT

https://mega.nz/#!dj5AWagD!_TgG79KA4Z21eJHTMoUPj3cYBQ2HNb60oLRE6m79dxU

Thanks. For the boat scene I prefer the ProNeg Std version, the restaurant scene was the only one where I found F-log to be useful:

This is F-Log (not a proper grade, just pushed DR for comparison):

And this is ProNeg Std, if you try to achieve the same:

Lots of noise. However, in situations like this you can let the background blow out a bit, or light the scene if possible.

2 minutes ago, BTM_Pix said:

If I get a chance I'll do an internal/external comparison with ProNegStd -2/-2 at some point.

 

Thanks, that would be nice! I was considering the BM Video Assist 4K too, but I read too many reviews about it being unstable. And yeah, if you can get a 4K60p capable recorder for a little more money, it's reasonable to save up for that one, as a recorder can easily outlive a camera body.

Yes, I tried both Fuji LUTs. I played many hours with the F-Log footage posted so far. I really wanted to see how it compares to ProNeg Std as I was about to purchase a Ninja Inferno. Now I'm in doubt. I could get decent results with F-Log and I liked what I saw but when I compared it to ProNeg Std, it was not better at all, just a bit different. I just don't see anything with F-Log that can't be done with ProNeg Std, unless you really need those extra stops. I'd still like to see some F-Log footage exposed to 90-95 IRE, but I doubt it would be different. Above middle gray the F-Log curve is almost a straight line so you won't have more code values between stops in the highlights. So a Ninja Inferno would be good for two things. First, for monitoring, but there are cheaper (and smaller) options for that. Second, for ProRes 4:2:2 recording, but I'm yet to see a comparison between ProNeg Std -2,-2 4:2:0 internal and 4:2:2 external. I wonder if there is any noticable benefit.

13 hours ago, Inazuma said:

@BTM_Pix Thanks, just had a little play around with that footage. I think you would benefit from exposing to 90 or 95 IRE rather than 100. I am not able to retrieve much colour information in the brightest parts of the image. Also can you just clarify whether or not this post has the images ordered as FLOG then ProNeg or FLOG then FLOG graded?

Can this be the reason of what I see here?

I applied a curve to the ProNeg Std JPG posted earlier, to see what's in the highlights. The sky is not pretty but this is a compressed JPG.

And I tried to grade the F-Log footage to something similar, and the sky is ugly with lots of artifacts.

Seems like it has a crop: "In full-frame shooting, the angle of view will be narrower under the following conditions: – When [File Format] is set to [XAVC S 4K] and [Record Setting] is set to [30p]."

I tried to give it a warmer, sunset type of look.