Cross Hatching Wont Go Away

The place for questions about shooting with Blackmagic Cameras.
  • Author
  • Message
Offline

Tony Rivera

Blackmagic Design

  • Posts: 3372
  • Joined: Mon Aug 12, 2013 4:52 pm
  • Real Name: Tony Rivera

Re: Cross Hatching Wont Go Away

PostFri Oct 21, 2016 6:00 pm

chris.white wrote:I'm still going back and forth with the BMD support folks on whether my camera needs to be RMA'd. They seem to be unwilling to believe that there's any issue with my camera. Last time around, I sent DNG's shot at 4 different resolutions... all of which exhibit the crosshatch. The attached jpeg with enlarged examples is from 1920x1080 RAW DNG. Yet, here was the last response I recieved...

"I took a look at your images from the camera and I dont see any patterns in these images that would justify RMA. You can still send it in for us to test but I think that would be a waste of your time."

Am I just hallucinating? That does show the crosshatch issue, right?
---
c
There seems to be have some confusion about this. We will be reaching out to you about this again to see what we can do for you.
Support: http://www.blackmagicdesign.com/support
Info: http://www.blackmagicdesign.com/company

Follow us on Instagram:
@blackmagicnewsofficial
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostFri Oct 21, 2016 7:23 pm

Earl R. Thurston wrote:
Earl R. Thurston wrote:Dmitry, are you able to post one DNG file from each test? I'd like to try some experiments in Adobe.

Thanks for sending me those DNG's, Dmitry.

I have also been able to confirm that the "BayerGreenSplit" fix in CornerFix (or ExifTool) improves the output on these DNG files. I first chose one notable color square from each of the darker sample images:

The attachment RAW_F8_orange_anim.gif is no longer available
The attachment RAW+B80_F5.6_lime_anim.gif is no longer available


Where I really saw a difference, though, was in the shadow on the wall behind the case:

The attachment RAW_F8_background_anim.gif is no longer available


This is the kind of natural part of an image that could really suffer from this problem, as it should otherwise be a smooth, soft gradation.


I also test this method today with ACR, Resolve and Raw Photo Processor apps. Resolve and RPP don't react to this at all. btw RPP has its own green-green pixel turning setting slider. As i know It mainly used for WB but i don't try to test it yet.
But i really like how image details looks after using this fix on Micro Camera. Grain became more cleaner and even FPN in pushed deepest shadows looks like slightly less visible too.
So it is worth to fix this for BMMCC camera too. C'mon BM, is it so complicated to add slider to Resolve to enable fine tune that green pixel setting ;) ?
Screen-Shot-2016-10-21-at-10.20.48-PM.jpg
Screen-Shot-2016-10-21-at-10.20.48-PM.jpg (609.39 KiB) Viewed 13638 times
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

James Alexander Barnett

  • Posts: 193
  • Joined: Sat Jan 09, 2016 5:48 pm
  • Location: UK

Re: Cross Hatching Wont Go Away

PostSat Oct 22, 2016 6:00 pm

Okay so I was doing some tests with my Video Assist rigged up acting as a Proxy recorder and when I came to look at the 4.6k footage straight from the camera I had crosshatch hell.

I have done tests before in good light and in all honesty I couldn't see anything, I have shot 3 paid jobs this week at 2k 16:9 and the images have been beautiful no noticeable issues.

The clip in this screen cast is:

4.6K
ProRes 422
3200K
1600 ISO
25FPS
Sayang 85mm (T1.5)

Screen capture resolution is 2560x1440 which is my iMac screen resolution.

The lighting was very poor so as others I think have mentioned this makes the issue more apparent, the image seems normal when scaled at 1/4 percentages so 25%, 50%, 75%, 100% but anything other than that and its horrible. This really can't be normal or how it's supposed to be, imagine doing a final output of a project and the client watches it full screen on a display which isn't an exact 25% increment of the full resolution its going to look like complete crap.

Sorry the images are a little crude I didn't intend to post this but thought I had too seeing just how bad it was.

Whats the score with this BM?

Screen capture Crosshatch video at the link below↓
https://vimeo.com/188432814/1635f6eefe
James Alexander Barnett
www.jamesalexanderdp.co.uk
URSA Mini 4.6k
Offline
User avatar

James Alexander Barnett

  • Posts: 193
  • Joined: Sat Jan 09, 2016 5:48 pm
  • Location: UK

Re: Cross Hatching Wont Go Away

PostSat Oct 22, 2016 7:45 pm

Just another thought, has anyone tried scaling an image recorded in Ultra HD using the 4K video assist?

Can anyone try testing? I'm presuming it will be the same, but worth a try.
James Alexander Barnett
www.jamesalexanderdp.co.uk
URSA Mini 4.6k
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostSat Oct 22, 2016 7:50 pm

Probably it will be the same in VA because this problem caused by in-camera debayering as we know now. The best you can do now is wrote one more problem ticket to BM support to stimulate them fix this problem quicker. Do you test Pocket camera for this Cross Hatching effect? Wonder is this only Micro/4.6k Camera problem, or older cameras where affected too?

P.S. BM support told me in reply that this problem is currently under investigation by their Development team.
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline

Travis Inghram

  • Posts: 8
  • Joined: Wed Jul 20, 2016 3:03 am

Re: Cross Hatching Wont Go Away

PostSat Oct 22, 2016 9:42 pm

So it looks like the problem is being mildly addressed. Hope to hear a resolution soon as this seems like something that warrants a recall, worse case scenario.
Offline
User avatar

James Alexander Barnett

  • Posts: 193
  • Joined: Sat Jan 09, 2016 5:48 pm
  • Location: UK

Re: Cross Hatching Wont Go Away

PostSun Oct 23, 2016 6:39 am

Dmitry Shijan wrote:Probably it will be the same in VA because this problem caused by in-camera debayering as we know now. The best you can do now is wrote one more problem ticket to BM support to stimulate them fix this problem quicker. Do you test Pocket camera for this Cross Hatching effect? Wonder is this only Micro/4.6k Camera problem, or older cameras where affected too?

P.S. BM support told me in reply that this problem is currently under investigation by their Development team.


Hi Dmitry,

Thanks for the info, I have noticed crosshatching once before on the BMPCC but it was only in certain lighting conditions, I tested scaling some properly exposed shots last night that I shot earlier in the week, and in outside lighting conditions the crosshatching is pretty much none existent, you can very slightly spot it in the blacks.

I will do a screen cast tomorrow morning comparing all the clips and the pocket camera also.
James Alexander Barnett
www.jamesalexanderdp.co.uk
URSA Mini 4.6k
Offline

roger.magnusson

  • Posts: 915
  • Joined: Wed Sep 23, 2015 4:58 pm
  • Location: Sweden

Re: Cross Hatching Wont Go Away

PostSun Oct 23, 2016 10:09 am

Dmitry Shijan wrote:C'mon BM, is it so complicated to add slider to Resolve to enable fine tune that green pixel setting ;) ?

I don't think that's the right approach. For one thing, would it work for ProRes/DNxHR? In RAW it's also a band aid at best since it seems to involve taking the average of two photosites (and doing so in the whole luminance range). Effectively losing detail. Hopefully camera calibration is performed for specific luminance values.

Dmitry Shijan wrote:P.S. BM support told me in reply that this problem is currently under investigation by their Development team.

Excellent! Thanks for letting us know.
Offline

Earl R. Thurston

  • Posts: 152
  • Joined: Sun Sep 04, 2016 4:52 pm
  • Location: Burnaby, BC

Re: Cross Hatching Wont Go Away

PostSun Oct 23, 2016 3:57 pm

roger.magnusson wrote:In RAW it's also a band aid at best since it seems to involve taking the average of two photosites (and doing so in the whole luminance range). Effectively losing detail.

That's not quite how it works. The DNG still contains two separate green channels. Changing the "Bayer Green Split" value calculates a new, full-resolution luminance channel, just with a different balance in the odd and even rows / columns. So, it's a suitable correction, but it's more work, involves additional programs, and only appears to work in an Adobe workflow.

When people use a 0.5 pan and 0.5 tilt in Resolve, then yes, that averages the pixels and lowers resolution.
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostSun Oct 23, 2016 5:44 pm

If someone interested here are few grey card DNGs frames from BMMCC. Even without noise reduction Cross Hatching pattern is visible there very nice at 200-300% zooming in and at custom ratio zooming out.
https://www.dropbox.com/s/1xmu479qh37fbqn/CAM1_2016-10-23_1953_C0005_000051.dng?dl=0
https://www.dropbox.com/s/dvjhe2m7afrjoj4/CAM1_2016-10-23_1953_C0005_000099.dng?dl=0
https://www.dropbox.com/s/qcf5bwkeqj1gp4x/CAM1_2016-10-23_1953_C0005_000168.dng?dl=0
Screen-Shot-2016-10-23-at-9.55.57-PM.jpg
Screen-Shot-2016-10-23-at-9.55.57-PM.jpg (262.04 KiB) Viewed 13393 times
Screen-Shot-2016-10-23-at-9.55.56-PM.jpg
Screen-Shot-2016-10-23-at-9.55.56-PM.jpg (256.38 KiB) Viewed 13393 times
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline

roger.magnusson

  • Posts: 915
  • Joined: Wed Sep 23, 2015 4:58 pm
  • Location: Sweden

Re: Cross Hatching Wont Go Away

PostSun Oct 23, 2016 8:21 pm

Earl R. Thurston wrote:
roger.magnusson wrote:In RAW it's also a band aid at best since it seems to involve taking the average of two photosites (and doing so in the whole luminance range). Effectively losing detail.

That's not quite how it works. The DNG still contains two separate green channels. Changing the "Bayer Green Split" value calculates a new, full-resolution luminance channel, just with a different balance in the odd and even rows / columns. So, it's a suitable correction, but it's more work, involves additional programs, and only appears to work in an Adobe workflow.

When people use a 0.5 pan and 0.5 tilt in Resolve, then yes, that averages the pixels and lowers resolution.


Ok, but if the balance is off only in darker parts, wouldn't a simple balance control mess up the rest of the image as well? I haven't seen any details from Adobe at all since the BayerGreenSplit parameter isn't very well documented.
Offline
User avatar

James Alexander Barnett

  • Posts: 193
  • Joined: Sat Jan 09, 2016 5:48 pm
  • Location: UK

Re: Cross Hatching Wont Go Away

PostSun Oct 23, 2016 10:14 pm

Okay so just filed a ticket with BM support UK regarding the Crosshatching and attached my screen cast with it.

Did some further testing this evening rolling my firmware back to 3.1 and as other people have mentioned it was a lot better, it was still slightly present but much better overall.

I have only been shooting ProRes, 2k & HD look fine, it's only above those resolutions that I'm encountering issues, as mentioned before poor light also adds to the problem.

My concern is because its a scaling issue everyone who watches the content could be seeing it differently, someone watching true 4k 1.1 pixels will be fine, others downscaling to fit iMac screen for example could be seeing horrible crosshatched images so its our of our hands in a way, I really don't feel comfortable shooting anything over 2k with this camera at the moment.
James Alexander Barnett
www.jamesalexanderdp.co.uk
URSA Mini 4.6k
Offline
User avatar

rick.lang

  • Posts: 11663
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Cross Hatching Wont Go Away

PostSun Oct 23, 2016 10:44 pm

Just an aside, but the iMac 5K Retina screen is 5120x2880, so all 4.6K footage can display 1:1 in a window on the desktop.

As OSO says, "It's all part of the plan." [If you don't know OSO, you don't have young children or grandchildren!]


Sent from my iPad using Tapatalk
Rick Lang
Offline

Earl R. Thurston

  • Posts: 152
  • Joined: Sun Sep 04, 2016 4:52 pm
  • Location: Burnaby, BC

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 12:25 am

roger.magnusson wrote:Ok, but if the balance is off only in darker parts, wouldn't a simple balance control mess up the rest of the image as well?

I would have thought that, too, but strangely it doesn't, at least not in the handful of DNG's I was able to test. I think the calculation is a bit more intrinsic than just a boost or reduction.

Would be interesting to hear if anyone had the side effect you mentioned, though.
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 1:49 am

If manually set incorrect second Green channel in RPP we can see exact same pattern artifacts. Also slight green cast is present but in Resolve probably it just compensates by adjusting overall WB.
Untitled-11111.jpg
Untitled-11111.jpg (457.26 KiB) Viewed 13357 times
Untitled-22222.jpg
Untitled-22222.jpg (559.51 KiB) Viewed 13357 times
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

James Alexander Barnett

  • Posts: 193
  • Joined: Sat Jan 09, 2016 5:48 pm
  • Location: UK

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 8:59 am

Okay so this is the response I have just received form BM, seems like a bit of a fob off to me and is totally different to what some others are being told on here too.

If it is a rescaling issue in my OS then how does using V3.1 improve the situation and how come it's only this camera I have ever had problems with?

----- Please reply above this line -----

Hello James,

thanks for your email,

this kind of behaviour is not actually in your footage, but it is more likely a rescaling behaviour in your OS. This is not something that we can currently control, but to confirm you should have a look at the picture on a 1:1 ratio through a reference monitor over an I/O card (like a decklink card).
In other words the rescale engine in OS can not always resolve the image resolution correctly and it creates some visual misbehaviour when trying to upscale or downscale the picture. Again this is not burned into your picture and your footage is perfectly safe to use.


Regards,

Michele De Benedetti
Technical Support Consultant
Blackmagic Design EMEA


----- End of message -----
James Alexander Barnett
www.jamesalexanderdp.co.uk
URSA Mini 4.6k
Offline

Larry Sullivan

  • Posts: 74
  • Joined: Thu Oct 17, 2013 4:18 am

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 9:49 am

James Barnett wrote:Okay so this is the response I have just received form BM, seems like a bit of a fob off to me and is totally different to what some others are being told on here too.

If it is a rescaling issue in my OS then how does using V3.1 improve the situation and how come it's only this camera I have ever had problems with?

----- Please reply above this line -----

Hello James,

thanks for your email,

this kind of behaviour is not actually in your footage, but it is more likely a rescaling behaviour in your OS. This is not something that we can currently control, but to confirm you should have a look at the picture on a 1:1 ratio through a reference monitor over an I/O card (like a decklink card).
In other words the rescale engine in OS can not always resolve the image resolution correctly and it creates some visual misbehaviour when trying to upscale or downscale the picture. Again this is not burned into your picture and your footage is perfectly safe to use.


Regards,

Michele De Benedetti
Technical Support Consultant
Blackmagic Design EMEA


----- End of message -----


Hmmm. The issue definitely goes away when I live feed out of Resolve to a full screen HD TV in cinema (full screen) mode, but there is something inherent in the footage itself that causes this behavior or else why doesn't our OS's have issues scaling other footage from other cameras?
Offline

David Hessel

  • Posts: 274
  • Joined: Mon Mar 10, 2014 7:53 pm

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 4:17 pm

That is a canned response from BM, we have seen that explanation many times before. If you want to know if your camera is affected open the footage in what ever program you want but make sure the timeline resolution matches the footage resolution. Zoom in to around 400%, if you see a screen door/crosshatching pattern then your unit is affected.
David Hessel
Offline
User avatar

Valentin Remy

  • Posts: 251
  • Joined: Thu Mar 24, 2016 3:52 pm
  • Location: Belgium

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 8:27 pm

I can't believe they answered that, outch...

If it appears, it's because there's something in the image that causes these scaling artefacts :/

That grid when we play the footage isn't showing just like that by itself, it's just a consequence.
Yeah, some players will handle that grid better (like in Premiere when we turn ON the High Quality Playback) when the footage is scaled. But this grid is still visible at the pixel level, it's there, why even arguing ?

My unit arrived today at BM, can't wait to see the results.
http://www.instagram.com/valentinremy.be/
Offline

Kyle Gordon

  • Posts: 403
  • Joined: Thu Dec 26, 2013 4:06 am

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 8:36 pm

What an awful response. I think BM people here should get in touch with whomever they have to above them, and get this straightened out from the top. If they are going to continue to pretend real issues don't exist, I think customers will look elsewhere for cameras.

What a huge, unnecessary mistake to make.
Kyle Gordon
Professional Singer/Composer/Producer and Director/Editor/Colorist
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 8:43 pm

As i see from replies at this and other forums it seems there are too many different people in BM support who all time answered different things for same questions. Some people from support are more involved in technical problems some less.
The main problem that there is still no any clear info from BM about WHEN and HOW this problem will be fixed. Nobody can't explain will it be fixed in next camera firmware and Davinci Resolve update, or is it a hardware problem that can be fixed only by sending back camera for recalibration.
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

James Alexander Barnett

  • Posts: 193
  • Joined: Sat Jan 09, 2016 5:48 pm
  • Location: UK

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 9:59 pm

It was a disappointing response yeah I had a few back and forths with them today but they really aren't giving the game away as to what's the problem.

I'm not fussed about returning my camera as I honestly believe it won't make a difference, if the cameras are debayering the image In a way that causes these scaling problems they will all be the same.

It's obvious to see that the problem isn't baked into the footage, my only major concern is I can't control the way my clips or films are viewed by the end user, which is worrying.
James Alexander Barnett
www.jamesalexanderdp.co.uk
URSA Mini 4.6k
Offline
User avatar

Valentin Remy

  • Posts: 251
  • Joined: Thu Mar 24, 2016 3:52 pm
  • Location: Belgium

Re: Cross Hatching Wont Go Away

PostMon Oct 24, 2016 10:38 pm

James Barnett wrote:It's obvious to see that the problem isn't baked into the footage


It is :/ It's all over my ProRes files.
http://www.instagram.com/valentinremy.be/
Offline
User avatar

James Alexander Barnett

  • Posts: 193
  • Joined: Sat Jan 09, 2016 5:48 pm
  • Location: UK

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 10:59 am

Valentin Remy wrote:
James Barnett wrote:It's obvious to see that the problem isn't baked into the footage


It is :/ It's all over my ProRes files.



I take my last comment back, I shot again last night at a gig, I shot most of it in 2k however for the main act I thought I would up the res to 4.6k and the results are horrible, really nasty patterns over all my footage.

All clips shot:

ProRes HQ
1600 ISO
3200K
270 Shutter
(Video Mode)

I have exported out 2 PNG's one at 4.6k & one at 2k, the 2k one, although low light I think looks pretty good, the 4.6k one on the other hand is horrible, even viewed at 1:1, if you blow it up as people have been suggesting its even worse.

This can't be normal surely??? If someone wouldn't mind downloading and taking a look I would greatly appreciate it.

2k Image
https://dl.dropboxusercontent.com/u/4047574/2k.png

4.6k Image
https://dl.dropboxusercontent.com/u/4047574/4.6k.png
James Alexander Barnett
www.jamesalexanderdp.co.uk
URSA Mini 4.6k
Offline
User avatar

Valentin Remy

  • Posts: 251
  • Joined: Thu Mar 24, 2016 3:52 pm
  • Location: Belgium

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 2:21 pm

It's not new ^^ We already know that footage downsampled in-camera look better than full res.

Try to shoot 2k again, but in window mode. It will be terrible again.
http://www.instagram.com/valentinremy.be/
Offline
User avatar

rick.lang

  • Posts: 11663
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Cross Hatching Wont Go Away

PostTue Oct 25, 2016 5:41 pm

James, looking at both the 2K and 4.6K image on my iPad Air Retina screen and blowing it up and I see zero instance of cross hatching or screen door effect. Not trying to upset you, but I don't know what you are talking about as the grain or noise or distribution of texture is random and not at all like it was shot through a screen door.

I have seen images where I can dial in various sizes of cross hatch pattern by varying how I view the image, but nothing amiss in yours.

Sent from my iPad using Tapatalk
Rick Lang
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 5:46 pm

Screen-Shot-2016-10-25-at-8.45.09-PM.jpg
Screen-Shot-2016-10-25-at-8.45.09-PM.jpg (426.73 KiB) Viewed 13093 times

Screen-Shot-2016-10-25-at-8.53.05-PM.jpg
Screen-Shot-2016-10-25-at-8.53.05-PM.jpg (343.94 KiB) Viewed 13091 times
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline
User avatar

rick.lang

  • Posts: 11663
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 6:42 pm

Just to confirm, Dmitry, on my iPad, I see horrid cross hatch on the images you posted, but none on James' images.


Sent from my iPad using Tapatalk
Rick Lang
Offline

David Hessel

  • Posts: 274
  • Joined: Mon Mar 10, 2014 7:53 pm

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 6:56 pm

Here is a 400% blow up, you can see the source of the issue. This tiny pattern can be greatly amplified by scaling depending on the actual amount and by the quality of the scaling method used. Many players will use simplier methods for scaling to maintain smooth playback so some handle it better than others. Regardless the problem lies in this cross hatch pattern baked into the footage.

4.6k.jpg
Crosshatch
4.6k.jpg (222.43 KiB) Viewed 13073 times
David Hessel
Offline

Earl R. Thurston

  • Posts: 152
  • Joined: Sun Sep 04, 2016 4:52 pm
  • Location: Burnaby, BC

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 10:08 pm

Here's another comparison, with individual pixels from each image enlarged 500%:

4.6k-vs-2k.png
4.6k-vs-2k.png (25.47 KiB) Viewed 13043 times


As expected, the 2k image is better because the camera has blended the uneven odd/even rows and columns together during interpolation. This is what the 4.6k image SHOULD look like -- the kind of noise one would expect, which looks quite normal and natural.

But the actual 4.6k image shows an image with a microscopic grid pattern on exact odd/even row and column boundaries, indicative of a debayering imbalance.

Again -- this is worth repeating -- evidence has already shown it can be easily and properly fixed in a DNG:

A006_09231522_C001_000007_anim.gif
A006_09231522_C001_000007_anim.gif (114.33 KiB) Viewed 13043 times


However, once it's in a ProRes file recorded by the camera, there's no way to adjust the balance properly. BMD would have to change something in the camera to fix that.

(The only other remedy is to sacrifice a bit of resolution, such as 0.5 pixel offsets, to blur the odd/even rows and columns together.)
Offline

Earl R. Thurston

  • Posts: 152
  • Joined: Sun Sep 04, 2016 4:52 pm
  • Location: Burnaby, BC

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 10:19 pm

rick.lang wrote:Just to confirm, Dmitry, on my iPad, I see horrid cross hatch on the images you posted, but none on James' images.

Dmitry's images show the secondary side effect (moiré) of scaling James' images. An error at the micro level becomes an error at the macro level.
Offline

Earl R. Thurston

  • Posts: 152
  • Joined: Sun Sep 04, 2016 4:52 pm
  • Location: Burnaby, BC

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 10:59 pm

Here's another test image. In this animation, I've taken the crops shown earlier and cropped them further to alternate between just the even rows/columns and the odd rows/columns for both sample images. (A blurred version of the samples appears at the bottom to highlight the average with noise removed.)

This demonstrates there's about a 3% average luminance difference between diagonal pixels on the 4.6k image:

4.6k-vs-2k-anim.gif
4.6k-vs-2k-anim.gif (41.45 KiB) Viewed 13061 times
Offline

Ryan Hamblin

  • Posts: 218
  • Joined: Sat Apr 18, 2015 10:04 pm
  • Location: LA/Nashville, TN

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 11:07 pm

Is this apparent on all cameras? I have only ran into it randomly like twice on my camera, and it seemed totally dependent upon scaling. Are you all seeing this consistently, as its only popped up on me in very random situations, and never on the final output. I have primarily used prores with the camera scaled down from the full sensor.
www.brainstem.tv
www.ryanhamblin.com
Offline
User avatar

PaulDelVecchio

  • Posts: 791
  • Joined: Wed Aug 22, 2012 5:33 am
  • Location: NY

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 11:29 pm

Ryan Hamblin wrote:Is this apparent on all cameras? I have only ran into it randomly like twice on my camera, and it seemed totally dependent upon scaling. Are you all seeing this consistently, as its only popped up on me in very random situations, and never on the final output. I have primarily used prores with the camera scaled down from the full sensor.



I've shot numerous things at 4.6k raw and when I take it through to completion and master at 4K, I don't see it. I did see it one or two time on random footage if I wasn't using a specific setting in Resolve but it hasn't really popped up after that. I haven't tried ProRes much but I should probably do a test to see if it's even an issue.
Paul Del Vecchio - Director/Producer
http://www.pauldv.net
http://www.youtube.com/user/pdelvecchio814
http://www.facebook.com/pauldv
http://instagram.com/pdelv
Twitter: @pauldv
Offline

Earl R. Thurston

  • Posts: 152
  • Joined: Sun Sep 04, 2016 4:52 pm
  • Location: Burnaby, BC

Re: Cross Hatching Wont Go Away

PostTue Oct 25, 2016 11:55 pm

Ryan Hamblin wrote:Is this apparent on all cameras?

I don't think anyone knows how widespread this is yet. Like all discussion forums, it all depends on who has reported it. The most prominent examples have been with the URSA Mini 4.6K's using more recent firmware. There have also been reports of less prominent patterns from a full size URSA and Micro Cinema Cameras. I myself got to test an URSA Mini 4K recently which did not exhibit the problem, but it had older firmware.

Ryan Hamblin wrote:I have only ran into it randomly like twice on my camera, and it seemed totally dependent upon scaling.

That's because the scaling moiré is a secondary side effect. Scaling simply reveals the problem more noticeably, or hides it.

To be honest, if you have seen it, chances are your camera is affected. We could confirm here if you were able to post some images (they would need to be 1:1 from the camera, such as a full 4.6K DNG or ProRes frame).
Offline
User avatar

rick.lang

  • Posts: 11663
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 1:46 am

David, I see it on your 400% blowup clearly, but when I expand James' on the iPad , it's not there.


Sent from my iPad using Tapatalk
Rick Lang
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 2:05 am

I find that Cross Hatching from 4.6K don't visible when i open image in Quicklook. Maybe it uses some kind of additional smoothing? It is possible that iPad uses same scaling method as Quicklook.
Also in Photoshop Cross Hatching is visible only with basic "Nearest Neighbor" resize method. It is strange because in VLC i use very quality "Lanczos" resize method but it don't smooths Cross Hatching effect at all.
Screen-Shot-2016-10-26-at-4.53.19-AM.jpg
Screen-Shot-2016-10-26-at-4.53.19-AM.jpg (261.18 KiB) Viewed 13006 times
Screen-Shot-2016-10-26-at-4.57.38-AM.jpg
Screen-Shot-2016-10-26-at-4.57.38-AM.jpg (109.32 KiB) Viewed 13006 times
Screen-Shot-2016-10-26-at-4.57.46-AM.jpg
Screen-Shot-2016-10-26-at-4.57.46-AM.jpg (79.27 KiB) Viewed 13006 times
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline

Ryan Hamblin

  • Posts: 218
  • Joined: Sat Apr 18, 2015 10:04 pm
  • Location: LA/Nashville, TN

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 3:47 am

PaulDelVecchio wrote:
Ryan Hamblin wrote:Is this apparent on all cameras? I have only ran into it randomly like twice on my camera, and it seemed totally dependent upon scaling. Are you all seeing this consistently, as its only popped up on me in very random situations, and never on the final output. I have primarily used prores with the camera scaled down from the full sensor.



I've shot numerous things at 4.6k raw and when I take it through to completion and master at 4K, I don't see it. I did see it one or two time on random footage if I wasn't using a specific setting in Resolve but it hasn't really popped up after that. I haven't tried ProRes much but I should probably do a test to see if it's even an issue.



I have only mastered in UHD with the camera so far and have not seen any issues with the final masters
www.brainstem.tv
www.ryanhamblin.com
Offline

David Hessel

  • Posts: 274
  • Joined: Mon Mar 10, 2014 7:53 pm

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 4:07 am

rick.lang wrote:David, I see it on your 400% blowup clearly, but when I expand James' on the iPad , it's not there.


Sent from my iPad using Tapatalk


I cannot speak for what the Ipad may be doing or why you cannot see it there. I used Nuke to pull that zoomed image, Nuke is a professional compositing program and the ability to zoom in on an image and see exactly what the rendered output will look like is absolutely critical.

Speaking from experience when you want to incorporate CG elements into a plate the edges of your alpha are very important, if you don't match them well with the footage the CG will appear to float over the image, not to mention roto and keying. You have to be able to zoom in and check your work and know that what you are seeing is true otherwise you end up with some nasty surprises later on. Nuke shows this pattern, along with after effects and Photoshop, so it IS there.

Besides scaling in programs has been around for how many years? Of the countless times people have scaled imagery why now with this camera is scaling suddenly and issue. If you have access to Photoshop I suggest you open the image in there instead of using an ipad.
David Hessel
Offline

Earl R. Thurston

  • Posts: 152
  • Joined: Sun Sep 04, 2016 4:52 pm
  • Location: Burnaby, BC

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 4:39 am

Folks, there are many reasons why one app or another won't reveal the problem, but no program worth its salt will create a moiré pattern out of thin air, especially when we're talking about smooth, detail-free areas of color like we're seeing in these images.
Offline
User avatar

timbutt2

  • Posts: 1297
  • Joined: Tue Nov 12, 2013 10:32 am

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 4:41 am

First time I saw cross hatching was in the 2.5K after the firmware update that enabled crop guides in March of 2015. I initially only saw it on the 2.5K LCD, but it didn't show up in the footage. I downgraded the Firmware and then re-upgraded it and it went away.

With the 4.6K I only see it in either Resolve or Premiere, where it is a scaling issue that doesn't go over to exported footage. On the LCD I sometimes see it when I engage false color and double tap zoom while in false color. That depends on the resolution settings. However, the cross hatching doesn't translate to any of my footage. It's really a scaling issue from my experience. All my tests have found that I can get it to show up in either Resolve or Premiere. In Premiere I make it disappear easily by turning on High Quality on.

So in summary it's a scaling issue with the displays, but not the footage is my experience.


Sent from my iPhone using Tapatalk
"I'm well trained in the art of turning **** to gold." - Tim Buttner (timbutt2)

Cameras: URSA Mini Pro G2, URSA Mini 4.6K EF & Cinema Camera 2.5K EF
Computers: iMac 5K (Late 2015) & MacBook Pro Retina 15.4in (Mid 2018)
Offline

roger.magnusson

  • Posts: 915
  • Joined: Wed Sep 23, 2015 4:58 pm
  • Location: Sweden

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 10:46 am

Many seem to be ignoring the fact that this is a well known artefact with Bayer pattern sensors that have a green channel imbalance. It's not some elusive magical scaling issue that creates pattern detail out of thin air. As several people have pointed out, the scaling issue is just a consequence of the pattern embedded in the image. If the worst examples shown here are accepted as the norm then we might as well call those particular 4.6K cameras HD cameras, because that's the downsampling needed for the issue to completely disappear.

Sorry for the rant, it's just so frustrating when the issue is so clear and ignored at the same time.
Offline
User avatar

Frank Glencairn

  • Posts: 1651
  • Joined: Wed Aug 22, 2012 7:07 am
  • Location: Germany

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 11:50 am

Dmitry Shijan wrote:If someone interested here are few grey card DNGs frames from BMMCC. Even without noise reduction Cross Hatching pattern is visible there very nice at 200-300% zooming in and at custom ratio zooming out.


Downloaded your files - can't see anything crosshatchy in Resolve (latest version) at any magnification.

Not in the Resolve preview monitor, and not on my video preview monitor.

Same goes for James' PNGs. Looks just fine in my browser and image viewer.
http://frankglencairn.wordpress.com/

I told you so :-)
Offline

David Hessel

  • Posts: 274
  • Joined: Mon Mar 10, 2014 7:53 pm

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 1:35 pm

timbutt2 wrote:First time I saw cross hatching was in the 2.5K after the firmware update that enabled crop guides in March of 2015. I initially only saw it on the 2.5K LCD, but it didn't show up in the footage. I downgraded the Firmware and then re-upgraded it and it went away.

With the 4.6K I only see it in either Resolve or Premiere, where it is a scaling issue that doesn't go over to exported footage. On the LCD I sometimes see it when I engage false color and double tap zoom while in false color. That depends on the resolution settings. However, the cross hatching doesn't translate to any of my footage. It's really a scaling issue from my experience. All my tests have found that I can get it to show up in either Resolve or Premiere. In Premiere I make it disappear easily by turning on High Quality on.

So in summary it's a scaling issue with the displays, but not the footage is my experience.


Sent from my iPhone using Tapatalk


Does your experience include downloading the examples posted in this thread or are you just basing you opinion on your own camera?
David Hessel
Offline
User avatar

Dmitry Shijan

  • Posts: 816
  • Joined: Wed Sep 17, 2014 7:15 pm
  • Location: UA

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 2:01 pm

Frank Glencairn wrote:
Dmitry Shijan wrote:If someone interested here are few grey card DNGs frames from BMMCC. Even without noise reduction Cross Hatching pattern is visible there very nice at 200-300% zooming in and at custom ratio zooming out.


Downloaded your files - can't see anything crosshatchy in Resolve (latest version) at any magnification.

Not in the Resolve preview monitor, and not on my video preview monitor.

Same goes for James' PNGs. Looks just fine in my browser and image viewer.


Thats strange. Try to add to DNGs some chroma noise reduction, some contrast, and 0.47 or 0.48 sharpen. That makes pattern more visible.
All my custom made accessories for BMMCC/BMMSC now available here https://lavky.com/radioproektor/
Offline

David Hessel

  • Posts: 274
  • Joined: Mon Mar 10, 2014 7:53 pm

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 3:31 pm

Dmitry Shijan wrote:
Frank Glencairn wrote:
Dmitry Shijan wrote:If someone interested here are few grey card DNGs frames from BMMCC. Even without noise reduction Cross Hatching pattern is visible there very nice at 200-300% zooming in and at custom ratio zooming out.


Downloaded your files - can't see anything crosshatchy in Resolve (latest version) at any magnification.

Not in the Resolve preview monitor, and not on my video preview monitor.

Same goes for James' PNGs. Looks just fine in my browser and image viewer.


Thats strange. Try to add to DNGs some chroma noise reduction, some contrast, and 0.47 or 0.48 sharpen. That makes pattern more visible.


I see it in your DNG's in resolve, it is not as severe as what I have seen from the 4.6k but it is there. I can also see it in adobe.
David Hessel
Offline
User avatar

timbutt2

  • Posts: 1297
  • Joined: Tue Nov 12, 2013 10:32 am

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 4:00 pm

David Hessel wrote:
timbutt2 wrote:First time I saw cross hatching was in the 2.5K after the firmware update that enabled crop guides in March of 2015. I initially only saw it on the 2.5K LCD, but it didn't show up in the footage. I downgraded the Firmware and then re-upgraded it and it went away.

With the 4.6K I only see it in either Resolve or Premiere, where it is a scaling issue that doesn't go over to exported footage. On the LCD I sometimes see it when I engage false color and double tap zoom while in false color. That depends on the resolution settings. However, the cross hatching doesn't translate to any of my footage. It's really a scaling issue from my experience. All my tests have found that I can get it to show up in either Resolve or Premiere. In Premiere I make it disappear easily by turning on High Quality on.

So in summary it's a scaling issue with the displays, but not the footage is my experience.


Sent from my iPhone using Tapatalk


Does your experience include downloading the examples posted in this thread or are you just basing you opinion on your own camera?

Haven't downloaded any footage examples on this thread. I'm basing my opinion on the camera's I've used and how I've witnessed the cross hatching phenomenon. I wanted to share it in case it helps others.


Sent from my iPad using Tapatalk
"I'm well trained in the art of turning **** to gold." - Tim Buttner (timbutt2)

Cameras: URSA Mini Pro G2, URSA Mini 4.6K EF & Cinema Camera 2.5K EF
Computers: iMac 5K (Late 2015) & MacBook Pro Retina 15.4in (Mid 2018)
Offline

Phillip Bergman

  • Posts: 148
  • Joined: Mon Nov 02, 2015 6:36 pm
  • Location: Fresno, CA

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 4:38 pm

roger.magnusson wrote:Many seem to be ignoring the fact that this is a well known artefact with Bayer pattern sensors that have a green channel imbalance. It's not some elusive magical scaling issue that creates pattern detail out of thin air. As several people have pointed out, the scaling issue is just a consequence of the pattern embedded in the image. If the worst examples shown here are accepted as the norm then we might as well call those particular 4.6K cameras HD cameras, because that's the downsampling needed for the issue to completely disappear.

Sorry for the rant, it's just so frustrating when the issue is so clear and ignored at the same time.



Yea that's what I've been realizing. I basically have a really expensive HD camera. Unless I shoot in a place that has absolutely no shadows, anything I shoot UHD and above exhibits crosshatching. And the only way to completely get rid of it in a rendered out file is to either edit it on a 1080 timeline and then upscale to UHd upon render (which loses detail), or downscale the render to 1080.
Offline

David Hessel

  • Posts: 274
  • Joined: Mon Mar 10, 2014 7:53 pm

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 4:55 pm

timbutt2 wrote:
David Hessel wrote:
timbutt2 wrote:First time I saw cross hatching was in the 2.5K after the firmware update that enabled crop guides in March of 2015. I initially only saw it on the 2.5K LCD, but it didn't show up in the footage. I downgraded the Firmware and then re-upgraded it and it went away.

With the 4.6K I only see it in either Resolve or Premiere, where it is a scaling issue that doesn't go over to exported footage. On the LCD I sometimes see it when I engage false color and double tap zoom while in false color. That depends on the resolution settings. However, the cross hatching doesn't translate to any of my footage. It's really a scaling issue from my experience. All my tests have found that I can get it to show up in either Resolve or Premiere. In Premiere I make it disappear easily by turning on High Quality on.

So in summary it's a scaling issue with the displays, but not the footage is my experience.


Sent from my iPhone using Tapatalk


Does your experience include downloading the examples posted in this thread or are you just basing you opinion on your own camera?

Haven't downloaded any footage examples on this thread. I'm basing my opinion on the camera's I've used and how I've witnessed the cross hatching phenomenon. I wanted to share it in case it helps others.


Sent from my iPad using Tapatalk


Maybe you should it might change your opinion of this issue. There are pleny of examples and information in this thread that shows this is not a scaling issue. Not only can it easily be seen baked into the footage but we have identified the source of it in the DNG which is a green channel inbalance due to alternating light and dark rows of pixels. Further we have seen that debayering the DNG's exibiting the crosshatch using setting designed to fix a green channel inbalance does indeed fix the issue.

While you are at it why not post a DNG or short clip from your camera? Since you have also witnessed the issue I wouldn't be surprised if your footage has it too, it might not be as glaringly obvious as others are. What is your final output resolution, 4.6k, UHD or HD?
David Hessel
Offline
User avatar

rick.lang

  • Posts: 11663
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Cross Hatching Wont Go Away

PostWed Oct 26, 2016 5:02 pm

Thanks, Frank. That's my experience. I can't explain why some have a problem and some do not with the same footage. I did see the screen door effect in Resolve in the early Spring, although it never affected my deliverables. It doesn't seem to be an issue for me now and for the last few months. I'm going to bow out of this discussion and leave it with BMD. It's not my goal to see this, to examine footage at 400% or to make it show up by tweaking settings in Resolve. I thought my observations viewing James' footage would be of interest since it was different than James' view of his footage. That's all except of course I hope a resolution is found quickly.


Sent from my iPad using Tapatalk
Rick Lang
PreviousNext

Return to Cinematography

Who is online

Users browsing this forum: albertenix, Brandon Fowler, Florent Piovesan, MSN [Bot] and 47 guests