New Ursa 12k sensor tech

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

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostSat Nov 14, 2020 7:13 pm

Mark Wyatt wrote:I'm glad Slashcam has posted this, as this is exactly what I am finding as well regarding moire (ie- there isn't any if dealt with properly in post). Too bad "the youtubers" didn't fully test their process.


Yeah it’s a shame, because I see this constant reference to the 12K having moire problems and being soft.

It’s what the internet now thinks is true.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Mark Wyatt

  • Posts: 196
  • Joined: Wed Sep 25, 2013 4:36 pm
  • Location: Vancouver

Re: New Ursa 12k sensor tech

PostSat Nov 14, 2020 7:16 pm

rick.lang wrote:Mark, are you typically recording in 12K or something else?


Hi Rick. Almost always in 8k (as I like the better readout speeds), and delivering in 4k.
www.wyattvisuals.com
Offline

John Paines

  • Posts: 3989
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: New Ursa 12k sensor tech

PostSat Nov 14, 2020 7:22 pm

rick.lang wrote:I’m concluding 4K in camera with full field of view is an excellent option and what I’d likely shoot on the UMP12K.


They appear to report that full sensor 4K yields the same as 12K down-sampled to 4K in Resolve (though, again, you have to wonder if this isn't more internet "information").

But, assuming they know what they're doing, this is unlike full sensor Prores on the BMPCC 4K, where you get better results in Resolve with 4K on an HD timeline, than in-camera 4K to HD. So, in this case, there's no reason to shoot 12K, even if you had the media for it.
Offline
User avatar

rick.lang

  • Posts: 13981
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

New Ursa 12k sensor tech

PostSat Nov 14, 2020 7:42 pm

Thanks, Mark.

I agree John it’s an interesting conclusion that bodes well for UMP12K shooters who call follow their preference and use either the camera of Resolve 17 Camera raw to get very similar results.

Also agree we’re better off recording 4K in the BMPCC4K rather downscaling to 2K/HD in camera. I’ve started to keep large timelines on the Mac Pro and downscale for Deliver.
Rick Lang
Offline

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostSat Nov 14, 2020 8:19 pm

John Paines wrote:
But, assuming they know what they're doing, this is unlike full sensor Prores on the BMPCC 4K, where you get better results in Resolve with 4K on an HD timeline, than in-camera 4K to HD. So, in this case, there's no reason to shoot 12K, even if you had the media for it.


Shot stabilising or shot re-framing contingency is one reason why you’d do it.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

rick.lang

  • Posts: 13981
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: New Ursa 12k sensor tech

PostSat Nov 14, 2020 8:27 pm

So true. I’m going to revisit your 12K footage in Resolve 17b to take a look at Camera raw in this light.
Rick Lang
Offline

CaptainHook

Blackmagic Design

  • Posts: 1765
  • Joined: Wed Aug 22, 2012 4:50 am
  • Location: Melbourne, Australia
  • Real Name: Hook

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 12:35 am

rick.lang wrote:The article mentions 7.8ms for full sensor 8K/4K but that’s the value for 6K crop sensor as I understand it. 8.5ms is still great.

URSA Mini Pro 12K:
Full 12K : ~15.5ms
12K 2.4:1 : ~12.3ms
8K/4K (full field of view) : ~8.5ms*
8K 2.4:1 (full width) : ~6.7ms
6K crop : ~7.8ms
4K crop : ~4.25ms*

We improved the speeds in an update. The 8.5ms you quote above actually has a * next to it to indicate we had hoped to improve it in an update, which we did:

12k 17:9 -> 15.55ms
12k 2.4:1 -> 12.27ms
8k 17:9 (full FOV) -> 7.78ms
4k 17:9 (full FOV) -> 7.78ms
6k Super16 -> 7.78ms
8k 2.4:1 (full width) -> 6.13ms
4k 2.4:1 (full width) -> 6.13ms
4k Super16 -> 3.89ms
**Any post by me prior to Aug 2014 was before i started working for Blackmagic**
Offline
User avatar

rick.lang

  • Posts: 13981
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

New Ursa 12k sensor tech

PostSun Nov 15, 2020 12:36 am

Thank you so much, Captain!

That 4K Super16 crop is very impressive too.
Rick Lang
Offline

John Paines

  • Posts: 3989
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 12:47 am

John Brawley wrote:Shot stabilising or shot re-framing contingency is one reason why you’d do it.

JB


I know, but at what cost.... I guess it depends on the use, and what compression ratios for 12K versus 4K and how it all compares? It's not a worry I'll have any time soon, but it'll be interesting to see how people end up using it.
Offline

newbie

  • Posts: 14
  • Joined: Fri Nov 06, 2020 5:23 am
  • Real Name: Dan Lane

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 1:59 am

I just shot some interviews for a client this week on the 12k and sigma art primes. Mastered at 4k and delivered 1080p.

Essentially a same day delivery so I didn't have much time to color. But it ended up looking pretty good.

Note - you can not trust the viewports in Resolve. It won't match after you render. The colors will be different and the footage does appear sharper in the output.

So you should never critically analyze BRAW in Resolve. Always render it first..
Offline

Wayne Steven

  • Posts: 3362
  • Joined: Thu Aug 01, 2013 3:58 am
  • Location: Earth
  • Warnings: 2

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 3:36 pm

John Brawley wrote:
John Paines wrote:
But, assuming they know what they're doing, this is unlike full sensor Prores on the BMPCC 4K, where you get better results in Resolve with 4K on an HD timeline, than in-camera 4K to HD. So, in this case, there's no reason to shoot 12K, even if you had the media for it.


Shot stabilising or shot re-framing contingency is one reason why you’d do it.

JB


Yes, just like I used to say? :D
aIf you are not truthfully progressive, maybe you shouldn't say anything
bTruthful side topics in-line with or related to, the discussion accepted
cOften people deceive themselves so much they do not understand, even when the truth is explained to them
Offline

Sean van Berlo

  • Posts: 279
  • Joined: Wed Jul 01, 2015 6:33 am
  • Location: The Netherlands

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 4:23 pm

newbie wrote:I just shot some interviews for a client this week on the 12k and sigma art primes. Mastered at 4k and delivered 1080p.

Essentially a same day delivery so I didn't have much time to color. But it ended up looking pretty good.

Note - you can not trust the viewports in Resolve. It won't match after you render. The colors will be different and the footage does appear sharper in the output.

So you should never critically analyze BRAW in Resolve. Always render it first..


What is a viewport? My viewer matches my output on resolve.
Offline

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 4:46 pm

This is kind of the perfect answer to those early 12k reviewers and the recent direction you tube reviewers are heading.



JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

John Paines

  • Posts: 3989
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 5:11 pm

Well jeez, give the youtubers some credit. They know their audience. If you have no money, no production design, no lighting, no grasp of cinema, cinematography or post-production, rest assured, your chances of failure with the movie you'll probably never make are no worse with a $1 iphone than a 12K. Or 5219. Or an Alexa 65.

OTOH, you have to wonder how many people prepared to take these reviews seriously are actually in the market for a $10K camera. My guess is, zero? Which makes these videos what? Entertainment? Consumer p***?

I say, we need "Marques" to review the 12k, and settle all questions once and for all! I mean, he so totally interviewed Bill Gates!
Offline

newbie

  • Posts: 14
  • Joined: Fri Nov 06, 2020 5:23 am
  • Real Name: Dan Lane

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 5:32 pm

Sean van Berlo wrote:What is a viewport? My viewer matches my output on resolve.
[/quote]

Blackmagic calls it a viewer. Other software calls it a viewport. And others call it the scene.

You can easily see the colors don't match in renders. The ones I'm attaching are rendered to TIFF which is a common frame format with clients I work with.

You can also see the renders are sharper than what you observe in Resolve. These are on a UHD timeline but I would expect the difference to also be significant in 12k.

Screen Shot 2020-11-15 at 9.15.56 AM.png
Screen Shot 2020-11-15 at 9.15.56 AM.png (889.43 KiB) Viewed 1030 times

Left is Resolve right is the render.

Also attached a sharpness test render.

On the other hand - I don't notice as drastic shifts when working with R3D files. Only BRAW appears to be off in Resolve.

These are not lab tests. Just my observations.
Attachments
Screen Shot 2020-11-15 at 9.28.09 AM.png
Screen Shot 2020-11-15 at 9.28.09 AM.png (842.73 KiB) Viewed 1030 times
Offline

John Paines

  • Posts: 3989
  • Joined: Tue Jul 28, 2015 4:04 pm

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 5:39 pm

You need to read the section in the manual which covers external monitoring. You can't judge image quality, resolution or color corrections in the Resolve source or program viewer. These viewers are for reference only, to show content, overlays, power windows, etc.
Offline

newbie

  • Posts: 14
  • Joined: Fri Nov 06, 2020 5:23 am
  • Real Name: Dan Lane

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 5:52 pm

Agreed.

Have to admit that is a bit ridiculous though. You're coloring and making creative decisions in a tool which you can't trust to respect them in the output. That's fine if you're just playing around. But clients (producers, editors, agencies, etc) can be very particular about what they want.
Offline

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 5:56 pm

newbie wrote:Agreed.

Have to admit that is a bit ridiculous though. You're coloring and making creative decisions in a tool which you can't trust to respect them in the output. That's fine if you're just playing around. But clients (producers, editors, agencies, etc) can be very particular about what they want.


Most other systems do similar things. They’re designed to be used with external monitors, external scopes.

Any Apple product playing back ProRes natively (like in FCP or just a viewer) has a sharper stop frame than when they’re in motion.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 6:06 pm

John Paines wrote:
OTOH, you have to wonder how many people prepared to take these reviews seriously are actually in the market for a $10K camera. My guess is, zero? Which makes these videos what? Entertainment? Consumer p***?


I dunno John....

I’d say the wrong information in those initial videos has hurt the 12K. I still see people posting about a chronic moire problem (Seems it’s affected by post processes as I mentioned and now Slashcam) and that camera is soft (badly executed test).

It’s hard to walk back those perceptions. But they’re persisting on facebook and other camera forums like here. I’ve had a lot of people message me privately bringing up the same concerns in recent weeks...

There was a perception that Blackmagic cameras used to overheat as well. They just don’t. But It’s still an idea that’s out there.

I think time will tell if the camera is going to be a success, but I think it’s already been very well received. I know that orders are increasing and they’re struggling to keep up with production. I know of several very high profile scenarios that the camera has already been embraced in and 6 or 12 months from now you’ll be watching that content. That’s the real test...

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Wayne Steven

  • Posts: 3362
  • Joined: Thu Aug 01, 2013 3:58 am
  • Location: Earth
  • Warnings: 2

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 7:47 pm

newbie wrote:Agreed.

Have to admit that is a bit ridiculous though. You're coloring and making creative decisions in a tool which you can't trust to respect them in the output. That's fine if you're just playing around. But clients (producers, editors, agencies, etc) can be very particular about what they want.


That's still going on.

What we need is a company that can provide reasonable 16 bit HDR, near rec2020 screens (as the Chinese can do). That you can use as work and reference screen. The cheapest HDR option for the moment, is one of the dual layer schemes, two 10 bit panels or so The way to do it, is to approach a TV company about doing a special edition at a TV price, with your labels on it (BM, Resolve DaVinci etc). A few Chinese conglomerates have been chaffing at the but, to join the big guys for years, putting effort into screen quality uniformity. Even if they can't get there, at 16 bits, you could probably calibrate it accurate enough. That's the beauty of it, even if it costs ten times more to get uniformity to 16 bits, the extra precision allows them to dusk out that stuff as long as it's ok enough. The fact that some parts of the screen might be 14 bits etc, is ok for a lower cost solution where you can't even get 10 bits right normally. The benefit fur the Chinese, is a prestige claim, at a prestigious price, without having to offer at a lower price. So, Dolby vision on Blu-ray 4k is the target 4000 brightness, as that is a bit of upper end use scenario (the blacks would look grey at that brightness, but most used would be closer to under 1000. One day in the future, one can switch over to a better technology, when cheaply available So, the sets are sold as normal premium TV's in order to bring down the price. A balsy plan, but Red never pulled it off.
aIf you are not truthfully progressive, maybe you shouldn't say anything
bTruthful side topics in-line with or related to, the discussion accepted
cOften people deceive themselves so much they do not understand, even when the truth is explained to them
Offline

Wayne Steven

  • Posts: 3362
  • Joined: Thu Aug 01, 2013 3:58 am
  • Location: Earth
  • Warnings: 2

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 7:56 pm

John Brawley wrote:This is kind of the perfect answer to those early 12k reviewers and the recent direction you tube reviewers are heading.



JB


You just can't tell how much drugs are involved in YouTube videos. :). I mean, the video front still makes the guy look like a manaquine, made to look a bit like a sasquatch. I even hate the parodies in YouTube.
aIf you are not truthfully progressive, maybe you shouldn't say anything
bTruthful side topics in-line with or related to, the discussion accepted
cOften people deceive themselves so much they do not understand, even when the truth is explained to them
Offline

Filmobsession

  • Posts: 30
  • Joined: Thu Feb 28, 2019 5:25 pm
  • Warnings: 2
  • Real Name: Daniel Black

Re: New Ursa 12k sensor tech

PostSun Nov 15, 2020 11:57 pm

Too many FAKE personalities on YouTube making videos for views/cash. The worst is when they laugh/chuckle during a formal sentence. Everything seems like joke.

Thank you for sharing that video J Brawley.

Film heritage. Young filmmakers need to be educated about the language and history of film.
Offline

Wayne Steven

  • Posts: 3362
  • Joined: Thu Aug 01, 2013 3:58 am
  • Location: Earth
  • Warnings: 2

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 3:10 am

Just vote them down. I got so sick of seeing utter rubbish being pushed on me on my streaming service, I just down vote now. If everybody down votes, then it's not going come up in the suggestion list as much, just to the people who like that stuff. Simple.


One thing that produces crap like that on fee streaming services, is the little financial return per view. It is really like one of those, fleeting momentary fame lowest common denominator travesty subjected productions from the dark days of the industry in the 1980's/1970's.

YouTube and the ad networks have destroyed the value of advertising by over supplying per page or content forcing the price given, because the advertising budgets are divided over more impressions, rather then be controlled, producing a negative attention feed back in the viewers. The highest -10% financial split to content suppliers should be also given, with less advertising. 1 or 2 seconds per minute. Under the old system, content impressions were controlled and space limited, giving adds more value and over less impressions. The solution to our video problem here.
Last edited by Wayne Steven on Mon Nov 16, 2020 2:12 pm, edited 2 times in total.
aIf you are not truthfully progressive, maybe you shouldn't say anything
bTruthful side topics in-line with or related to, the discussion accepted
cOften people deceive themselves so much they do not understand, even when the truth is explained to them
Offline

CaptainHook

Blackmagic Design

  • Posts: 1765
  • Joined: Wed Aug 22, 2012 4:50 am
  • Location: Melbourne, Australia
  • Real Name: Hook

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 3:18 am

Please keep the discussion on topic.
**Any post by me prior to Aug 2014 was before i started working for Blackmagic**
Offline

Filmobsession

  • Posts: 30
  • Joined: Thu Feb 28, 2019 5:25 pm
  • Warnings: 2
  • Real Name: Daniel Black

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 8:23 am

Morning Hook,

The problem is, like what Brawley suggested, some ‘Manic Laughing YouTube Filmmakers’ are trashing the 12K sensor with INACCURATE information.

I believe that spoof video is quite fitting.

I agree with you, Wayne.

Anyway, keep striving and loving the craft, Folks & BMD!
Offline

CaptainHook

Blackmagic Design

  • Posts: 1765
  • Joined: Wed Aug 22, 2012 4:50 am
  • Location: Melbourne, Australia
  • Real Name: Hook

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 1:16 pm

The spoof video is fine. But we don't need multiple posts attacking Youtube content creators for their general approach or personalities. If you disagree/agree with a video related to the 12k and want to share your reasons that's fine, but keep it civil and on topic please.
**Any post by me prior to Aug 2014 was before i started working for Blackmagic**
Offline

Wayne Steven

  • Posts: 3362
  • Joined: Thu Aug 01, 2013 3:58 am
  • Location: Earth
  • Warnings: 2

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 2:09 pm

Thank you Daniel, for agreeing with my deleted post agreeing that innacurate YouTube videos and spoofs, certainly are a waste of time, and hence was that subtopic (and the exaggeratedly altered title image of that link). However, people who lie, and crazily make up things about others here, their posts can remain?
Last edited by Wayne Steven on Mon Nov 16, 2020 6:02 pm, edited 1 time in total.
aIf you are not truthfully progressive, maybe you shouldn't say anything
bTruthful side topics in-line with or related to, the discussion accepted
cOften people deceive themselves so much they do not understand, even when the truth is explained to them
Offline

Filmobsession

  • Posts: 30
  • Joined: Thu Feb 28, 2019 5:25 pm
  • Warnings: 2
  • Real Name: Daniel Black

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 3:24 pm

Hook,

I don’t have the energy and desire to attack. A spade is a spade. Call it what it is.

Hopefully, a young or old filmmaker will read these comments and believe he/She CAN create serious content because people like me back it!!!!

That’s my take, Hook...



Edit: Hook, you gave me a warning because of my above comment. I stand by what I say. No reason to be upset, Hook.
Last edited by Filmobsession on Wed Nov 18, 2020 5:19 pm, edited 4 times in total.
Offline

mlakner

  • Posts: 1
  • Joined: Sun Nov 15, 2020 3:44 am
  • Real Name: Max Lakner

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 7:38 pm

John Brawley wrote:
Mark Wyatt wrote:I'm glad Slashcam has posted this, as this is exactly what I am finding as well regarding moire (ie- there isn't any if dealt with properly in post). Too bad "the youtubers" didn't fully test their process.


Yeah it’s a shame, because I see this constant reference to the 12K having moire problems and being soft.

It’s what the internet now thinks is true.

JB


In your experience, is there no moire problem with the 12k, worse than other cameras? If so, why do you think those youtubers had that problem? - High compression rate in camera?

Thanks.
Offline
User avatar

Robert Niessner

  • Posts: 3228
  • Joined: Thu Feb 21, 2013 9:51 am
  • Location: Graz, Austria

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 11:25 pm

Eugene Belsky has released a nice video showcasing the UMP12k with a Zeiss Otus 55mm f/1.4
and several braw files to download:

Robert Niessner
LAUFBILDkommission
Graz / Austria
---------------------------------------------------------------------
Blackmagic Camera Blog (German):
http://laufbildkommission.wordpress.com

Read the blog in English via Google Translate:
http://tinyurl.com/pjf6a3m
Offline

Dan Cotreau

  • Posts: 18
  • Joined: Wed Aug 12, 2020 8:43 pm
  • Location: Los Angeles
  • Real Name: Dan Cotreau

Re: New Ursa 12k sensor tech

PostMon Nov 16, 2020 11:43 pm

Thanks for sharing Robert.
Offline

Ryan Earl

  • Posts: 270
  • Joined: Sun Jan 17, 2016 6:56 pm
  • Location: Philadelphia, USA

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 12:31 am

Robert Niessner wrote:Eugene Belsky has released a nice video showcasing the UMP12k with a Zeiss Otus 55mm f/1.4
and several braw files to download:


Yes, thanks for posting this here Robert. The closeups look really great, the Otus lenses with this sensor really pop and I'm actually into the 60 fps look.

Have you noticed moire in these clips? There is even a ball of yarn shot which would normally be tricky without something popping up.
Offline
User avatar

Robert Niessner

  • Posts: 3228
  • Joined: Thu Feb 21, 2013 9:51 am
  • Location: Graz, Austria

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 12:52 am

Ryan Earl wrote:Yes, thanks for posting this here Robert. The closeups look really great, the Otus lenses with this sensor really pop and I'm actually into the 60 fps look.

Have you noticed moire in these clips? There is even a ball of yarn shot which would normally be tricky without something popping up.


None so far. I am testing on my laptop which has a UHD display, but unfortunately only a Quadro with 4GB which doesn't allow me to switch to 12k demosaicing. 8k is the maximum.
The closeup of the woman is great - in a HD timeline I can zoom in 6x and still get a very sharp image with 8k debayer.
Robert Niessner
LAUFBILDkommission
Graz / Austria
---------------------------------------------------------------------
Blackmagic Camera Blog (German):
http://laufbildkommission.wordpress.com

Read the blog in English via Google Translate:
http://tinyurl.com/pjf6a3m
Offline

MishaEngel

  • Posts: 1413
  • Joined: Wed Aug 29, 2018 12:18 am
  • Real Name: Misha Engel

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 1:09 am

mlakner wrote:In your experience, is there no moire problem with the 12k, worse than other cameras? If so, why do you think those youtubers had that problem? - High compression rate in camera?

Thanks.


Just read the Slashcam.de articles.
https://www.slashcam.de/artikel/Test/Der-Sensor-der--Blackmagic-URSA-Mini-Pro-12K---Demosaicing-und-Rolling-Shutter.html
https://www.slashcam.de/artikel/Test/URSA-Mini-Pro-12K---Dynamik-und-Moir-es.html
When you don't understand German, just use a translator like Google translate.

Slashcam.de used the swiss army knife of lenses (sigma 18-35).

Eugene Belski used the otis 55 f1.4 which is a pretty good lens.

From LensRentals
Image

Offline

Ryan Earl

  • Posts: 270
  • Joined: Sun Jan 17, 2016 6:56 pm
  • Location: Philadelphia, USA

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 1:40 am

MishaEngel wrote:
mlakner wrote:In your experience, is there no moire problem with the 12k, worse than other cameras? If so, why do you think those youtubers had that problem? - High compression rate in camera?

Thanks.


Just read the Slashcam.de articles.
https://www.slashcam.de/artikel/Test/Der-Sensor-der--Blackmagic-URSA-Mini-Pro-12K---Demosaicing-und-Rolling-Shutter.html
https://www.slashcam.de/artikel/Test/URSA-Mini-Pro-12K---Dynamik-und-Moir-es.html
When you don't understand German, just use a translator like Google translate.

Slashcam.de used the swiss army knife of lenses (sigma 18-35).

Eugene Belski used the otis 55 f1.4 which is a pretty good lens.

From LensRentals
Image



I'm having to piece this together from Robert and a Google Translate page from German to English. . .

It would be helpful if Blackmagic published ARRI like White Papers. https://www.arri.com/en/learn-help/lear ... ite-papers

Things are broken down with workflow examples that become more official guidance. It's great what Hook and team did with color (and explantations on the forum) in Resolve for example.
Offline

Uli Plank

  • Posts: 8169
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 2:09 am

The Otus is as good as it gets. It's not a cine lens but optimized as a photographic lens for resolution and bokeh without compromises. Zeiss is really showing off that they still know what they are doing.
Resolve Studio 16.2.7 and Fusion Studio under MacOS Mojave 10.14.6
iMac 2017 Radeon Pro 580 8 GB VRAM and 32 GB RAM
2018 Mac mini 16 GB RAM plus eGFX Breakway Radeon RX 580
(currently my test system for DR 17 beta 2 under Catalina 10.15.7)
Offline

Wayne Steven

  • Posts: 3362
  • Joined: Thu Aug 01, 2013 3:58 am
  • Location: Earth
  • Warnings: 2

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 5:56 am

Robert Niessner wrote:Eugene Belsky has released a nice video showcasing the UMP12k with a Zeiss Otus 55mm f/1.4
and several braw files to download:



Still seeing the same problems with the face.
aIf you are not truthfully progressive, maybe you shouldn't say anything
bTruthful side topics in-line with or related to, the discussion accepted
cOften people deceive themselves so much they do not understand, even when the truth is explained to them
Offline

John Griffin

  • Posts: 851
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 3:58 pm

Viewing the clips now in resolve on a UHD timeline and 27" UHD screen 12" away - detail is amazing. Almost too much it's distracting and becomes a feature in itself. Color fidelity I'm not so sure about. It looks OK but not 'Wow' like the resolution. Skintone look fine once it's properly WB'd.
I'm viewing in REC709 in ACES using Gen5. Oh and no moiré issues on stuff that would normally cause it on a camera without an OLPF.
i7 6950x, 64gb RAM, GTX1080ti.
Offline
User avatar

rick.lang

  • Posts: 13981
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

New Ursa 12k sensor tech

PostTue Nov 17, 2020 5:14 pm

Glad colour looks better once graded properly. I’m still downloading and might make a short video to post on Vimeo. Looking forward to judging 12K, 8K, 6K, 4K, 2K with HDR 4000/1600/1000 nits and SDR 100 nits (displayed up to 1600 nits on the Apple Pro XDR 6K display). My HDTV can handle 2K SDR so I’ll just have to imagine the rest.
Rick Lang
Offline

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 5:58 pm

mlakner wrote:
In your experience, is there no moire problem with the 12k, worse than other cameras? If so, why do you think those youtubers had that problem? - High compression rate in camera?

Thanks.


In my view the 12K has FAR LESS of a moire problem than any previous Blackmagic camera.

I believe that the reason moire is coming up is because people are SCALING shots a lot more in non-native and unusual timeline resolutions and not decoding for the best possible outcomes. The very high resolution basically FORCES many users to choose lesser than camera native resolution timelines. Most of what I rendered was from an 8K timeline. Though of course I edited and graded at 4K. It takes a lot longer to render an 8K timeline than a 4K timeline so I think most aren’t bothering.

I think this is basically post workflow induced moire. I’ve shot a lot of material with this camera now and I just don’t see this problem coming up for me at all.

I don’t think compression choices we as end users make has too much to do with if moire comes up or not.

What makes a difference is the extra resolution (which reduces the chance of it happening) and the de-mosaic algorithm.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

John Griffin

  • Posts: 851
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 6:06 pm

12k moire.jpg
12k moire.jpg (138.48 KiB) Viewed 487 times

Pixel peeping at full res in 12k RAW / 8k timeline at 100% and yes there is some moiré but it's confined to small areas of the image with hard fine diagonal lines. To be fair any sensor without an OLPF is going to do this and lower res sensors much more. At 4k it's not going to be much of an issue.
i7 6950x, 64gb RAM, GTX1080ti.
Offline

Andrew Kolakowski

  • Posts: 6842
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 10:00 pm

John Brawley wrote:
I think this is basically post workflow induced moire. I’ve shot a lot of material with this camera now and I just don’t see this problem coming up for me at all.

I don’t think compression choices we as end users make has too much to do with if moire comes up or not.

What makes a difference is the extra resolution (which reduces the chance of it happening) and the de-mosaic algorithm.

JB


Scaling down will never introduce moire as far as I can tell, so there is no such a thing like post scale moire. It has to be there in first place. Any scaling down will actually cover moire.
If there is any problem then it's somewhere in Resolve which takes RAW data and processes it.

That paper about new sensor layout was very good and shown that it needs different debayering than Bayer sensors. No idea where BM is with this.
Offline

Hendrik Proosa

  • Posts: 1256
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Estonia

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 10:24 pm

Andrew Kolakowski wrote:Scaling down will never introduce moire as far as I can tell, so there is no such a thing like post scale moire. It has to be there in first place. Any scaling down will actually cover moire.
If there is any problem then it's somewhere in Resolve which takes RAW data and processes it.

Scaling down can introduce moire depending on what kind of sampling filter is used.
I do stuff.
Offline

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 10:26 pm

Andrew Kolakowski wrote:Scaling down will never introduce moire as far as I can tell,


It's easy to test. Take a shot of a zone plate or resolution chart and then drag the playback window on whatever system you use and resize it on the fly and watch the moire come and go independent of the camera used.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Andrew Kolakowski

  • Posts: 6842
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 10:30 pm

Yes, because too many apps use simple nearest neighbour scaling for preview window. This is done to save resources, but it's so misleading. You are seeing side effect of one the the simplest scaling methods, not real issues with your footage. Change it to any decent scaling and it should be gone.
Resolve final export should never touch nearest neighbour.
Resolve SDI/HDMI preview is scaled with basic scaling, so watch this as well. At best it's Bilinear method, which is fairly simple as well.
Offline

John Brawley

  • Posts: 2674
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: New Ursa 12k sensor tech

PostTue Nov 17, 2020 11:18 pm

Andrew Kolakowski wrote:Yes, because too many apps use simple nearest neighbour scaling for preview window. This is done to save resources, but it's so misleading. You are seeing side effect of one the the simplest scaling methods, not real issues with your footage.


Right. So you CAN induce moire with scaling. You categorically posted that's not the issue. I disagree.

That was the point.

The results from slash cam article when translated say

"The main disadvantage of this concept, however, is that the user also needs to understand it. Because the recorded resolution and timeline resolution do not match, the conversion can lead to artifacts such as moirées. We had already suspected this when the problem was discussed in our forum, now we can also prove it. Thanks to the brand new Resolve Beta 17. Because with this beta you can set the resolution with which you want to demosaicing in the RAW development."

And their theory as to why others are getting moire....

"Every camera has certain detail frequencies in which moirées appear. However, even with the switchover and a sweep pattern with different distances, we did not succeed in generating moirées as clear as a fewYoutuber recently presented on the net .

Since these Youtubers worked with Premiere Pro, our guess is that Premiere will scale "differently" here if the recording resolution is significantly above the timeline resolution. In the Blackmagic RAW settings, Premiere did not allow us to determine the decoding resolution ourselves - as is possible under Resolve. Maybe this will be possible with a future Blackmagic RAW plugin update in the future. However, it could also be that the plug-in API under Premiere does not even allow dynamic changes to the resolution of a clip."


Which also backs up what I was pitching. I can't see why some get terrible moire, far worse than what they should with a 12K camera (the greater the resolution, the less the camera should moire) when other's don't with this camera.

I posted a resolution chart with two different lenses right in the sweet spot of exposure (as far as diffraction) and it exhibits very little moire. That should show up the worst in that case and it just didn't.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Hendrik Proosa

  • Posts: 1256
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Estonia

Re: New Ursa 12k sensor tech

PostWed Nov 18, 2020 8:25 am

Only widely used filters that definitely don’t introduce potential moire are bilinear and gaussian. Rescaling with sharpening effect that Lanczos variants and others do, using negative lobes, can probably produce moire when they hit specific detail frequencies.
I do stuff.
Offline

John Griffin

  • Posts: 851
  • Joined: Sun Aug 23, 2015 4:18 pm

Re: New Ursa 12k sensor tech

PostWed Nov 18, 2020 8:42 am

I ran through various timeline, RAW resolution and rescaling types in R17 and nothing changed the behaviour of that small patch of moiré on the cars. Also if it's a display problem it would change with zooming in the window and it doesn't. I can't find moiré in any other part of that or any other clip despite them containing the kind of resolution and fine detail that would provoke it in any other camera without an OLPF so I think it's maybe a specific issue with slanted high contrast fine lines. Maybe it can be ironed out in firmware with a change to the in-camera debayering. From the evidence of playing with these clips it's clearly an astonishing camera in terms or rendering clean high res detail.
i7 6950x, 64gb RAM, GTX1080ti.
Offline

Andrew Kolakowski

  • Posts: 6842
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: New Ursa 12k sensor tech

PostWed Nov 18, 2020 12:06 pm

John Brawley wrote:
Andrew Kolakowski wrote:Yes, because too many apps use simple nearest neighbour scaling for preview window. This is done to save resources, but it's so misleading. You are seeing side effect of one the the simplest scaling methods, not real issues with your footage.


Right. So you CAN induce moire with scaling. You categorically posted that's not the issue. I disagree.

That was the point.

The results from slash cam article when translated say

"The main disadvantage of this concept, however, is that the user also needs to understand it. Because the recorded resolution and timeline resolution do not match, the conversion can lead to artifacts such as moirées. We had already suspected this when the problem was discussed in our forum, now we can also prove it. Thanks to the brand new Resolve Beta 17. Because with this beta you can set the resolution with which you want to demosaicing in the RAW development."

And their theory as to why others are getting moire....

"Every camera has certain detail frequencies in which moirées appear. However, even with the switchover and a sweep pattern with different distances, we did not succeed in generating moirées as clear as a fewYoutuber recently presented on the net .

Since these Youtubers worked with Premiere Pro, our guess is that Premiere will scale "differently" here if the recording resolution is significantly above the timeline resolution. In the Blackmagic RAW settings, Premiere did not allow us to determine the decoding resolution ourselves - as is possible under Resolve. Maybe this will be possible with a future Blackmagic RAW plugin update in the future. However, it could also be that the plug-in API under Premiere does not even allow dynamic changes to the resolution of a clip."


Which also backs up what I was pitching. I can't see why some get terrible moire, far worse than what they should with a 12K camera (the greater the resolution, the less the camera should moire) when other's don't with this camera.

I posted a resolution chart with two different lenses right in the sweet spot of exposure (as far as diffraction) and it exhibits very little moire. That should show up the worst in that case and it just didn't.

JB


Well exactly- Resolve engine can be not optimised properly for this new sensor, but this is BM role to sort it out.
User by exporting 4K downscaled version of 12K can't really add non-existing moire (unless Resolve internal engine does it processing RAW in bad way). It's not really users fault, but Resolve/Premiere.
Not exactly the same and not exactly related to "seeing moire on preview window".
Last edited by Andrew Kolakowski on Wed Nov 18, 2020 12:14 pm, edited 3 times in total.
Offline

Andrew Kolakowski

  • Posts: 6842
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: New Ursa 12k sensor tech

PostWed Nov 18, 2020 12:10 pm

John Griffin wrote:I ran through various timeline, RAW resolution and rescaling types in R17 and nothing changed the behaviour of that small patch of moiré on the cars. Also if it's a display problem it would change with zooming in the window and it doesn't. I can't find moiré in any other part of that or any other clip despite them containing the kind of resolution and fine detail that would provoke it in any other camera without an OLPF so I think it's maybe a specific issue with slanted high contrast fine lines. Maybe it can be ironed out in firmware with a change to the in-camera debayering. From the evidence of playing with these clips it's clearly an astonishing camera in terms or rendering clean high res detail.


Yes, it will be quite complex and you can have 50 recordings without and 51 will have it once hitting certain image pattern. Not a sensor expert, but looks like new 12K pattern design (color samples are even more apart from each other?) is more prone to it than Bayer. It definitely needs new approach to debayering, like this study paper shown.
PreviousNext

Return to Cinematography

Who is online

Users browsing this forum: Google Feedfetcher and 26 guests