Will Blackmagic ever fix the FPN on the 4.6K

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

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 07, 2017 5:59 pm

This is a legitimate question I think most of us want to know.
I'd like to start a discussion on the process of removing FPN. I've been doing tons of research but Blackmagic are very hush hush when it comes to how they process their images.
As FPN is an inherent CMOS problem it's hard to fix but Red, Arri, Sony and Panasonic have it figured out so what's taking Blackmagic so long to implement a fix and is there any glimmer of hope to a future solution?

Yeah the 4.6k can create beautiful images but there are a good 4 stops of decent lattitude going to waste. This camera would be truly perfect if it weren't for that FPN.

I'd love for a BM employee to give us some info on this infamous FPN issue. ;)
Offline

Dante Harbridge Robinson

  • Posts: 61
  • Joined: Mon Jun 30, 2014 9:41 pm
  • Location: London

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 07, 2017 7:36 pm

It must be pretty difficult to solve. Isn't it something to do with black shading?
Red solve it by you having to set off a black shade which takes up to 40 minutes, and you save it as a profile. You have to make different profiles for different sensor temperatures (which you have to select depending on your shooting conditions)
Arri does automatic black shading that I don't think you have to wait for, and internal cooling keeps the sensor at a constant temperature, so you never have to choose one.

As far as I know the URSA 4.6K has some kind of auto black shading. I wonder if that is optimised for a certain sensor temperature?

Maybe Sony and Panasonic have solved the FPN issue but they may have drastically different sensors, that are really to my eye at least inferior to the URSA 4.6K due to magenta cast across the whole image. If you try and remove magenta you will take it from the skin, and leave it looking dead. To me it seems the 4.6K and Alexa are unique in that the skin colours are fully separated and this cast is not present. Even Red's skin tones look dead by comparison and kind of look like there is a polarising filter removing the sheen from someone's face. When filming THE HOBBIT for example special makeup had to be used to compensate for the Red Dragon's colour inadequacies. (Maybe Red gave Peter Jackson those 60 Red cameras for free?) Conversely, Sony appears to favour an unnatural amount of 'sheen'. Alexa is spot on like reality and Blackmagic is close to that.

So really, even if other companies have solved it, it might be at the expense of some other qualities. Maybe the 4.6K sensor creates 2x as much heat because it is 2x as sensitive to skin colours than the Panasonic and Sony sensors, and gets fpn because the heat dissipation system is under more stress?

But yes I love to have no fpn too :)
Last edited by Dante Harbridge Robinson on Fri Jul 07, 2017 8:10 pm, edited 1 time in total.
Offline

thomas bruegger

  • Posts: 246
  • Joined: Thu Oct 18, 2012 5:43 pm
  • Location: switzerland

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 07, 2017 8:01 pm

+1 a statement from BMD would be very appretiated.
Thomas Bruegger / garage5 GmbH
www.garage5.ch
Offline
User avatar

rick.lang

  • Posts: 13843
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 07, 2017 9:58 pm

The Black Shading calibration on the camera isn't automatic but does help if done correctly for most people.


Sent from my iPhone using Tapatalk
Rick Lang
Offline

James Harkness

  • Posts: 287
  • Joined: Wed Jul 13, 2016 5:04 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 1:12 am

I use the 4K EF and dont have a problem with FPN. Not sure if its supposed to be better or worse than the 4.6k. It is something I have seen before, but I seem to be able to completely avoid it praise GOD. Light your scenes well, or atleast have a super fast lens (I use Sigma 30mm 1.4) and I think turning the camera on for a minute or two before you use it helps I might've read.
Offline

VicHarris

  • Posts: 236
  • Joined: Sat Aug 24, 2013 5:36 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 2:27 am

I have no problem using most of the stops on my URSA Mini Pro. I use false colors, light correctly and give the sensor as much of a high ass key as I can and the shadows are fine on my camera. 2 others cameras I use in LA are fine as well.
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 2:45 am

Jakelutz wrote:This is a legitimate question I think most of us want to know.


This is an issue that comes up a lot.

Yes, it's possible to see FPN on the camera, but it's _usually_ something that can be avoided or reduced depending on how you process your footage.

Here's a recent thread that kind of started the same as this one and then the OP ended up getting to this point.
viewtopic.php?f=2&t=57403&p=348229#p348193

You'll find some very low lights shots I did recently on multiple ursa bodies that look great and have very little in the way of FPN.

So what would be really useful is if you can post some original files as examples and how you're grading / processing and we can maybe see if there's something we can do to make it look better ?


Jakelutz wrote:I'd like to start a discussion on the process of removing FPN. I've been doing tons of research but Blackmagic are very hush hush when it comes to how they process their images.


What EXACTLY did you want to know ? I also presume you have done the "sensor calibration" function ?

Jakelutz wrote:As FPN is an inherent CMOS problem it's hard to fix but Red, Arri, Sony and Panasonic have it figured out so what's taking Blackmagic so long to implement a fix and is there any glimmer of hope to a future solution?


Well let's start here. An AC colleague of mine tells me he just did a series for Netflix on RED (not yet aired). He complained about a lot of problems with the cameras they had but one that came up for him was FPN. To calibrate on the camera it's a very lengthy calibration process and that you need to do it several times for many different potential operating temperatures. In other words if the ambient temperature range varied more than about 10 degrees C they had to have a different temperature calibration. You have 5 memories in the camera available and he found it wasn't enough because it ALSO was supposed to use the same settings when working off speed, but that function didn't seem to work. Their solution was to shoot HFR with a 360Deg shutter.

Does that sound like they have it figured out ? I'm not sure it does to me. More like managed.

Arri. Well they do indeed have it very well sorted but in the right conditions it also will show up there. https://www.arri.com/forum/viewtopic.php?t=267&p=682

Sony have another process they call APR, which you have to do every day, every time you turn the camera on in the morning. If you don't and skip it (which I've done) you also get nice stuck pixels instead.

I haven't worked with the Varicam for a while so I haven't seen what Panny are doing these days, but I imagine it's like sony, that a black balance sorts it once you're at operating temperature.

I think that maybe the problem is is that BMD firstly give you everything from the sensor. What they maybe should do is truncate the bottom end and just clip it and have a bit less dynamic range and be done with it.

The second issue is that BMD's log is especially flat. You look at a low con subject in BMD film on any LCD monitor and you'll see noise and FPN that end up disappearing in a normalised grade.

Now, you may well have an abnormal camera too. So post some shots !

Jakelutz wrote:Yeah the 4.6k can create beautiful images but there are a good 4 stops of decent latitude going to waste. This camera would be truly perfect if it weren't for that FPN.


If you think you're losing 4 stops to FPN then I'd say there's something extremely wrong with your camera.


Jakelutz wrote:I'd love for a BM employee to give us some info on this infamous FPN issue. ;)


You may get someone to respond, but I doubt it :-)

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 4:11 am

Thanks everyone for the huge amount of info, I've been researching this since I got my camera a year ago, to be clear a well lit scene with correct lighting a 400 ISO looks great and so does footage shot at day, no FPN whatsoever. Switching to 800 ISO in every other situation is unusable.
Warwick my local BM support tech assures me Blackmagic deem the sensor calibration on my specific unit to be the best it can possibly be.

I've even said I would pay the extra labour to have it carefully calibrated over a long period of time, Warwick said that would make no difference so they must have a step by step process that's not very specific, maybe a warm up and click calibrate process? Who knows.

Implementing a user calibrate settting would be a good start.

I've sent many DNGs but I'm yet to upload a 4k sample to Vimeo as moving shots show it the best when your eyes are scanning the scene.

In my honest opinion, Blackmagic designed a cost effective sensor from a manufacturer, cooled it to the point the noise is nearly non existent but was then left with the FPN. Still sadly, they promised 15 stops of dynamic range which is technically correct but more like 12 usable stops when crushing the footage to hide FPN.

Btw your pier footage looked amazing John, I could push it to 6400 ISO equivalent and couldn't see FPN which constantly had me worried about my specific unit being different. I'll upload some DNGs and you can take a look.
Offline

Mikko Parttimaa

  • Posts: 10
  • Joined: Sat Jun 03, 2017 2:28 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 11:41 am

From my experience, it feels like the quality of sensors or calibration varies. I have shot with two Pro bodies which both show quite similar amounts fpn. I've sent dngs to support and they say they are normal. But i've also graded some footage from a older 4.6k body which had much worse fpn issues in similar settings.

The very first examples posted public shot by John can be pushed like any other camera. You get noise but no fpn in the blacks. Expect for the car shot, which I think is a different EF mount body while other footage is PL?

Other cameras do also exhibit fpn when pushed to extremes. But with others, like Alexas or FS7s, the noise or compression breaks the picture long before fpn. I've shot and graded quite some footage from both and saved some crazily underexposed shots, which with Ursa bodies I have used would have been unusable. Alexa gets noisy and FS7 muddy but that can be acceptable, fixed pattern over the whole picture can not.

At the moment I can quite happily shoot at 800 with good exposure. Then there are no problems and images are gorgeous (or at least the IQ is :D ). Still a confirmation would be nice: is this as good as the sensor is ever going to get, or is BM working on the sensor calibration to make it better in the future? Or is this just about the varying quality of the sensors out there? Or if the camera processing can't deal with subtracting the fpn maybe it could be done in Resolve instead?
Offline

Ryan Payne

  • Posts: 247
  • Joined: Sat Apr 15, 2017 6:48 am

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 12:24 pm

Sorry to go slightly OT but on the topic of FPN why would it be that seemingly the Pocket has such little FPN in comparison to the Micro?

Would it be that most people are simply viewing the Micro on the better monitor and there for see more imperfections? Or is there an actual difference and why hasn't it been solved yet? Much like the rest of the micros issues and shortcomings.
Offline

Dan Shay

  • Posts: 36
  • Joined: Tue Apr 12, 2016 4:06 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 1:58 pm

The blackshading on the URSA 4.6k doesn't remove fpn. It fixes color shift to blue as you underexpose. FPN noise occurs because of small imbalances in gain at sensor sites. It could be removed by sampling numerous frames(in darkness) and creating a map. Then using Resolve to subtract the fpn noise map from the images. If BM created an automatic method of doing this it would be great. URSA mini 4.6k also has a problem with cmos smear, which is often confused with fpn. If high contrast images are underexposed, starting at about 3 stops under, the bright objects will create brightening horizontal streaks, similar to a lens flare. The Canon C300 mark II has the same problem.
Offline

Mikko Parttimaa

  • Posts: 10
  • Joined: Sat Jun 03, 2017 2:28 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 2:10 pm

Ryan: they have different sensors. I've read some speculation around that the global shutter circuitry is what is causing fpn in BMPC and all the other cameras made after it. Even though it's not in use, it's still there. But can't know for sure.
Offline
User avatar

Nick Gombinsky

  • Posts: 248
  • Joined: Fri Jul 19, 2013 5:30 pm
  • Location: Buenos Aires, Argentina

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 4:16 pm

I've had two UM46K. The first one, which was from the first preorder batches, had some horrible FPN if you looked for it in a low key scene. It all went away after a couple of firmware updates, and having the black shading tool in the menu. Just do it every time the temperature and humidity changes around you, and you're golden.

If there is a lot of noise and lines and sh...tuff, and stuff, maybe you need to send it back.
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 4:55 pm

Nick Gombinsky wrote:I and having the black shading tool in the menu. Just do it every time the temperature and humidity changes around you, and you're golden.


As mentioned earlier in this thread, the sensor calibration isn't really a black shading calibration tool. They're different.

It's not necessary to run the sensor calibration tool as often as you would for other camera that require a near daily black shade. They're not really the same thing. So changing temperature or humidity won't be a trigger for running this process.

The sensor calibration process is more like every few months.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 4:59 pm

Dan Shay wrote: It could be removed by sampling numerous frames(in darkness) and creating a map. Then using Resolve to subtract the fpn noise map from the images. If BM created an automatic method of doing this it would be great.


Dark frame subtraction is a great technique. For it to work the dark frame has to be generated by the same camera that you shot the footage on. But it's probably the stand alone best way to reduce FPN.

Dan Shay wrote:URSA mini 4.6k also has a problem with cmos smear, which is often confused with fpn. If high contrast images are underexposed, starting at about 3 stops under, the bright objects will create brightening horizontal streaks, similar to a lens flare. The Canon C300 mark II has the same problem.


Well Alexa and RED too...a post from 2011 mind....

https://johnbrawley.wordpress.com/2011/ ... ind-trick/

All CMOS cameras have this smear issue.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 5:20 pm

Jakelutz wrote:Switching to 800 ISO in every other situation is unusable.


Then for sure something is wrong with your camera. You should never see FPN in normally exposed or even low contrast scenes that are shot at and exposed for 800 ISO. If you are then you've got an issue with your camera and you should be taking it up with BMD.

What would be best is to post some examples, or even better, shoot some for us to best illustrate what you're seeing. These will also help if you're taking it up with BMD and they're not seeing the same results.

Jakelutz wrote:said that would make no difference so they must have a step by step process that's not very specific, maybe a warm up and click calibrate process? Who knows.

Implementing a user calibrate settting would be a good start.


There's an ongoing assumption that a user activated calibration would help. Alexa's don't have a user generated or initiated black shade or calibration. That's because they have a different way of managing heat and noise. BMD have followed the same approach.

The Arri model is to maintain thermal stability in all shooting conditions using solid state cooling. Other companies like RED use a fan so there's a lot more variation (hence the need for 5 memory settings) and the way the fan noise can change over a long take. Arri say they maintain a sensor temperature that's within + or- 0.5 Deg C no matter what else is happening around the camera. That includes HEATING the sensor if it's super cold.

The fan on the Alexa doesn't so much cool the sensor. It moves heat away from the heatsink that's had heat piped to it from a peltier / solid state cooling system. That system VARIES the cooling depending on how hot (or cold) the sensor is getting. An Alexa Mini is the same as a regular alexa with a lot less heat sink and a more direct fan based cooling so we get increased and now variable fan noise on an Alexa Mini.

By maintaining a very narrow and very stable window of operating temperature they can fine tune the noise and FPN characteristics.

Jakelutz wrote: Still sadly, they promised 15 stops of dynamic range which is technically correct but more like 12 usable stops when crushing the footage to hide FPN.


Again, I haven't seen your shots, but if you really are losing 3 stops (and you said 4 before) to FPN then your camera has a major problem.


Jakelutz wrote:Btw your pier footage looked amazing John, I could push it to 6400 ISO equivalent and couldn't see FPN which constantly had me worried about my specific unit being different. I'll upload some DNGs and you can take a look.


FPN comes out in different ways. It depends on what you're shooting and how you expose for it. My dailies colourist from the last series that I just finished ended up telling me he could do a BETTER job of salvaging very underexposed BMD images compared to Arri Alexa Images shot at the same time. I suspect this is perceptual because the UM4.6K has more pixels and therefore less noise once scaled to 1920, but I digress.

Here's some recent salvage work for those that haven't seen what can be done. There is FPN in there but it's such an extreme example, I'm frankly amazed that this shot survived. Not only that it made it to air.

JB

ExDStill_QOTS_170131_010438.jpg
ExDStill_QOTS_170131_010438.jpg (262.2 KiB) Viewed 16919 times

ExDStill_QOTS_170131_010447.jpg
ExDStill_QOTS_170131_010447.jpg (762.92 KiB) Viewed 16919 times
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Ryan Payne

  • Posts: 247
  • Joined: Sat Apr 15, 2017 6:48 am

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 08, 2017 11:54 pm

Mikko Parttimaa wrote:Ryan: they have different sensors. I've read some speculation around that the global shutter circuitry is what is causing fpn in BMPC and all the other cameras made after it. Even though it's not in use, it's still there. But can't know for sure.


If that's the case for the micro then that's a real shame to get no benefit only downsides. But with the lack of updates then you may just be correct, either way it's a shame. At this point I'm expecting a micro V2 before a firmware update..
Offline

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSun Jul 09, 2017 10:43 am

John, Is it unfair for me to request a full replacement of my Blackmagic camera if it does not look as good as yours? BM's weakest point seems to be consistency between cameras.
Your footage looks flawless in regards to FPN! I'd expect to see large bands in the sky if it were my unit.
I'll add some DNGs to my dropbox soon.

Here's the FPN problem, compression helps:


Plus a little bonus treat that has shown up in the last few days:


Thanks in advance, Jake.
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSun Jul 09, 2017 4:34 pm

Jakelutz wrote:John, Is it unfair for me to request a full replacement of my Blackmagic camera if it does not look as good as yours? BM's weakest point seems to be consistency between cameras.
Your footage looks flawless in regards to FPN! I'd expect to see large bands in the sky if it were my unit.
I'll add some DNGs to my dropbox

Thanks in advance, Jake.



Very hard to tell on Vimeo. DNGs would be best, preferably of that scene.

As I've said in previous examples I look for blacks that are black in the image. I don't see anything sitting on the bottom of the waveform in your vimeo grade. Also the highlights look compressed in a way that feels a bit odd to me. Can you talk about your post process for these clips ?

The second clip is ? What. Raw ? ProRes ?

What I do see is some pretty heavy shading (discolouration / magenta edge on the right)I would submit that to BMD and ask about that.

It's up to BMD in how they deal with it, but you have to start that process. If they've turned you back previously just ask again politely and persistently with the shots you're having double with.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSun Jul 09, 2017 6:15 pm

The first clip in the dark was also in fog which is why the highlights seem to have a glow on them. I shoot everything in RAW 4:1, the dark clip being 'Blackmagic Design 4.6K Film' gamma with no curve added.

The sensor discolouration clip was processed using 'Blackmagic Design 4.6K Video' gamma to pump the sat and contrast which slightly clipped the highlights but I didn't care much for the grade as it was only shot as an example of what's happening to the sensor.

Delivered in Quicktime - MPEG4 Video

Here's a link to the DNG's for the relevant shots,
https://www.dropbox.com/sh/djqzqzb4c0lq ... IH2Qa?dl=0

Let me know if the FPN looks the same as it would on your cam, if so that's good, as for the sensor turning Hulk, I'll send BM support another email.

Interestingly I shot the 2.5K and the URSA mini side by side when the URSA first arrived and the 2.5k had more usable dynamic range if I were excluding stops with FPN. It's a bummer the BMCC didn't shoot 60fps :cry:
Offline
User avatar

Fabio Lanzone

  • Posts: 170
  • Joined: Mon Jan 14, 2013 10:26 am
  • Location: Milan, Italy

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostMon Jul 10, 2017 8:10 am

John Brawley wrote:Well let's start here. An AC colleague of mine tells me he just did a series for Netflix on RED (not yet aired). He complained about a lot of problems with the cameras they had but one that came up for him was FPN. To calibrate on the camera it's a very lengthy calibration process and that you need to do it several times for many different potential operating temperatures. In other words if the ambient temperature range varied more than about 10 degrees C they had to have a different temperature calibration. You have 5 memories in the camera available and he found it wasn't enough because it ALSO was supposed to use the same settings when working off speed, but that function didn't seem to work. Their solution was to shoot HFR with a 360Deg shutter.

Well, that was true with older models, which are not sold anymore. DSMC2 cameras' auto black shade creates maps for every temperature/integration time combination (the process takes up to 40 minutes, but you never do it again) and applies the correct map in the background as you shoot.

I think that BMD is aware of the problem, but a process like this needs horsepower and could raise camera's price.

Doing that in Resolve could be a solution, as long as the black map you feed is the one shot at the same temperature you shot the image you are going to correct. And I bet it will take a lot of render time. A messy process in the end.

Hope things will be better on next camera
URSA4K v2 EF - BMPCC 4K - BMCC EF - BMPCC - RED DRAGON
Offline

Mikko Parttimaa

  • Posts: 10
  • Joined: Sat Jun 03, 2017 2:28 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostTue Jul 11, 2017 8:57 pm

I think also it's about lack on processing power. Yes, post fix would be a bit messy but fortunately it would only be needed with really underexposed footage. It would be better than nothing!

As the sensor is probably maintained in quite stable temperatures, it might be possible to just use one map. Should do some tests to see if the FPN changes with temperature.

I hope they are working on something but all the silence makes me doubt it.
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostTue Jul 11, 2017 9:12 pm

Fabio Lanzone wrote:Well, that was true with older models, which are not sold anymore. DSMC2 cameras' auto black shade creates maps for every temperature/integration time combination (the process takes up to 40 minutes, but you never do it again) and applies the correct map in the background as you shoot.


My favourite answer when dealing with any problem you bring up with a RED camera. Use the new model.

Fabio, how many memories for black shade on your dragon ? How much temperature variation before it gives you the warning that you're operating the camera out of calibration ? More than 5 degrees right ?

Fabio Lanzone wrote:
I think that BMD is aware of the problem, but a process like this needs horsepower and could raise camera's price.


No. As I explained. They do it the same way Arri do it on their Alexa. It has nothing to do with horsepower. And everything to do with thermal stability of the sensor. I think you're failing to understand how that works and how different that is from RED and Sony. Go look up how to do a black shade on an Alexa in the manual. It's not there because it's done at the factory. Once.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostTue Jul 11, 2017 9:15 pm

Jakelutz wrote:The first clip in the dark was also in fog which is why the highlights seem to have a glow on them. I shoot everything in RAW 4:1, the dark clip being 'Blackmagic Design 4.6K Film' gamma with no curve added.

The sensor discolouration clip was processed using 'Blackmagic Design 4.6K Video' gamma to pump the sat and contrast which slightly clipped the highlights but I didn't care much for the grade as it was only shot as an example of what's happening to the sensor.

Delivered in Quicktime - MPEG4 Video

Here's a link to the DNG's for the relevant shots,
https://www.dropbox.com/sh/djqzqzb4c0lq ... IH2Qa?dl=0

Let me know if the FPN looks the same as it would on your cam, if so that's good, as for the sensor turning Hulk, I'll send BM support another email.

Interestingly I shot the 2.5K and the URSA mini side by side when the URSA first arrived and the 2.5k had more usable dynamic range if I were excluding stops with FPN. It's a bummer the BMCC didn't shoot 60fps :cry:


Jake I'll take a look.

But a couple of observations from your vimeo.

You shot at 1600 ? I always always always go from 800. It works better even in low light to lift after. And your blacks aren't black. They're all elevated and sitting higher. than where I'd expect. I'll have a look and let you know what I think. I'm on the road at the moment so it's not the most of idea circumstances for me.

JB

- currently in Atlanta
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Florian Duning

  • Posts: 43
  • Joined: Wed May 11, 2016 10:41 am

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostTue Jul 11, 2017 9:38 pm

Fog provokes FPN on the Ursa mini. Have noticed it myself. Or let's say, it makes it more visible because of the more uniformly lit areas in the image due to the fog.

But there are ways to actually increase the FPN! If you have one of the image very dark and have some brighter, but still very underexposed parts on the other half of the image, then the FPN in the very dark image areas will increase dramatically. I think it has something to do with the horizontal CMOS smear. One explanation would be, that the smear offsets the pixels' value before they are read out, therefore offsetting the brightness values and in turn throwing the factory sensor calibration data off, which then leads to increased FPN.

I might be able to replicate this and share some images.

best regards,

Florian
Offline

VicHarris

  • Posts: 236
  • Joined: Sat Aug 24, 2013 5:36 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostTue Jul 11, 2017 10:28 pm

Jakelutz wrote:The first clip in the dark was also in fog which is why the highlights seem to have a glow on them. I shoot everything in RAW 4:1, the dark clip being 'Blackmagic Design 4.6K Film' gamma with no curve added.

The sensor discolouration clip was processed using 'Blackmagic Design 4.6K Video' gamma to pump the sat and contrast which slightly clipped the highlights but I didn't care much for the grade as it was only shot as an example of what's happening to the sensor.

Delivered in Quicktime - MPEG4 Video

Here's a link to the DNG's for the relevant shots,
https://www.dropbox.com/sh/djqzqzb4c0lq ... IH2Qa?dl=0

Let me know if the FPN looks the same as it would on your cam, if so that's good, as for the sensor turning Hulk, I'll send BM support another email.

Interestingly I shot the 2.5K and the URSA mini side by side when the URSA first arrived and the 2.5k had more usable dynamic range if I were excluding stops with FPN. It's a bummer the BMCC didn't shoot 60fps :cry:


Like John keeps saying, the blacks are too lifted. Not sure what you were expecting from this camera in that light.


https://www.dropbox.com/s/50ph7v7xz2zlg ... 1.png?dl=0
Offline
User avatar

Tommaso Alvisi

  • Posts: 144
  • Joined: Wed Aug 22, 2012 7:53 am
  • Location: ITALY

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostTue Jul 11, 2017 10:29 pm

Florian Duning wrote:I might be able to replicate this and share some images.

best regards,

Florian


I'd be really curious!
Offline

Florian Duning

  • Posts: 43
  • Joined: Wed May 11, 2016 10:41 am

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostWed Jul 12, 2017 9:42 pm

I have something to look at for you all ;) Just out of curiosity I dusted off my old EOS 550D and hooked it up to a recorder. ISO1600, look at that FPN! ;)
https://we.tl/1YeaTyiVFr

I have not expected that much. Looks basically the same as the Blackmagic FPN.

After all I am starting to think, that many cameras exhibit FPN under certain conditions, but we are just drilled to search for it on BM cameras.
Offline

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostThu Jul 13, 2017 8:33 pm

Florian, I have actually experienced exactly what you're talking about with the smearing causing more FPN. I pointed my camera at a pc monitor and the entire frame horizontally from the pc screen had stronger FPN which confirms this.

Vic, thankyou it's already been mentioned that the blacks are too lifted. A common thing when using Blackmagic Film log.
When I shot in the fog I knew there'd be FPN.
The big question is, are all URSA minis made equally?
We're all trying to figure how Blackmagic would fix this in a future update for some peace of mind.
I understand if people are mad they didn't get the dynamic range that was promised.
It's like selling anything and saying it can do what it really can't.

John, thanks for taking the time to help out wherever you can, I know you must be busy.
Offline

Florian Duning

  • Posts: 43
  • Joined: Wed May 11, 2016 10:41 am

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostThu Jul 13, 2017 9:56 pm

I believe that all Ursa Mini 4.6Ks have sensors that perform nearly the same. I think the difference in FPN performance from camera to camera comes from the factory calibration (NOT user blackshading). The hardware is most likely fine. Its just, that the calibration process in the factory was a little off. I had my camera recently recalibrated by BM and it improved noise uniformity and decreased FPN. It is still there and on a level what I see from other concerned users.

Lets not make FPN the counterpart of a hypochondriacs' cancer. You notice a symptom, you google it and from then on it goes downhill. You believe that your camera exhibits a symptom stronger than other cameras but thats not true. As a past professionell hypochondriac I know what I am talking about :D

BTW: User accessible blackshading has no influence on FPN on the Ursa minis. It just calculates a global offset for each color channel to remove color casts in the blacks. So its like a whitebalance for blacks. To adress FPN you would need a calibration that can calculate offsets for each and every photosite itself. This is what RED cameras do.
Offline

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 14, 2017 3:51 am

Once you see it, it's extremely hard to unsee. Especially when using raw and uncompressed formats.
Google image search any webcam and you'll see FPN, even the old DSLRS like you mentioned had plenty.
In 2015 -2017 you'd be lucky to find a camera over $5000 which has the same problem. It's not good enough when a "professional cinema camera" has this problem to such a degree. Not to mention my sensor has now gone way off calibration on the RH side. So now it could easily happen again.

Here's a great webcam image but hey I'm sure it's ok when correctly exposed..
Image
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 14, 2017 3:52 am

Jakelutz wrote:The big question is, are all URSA minis made equally?
We're all trying to figure how Blackmagic would fix this in a future update for some peace of mind.


I looked at your DNG's.

It's kind of the worst possible scenario. Large underexposed (dark) areas with no visible detail to recover with different and changing transient light areas overlayed that are translucent (fog) and moving along with also, a moving camera. Seriously I couldn't come up with a more difficult scenario if I tried. You're trying to see in the dark with thin semi translucent white veil over the top.

I could see lines when I really looked for them on a normalised grade. If the shot was locked off you probably wouldn't notice them. If there was no fog you definitely wouldn't see any. I think I'd probably get away with it in my environments (even hand held)... I mean it would probably pass a tech check. But I think you'd have to do something more obvious than that for BMD to change the body. I'm pretty sure If I did the same thing on the bodies I own I'd get the same result.

I'd try do a less stressful test than that to get them to swap over your body. If your body is out of spec, i don't think it's by much is what I'm trying to say, but that's just me guessing off a DNG.


Florian Duning wrote:BTW: User accessible blackshading has no influence on FPN on the Ursa minis. It just calculates a global offset for each color channel to remove color casts in the blacks. So its like a whitebalance for blacks. To adress FPN you would need a calibration that can calculate offsets for each and every photosite itself. This is what RED cameras do.


Which is why the process is called "sensor calibration" in BM's menu not "black shading".

JB
Last edited by John Brawley on Fri Jul 14, 2017 3:55 am, edited 1 time in total.
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

rick.lang

  • Posts: 13843
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 14, 2017 3:53 am

Thanks, Florian. I'm guilty of thinking that the Black Shading Calibration affected the visibility of the fixed pattern noise. Recently John Brawley posted that they were different items which supports your statement.


Sent from my iPhone using Tapatalk
Rick Lang
Offline

John Brawley

  • Posts: 2634
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Jul 14, 2017 4:16 am

Jakelutz wrote:Once you see it, it's extremely hard to unsee. Especially when using raw and uncompressed formats.
Google image search any webcam and you'll see FPN, even the old DSLRS like you mentioned had plenty.


It's all about degrees.

The first shot in this clip is Ursa mini, as are most of the super close subjective closeups. I point out the first shot because it's shot in a dark nightclub with only LED strip lightings and a couple of crappy nightclub movers and a single tungsten Kino Tube with CTS and a lot of atmosphere from DF50 hazers in the air. aka fog.

There's some super close macro shots that I think are also Ursa Mini, and even a BM micro camera for the body rig shot as well.

From memory I pulled focus myself on the first shot on a 35mm Ultraspeed at T1.4

http://www.usanetwork.com/queenofthesou ... l#fb.71217

You can shoot handheld in low light with haze, I think it's about degrees and expectation. You're comparing very extreme exposure differences.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Jul 15, 2017 7:32 am

There's a lot of information in this thread, exactly what I was hoping for. This should serve as a good starting point for people looking for info on FPN.
That footage looks great John, if you think my stuff looks the same as what you'd get I'm happy to deal with the amount of FPN I'm seeing. Even with the loss of DN, URSAs are the best bang for the buck. Especially for us who cannot afford alexas or high end Reds, nothing comes close.
Until anyone from Blackmagic can shed some light on the subject I guess we wait until the hopeful firmware "added advanced user sensor calibration" into the menu. I wouldn't mind the ten minutes it takes on location to get rid of the FPN.

Thanks all.
Offline

Denny Smith

  • Posts: 12886
  • Joined: Thu Aug 01, 2013 4:19 pm
  • Location: USA, Northern Calif.

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSun Jul 16, 2017 5:53 pm

The bottom line is, all CMOS sensors have FPN to some degree... and from,what I have seen form JB and others, is is not any more of an issue with the BM 4:6 than it is with other cameras. :roll:
Cheers
Denny Smith
SHA Productions
Offline

Mikko Parttimaa

  • Posts: 10
  • Joined: Sat Jun 03, 2017 2:28 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSun Jul 16, 2017 11:29 pm

Denny Smith wrote:The bottom line is, all CMOS sensors have FPN to some degree... and from,what I have seen form JB and others, is is not any more of an issue with the BM 4:6 than it is with other cameras. :roll:
Cheers


I would say most of the time it's not an issue. But it is more of an issue than with other cinema cameras which have a way of dealing with it.
Offline

Denny Smith

  • Posts: 12886
  • Joined: Thu Aug 01, 2013 4:19 pm
  • Location: USA, Northern Calif.

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostMon Jul 17, 2017 6:11 am

I started shooting digital when the first CCD sensor cameras were released, never had an FPN or other artifact issue, except smearing with fast pans and bright lights. CMOS sensors on the other hand can exhibit several types of artifacts, FPN being one, and if you look hard enough you can see it in most CMOS sensor cameras. Yes, some cameras seem more prone to FPN than others, BM cameras are no exception.

But if you set up the camera correctly, as John Brawley mentioned, expose correctly, run the black balance program, and allow the camera to reach its operating temp before using, FPN will not be an issue is most situations. You are not going to find a FPN free CMOS sensor camera, short of an ARRI Alexa.
Cheers
Denny Smith
SHA Productions
Offline

Jakelutz

  • Posts: 14
  • Joined: Wed Apr 24, 2013 5:49 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Aug 18, 2017 3:27 am

An update,
So after the RH side of my sensor started turning a strong green colour they deemed it out of spec and replaced with new.

The new unit still has some green on the RH side BUT the FPN seems so much better to the point it is indistinguishable to the noise. Great, so I can now fix some of the sensor irregularities in post which is good.

If anyone else is interested in the green tint I'm seeing, check these links.
BM has deemed this sensor to be within the same spec as all other 4.6k units.
Interested to see if this is on a spectrum of other units or if all are the same.

https://www.dropbox.com/sh/sm42ledzhi7z ... XQHna?dl=0

Thanks, Jake
Offline
User avatar

Frank Glencairn

  • Posts: 1662
  • Joined: Wed Aug 22, 2012 7:07 am
  • Location: Germany

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Aug 18, 2017 7:39 am

Just out of curiosity, what is you grading workflow in Resolve?

I usually apply a LUT (not BMs video LUT) first, and than I do most of my grading in the raw tab.
Only secondary correction is done on the RGB image.

I run only once or twice into FPN (and that was on material I haven't shot myself), so my workflow may have something to do with it, since I doubt, that all my BM cameras have magic sensors.
http://frankglencairn.wordpress.com/

I told you so :-)
Offline

Uli Plank

  • Posts: 7761
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Aug 18, 2017 8:18 am

Denny Smith wrote: You are not going to find a FPN free CMOS sensor camera, short of an ARRI Alexa


I'd include the Red as well, but at the cost of very long calibration times.
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
Offline

Mikko Parttimaa

  • Posts: 10
  • Joined: Sat Jun 03, 2017 2:28 pm

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Aug 18, 2017 1:52 pm

Haven't seen it on FS7 either. You can get crosshatching kind of pattern with flares, but no lines when lifting shadows.
Offline

Ryan Hamblin

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

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostFri Aug 18, 2017 3:33 pm

I have provoked fpn on RED as well. One of our biggest issues with red outside of it being rather soft for the amount of pixels your dealing with is that you can very quickly provoke chroma noise. And it's just real ugly
www.brainstem.tv
www.ryanhamblin.com
Offline

Uli Plank

  • Posts: 7761
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Aug 19, 2017 7:57 am

True, the older Red sensors have more chroma noise than a BM, but when properly calibrated it has no regular pattern.
The softness, OTOH, is the result of the OLPF, which avoids the moiré you see with BM's cameras.
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
Offline

Ryan Hamblin

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

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSat Aug 19, 2017 7:17 pm

Uli,

Shot and owned RED gear since basically their inception. We have shot or tested every version of their cameras. The Chroma noise is still their even on the 8k... its way more subdued, but as we are delivering 4k now to all our clients it just is a real pain to have to denoise so much footage. The 8k also was provoke moire more than the 4.6k sensor... I nor any of my colleages could figure out why, but it put the camera out of the running as an additional camera to our 4 4.6k's on a live concert shoot. I am not knocking what RED allowed for me and my carrer, but they just seem to not be holding up in the noise and detail world at the moment. That of course is my opinion and the post side of our company. The DR was showing about 1 stop less in the highlights as well.

This is a complete sidestep on things, but do you think that the compression level put on the 8k also has a significant detriment to the detail level? In our test the 4k prores that was kicked out from the camera at times seemed to hold a bit more detail. I am wondering if that has to do with the downscale and then compression ( basically having to compress less data so you in a sense get less compression ).
www.brainstem.tv
www.ryanhamblin.com
Offline

Uli Plank

  • Posts: 7761
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSun Aug 20, 2017 12:42 pm

Well, I also have worked with all Red cameras from the Red One, firmware 12 or so.
Only had a short test of an 8K Red and concluded to pass. We have two Scarlet MX and one Epic Dragon and are quite happy with those. If we'd need another camera I'd rather consider a UM46 Pro now.

Regarding your observations: yes, the noise seemed not very low on the 8K. This is to be expected when photo cells get smaller, you can't cheat the laws of physics. Plus, wavelet compression is making an image softer, that's right.

8K from a Bayer sensor are not 8K resolution anyway. If you supersample and downscale, you'll need to "choose your poison" regarding scaling algorithms and may need some mild sharpening in the end.

But I get the feeling we are derailing this thread a bit…
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
Offline

Ryan Hamblin

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

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostSun Aug 20, 2017 1:25 pm

Indeed, and now to our previously scheduled programming...
www.brainstem.tv
www.ryanhamblin.com
Offline

gigi_logfilm

  • Posts: 5
  • Joined: Wed Apr 11, 2018 10:19 am
  • Real Name: Gigi Tufano

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostWed Apr 11, 2018 10:37 am

Hi,
I just bought a second hand Production Camera 4k. It has a terrible FPN. I upgraded the firmware to the last version (it had the 2.01 firmware) but nothing changed. Is there a CMOS replacement program for very bad condition's CMOS or something I can do to fix it? Sorry but I am pretty new to Blackmagic as a owner. I used this camera many times renting it from local renthouse but I never had any kind of problems.
thanks
ginuz.
Offline
User avatar

Xtreemtec

  • Posts: 4503
  • Joined: Wed Jan 02, 2013 11:48 am
  • Location: The Netherlands

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostThu Apr 12, 2018 11:42 am

The Production Camera 4K is not supported anymore for about 2 years now... Since more model of the Ursa Mini came..

You could put a requast at BMD Support.. But i guess the Camera is outside it's waranty periode.. And thus likely there will be no support on these old models anymore..

Buying a 2nd hand BMD camera is always risky as most of the times the waranty has been expired.. Meaning the support becomes very difficult.. (Same applies to almost all BMD hardware outside the waranty periode..)

You should buy new gear and sell it at the point waranty will expire.. That way you always work with the newest gear and always are able to get your gear repaired by BMD.. ;)
Daniel Wittenaar .:: Xtreemtec Media Productions ::. -= www.xtreemtec.nl =-
4K OBV Trailer, ATEM TVS HD, 4M/E Broadcast Studio 4K, Constelation 8K, Hyperdeck Studio 12G, Ursa Broadcast 4K, 4K fiber converters with Sony Control
Offline
User avatar

Frank Glencairn

  • Posts: 1662
  • Joined: Wed Aug 22, 2012 7:07 am
  • Location: Germany

Re: Will Blackmagic ever fix the FPN on the 4.6K

PostMon May 07, 2018 7:45 am

Will Blackmagic ever fix the FPN on the 4.6K


It goes away, the day folks learn to light and grade :D
http://frankglencairn.wordpress.com/

I told you so :-)
Next

Return to Cinematography

Who is online

Users browsing this forum: bohndigga and 35 guests