URSA Mini 4.6k The Original™ Magenta Cast Issue

http://imgur.com/a/mZaE5

Posted in the other thread as well. I tested my new 4.6k MINI today and see no magenta cast anywhere.

Tested

RAW
prores
prores w/no lens

Hey guys,

Had a chance to do a quick test with my UM46. Keep in mind I don't have the EVF or a spot meter right now, so the lighting balance levels across the frame were done by eye; not a super scientific test for the vignetting. It is however a decent test regarding color cast.

Camera balanced to 5600K ISO 800, lit with Dracast LED 5600K balanced lights. Rokinon 24mm T5.6. Keep in mind I don't have a color temp meter either, so I cannot guarantee the Dracasts are balanced perfectly.

Here's a link to the DNG: http://bit.ly/1MCaeSs

First image is straight from the camera with only exposure/saturation adjustments.

Second image the same as above, but WB set to grey card in the RAW tab--TEMP: 4620 TINT: -5.20

I agree other cameras have color casts and the issue can be corrected in Resolve, however I didn't experience the vignetting with my BMPC. Could this be due to the larger sensor size? Are there better lenses for this? Thanks guys.

For what it's worth, I tested my MINI 4.6k I got last week and see no magenta cast anywhere:

http://imgur.com/a/mZaE5

Tested with a white card outside.

RAW
prores 444
and one with no lens

Were these graded? I have not seen whites like that straight out of the camera under any lighting conditions.

Just for giggles can we see one of these tests with an IR cut filter

http://imgur.com/a/mZaE5

Posted in the other thread as well. I tested my new 4.6k MINI today and see no magenta cast anywhere.

Tested

RAW
prores
prores w/no lens

Cool! I take it you didn't use an IR cut filter?

For what it's worth, I tested my MINI 4.6k I got last week and see no magenta cast anywhere:

http://imgur.com/a/mZaE5

Tested with a white card outside.

RAW
prores 444
and one with no lens

Were these graded? I have not seen whites like that straight out of the camera under any lighting conditions.

no grading
iso200
wb 4500k
f22
shot in direct sun
video dynamic range

http://imgur.com/a/mZaE5

Posted in the other thread as well. I tested my new 4.6k MINI today and see no magenta cast anywhere.

Tested

RAW
prores
prores w/no lens

Cool! I take it you didn't use an IR cut filter?

no, sure didn't

http://imgur.com/a/mZaE5

Posted in the other thread as well. I tested my new 4.6k MINI today and see no magenta cast anywhere.

Tested

RAW
prores
prores w/no lens

Great news! I'll be testing mine soon.

Just inspected the sensor, i most definitely can see a finger like smudge on that protection glass. Just a wild thought, could that be the problem? Maybe a residue left on that glass? My 3 year old bmcc has a cleaner glass.

One thing is sure, it's not mine. I'm very careful with it, in a dust free envoirment.

Just inspected the sensor, i most definitely can see a finger like smudge on that protection glass. Just a wild thought, could that be the problem? Maybe a residue left on that glass? My 3 year old bmcc has a cleaner glass.

One thing is sure, it's not mine. I'm very careful with it, in a dust free envoirment.

The plot thickens. A fingerprint! This sounds like an episode of CSI now - ENHANCE!

Just inspected the sensor, i most definitely can see a finger like smudge on that protection glass. Just a wild thought, could that be the problem? Maybe a residue left on that glass? My 3 year old bmcc has a cleaner glass.

One thing is sure, it's not mine. I'm very careful with it, in a dust free envoirment.

The plot thickens. A fingerprint! This sounds like an episode of CSI now - ENHANCE!

Well at this state, i can't really laugh about this at all. I'm glad your ursa is fine. I'm just trying to help since we don't hear a word from bmd till now. I feel quite sad about this, and I'm just trying to enjoy my ursa which is very hard now. Can't use it for paid work now.

Hey guys,

Had a chance to do a quick test with my UM46. Keep in mind I don't have the EVF or a spot meter right now, so the lighting balance levels across the frame were done by eye; not a super scientific test for the vignetting. It is however a decent test regarding color cast.

Camera balanced to 5600K ISO 800, lit with Dracast LED 5600K balanced lights. Rokinon 24mm T5.6. Keep in mind I don't have a color temp meter either, so I cannot guarantee the Dracasts are balanced perfectly.

Here's a link to the DNG: http://bit.ly/1MCaeSs

First image is straight from the camera with only exposure/saturation adjustments.

Second image the same as above, but WB set to grey card in the RAW tab--TEMP: 4620 TINT: -5.20

I agree other cameras have color casts and the issue can be corrected in Resolve, however I didn't experience the vignetting with my BMPC. Could this be due to the larger sensor size? Are there better lenses for this? Thanks guys.

Is it just my monitor or am I still seeing it in the corners of the corrected one?

Just inspected the sensor, i most definitely can see a finger like smudge on that protection glass. Just a wild thought, could that be the problem? Maybe a residue left on that glass? My 3 year old bmcc has a cleaner glass.

One thing is sure, it's not mine. I'm very careful with it, in a dust free envoirment.

The plot thickens. A fingerprint! This sounds like an episode of CSI now - ENHANCE!

Well at this state, i can't really laugh about this at all. I'm glad your ursa is fine. I'm just trying to help since we don't hear a word from bmd till now. I feel quite sad about this, and I'm just trying to enjoy my ursa which is very hard now. Can't use it for paid work now.

My apologies -

Well, I done goofed. I was running a slightly older version of V12 and using the BMD film gamma instead of the 4.6 gamma. After using it, the results were much easier to attain. A little work with the power windows and offset and I was able to eliminate the lens vignetting. Just another example of how important working with Resolve is to this camera, which is not a surprise:

Also wanted to add the disclaimer that I'm not a professional colorist, so someone from Company 3 or the like can attain much better results; take my quick test with a grain of salt. Anyway, been loving the ease of use with the UM46!

This is a low light grade 1600 iso

MiniMovie_1.27.3.jpg

I like this low light shot. You should add it to the lowlight thread on BMCuser.

Well this is disappointing I really thought we'd be hearing something from blackmagic today due to the severity of the issue. Especially when they have been pretty active and present on other threads.

I also have this issue with my 4.6K EF.

We do as well


Sent from my iPhone using Tapatalk

Well, I done goofed. I was running a slightly older version of V12 and using the BMD film gamma instead of the 4.6 gamma. After using it, the results were much easier to attain. A little work with the power windows and offset and I was able to eliminate the lens vignetting. Just another example of how important working with Resolve is to this camera, which is not a surprise:

Did you guys read this and confirm using the latest resolve with the BMD 4.6k gamma. Just want to eliminate other variables that may be helping to cause this as well.

For any BM folks looking into this, here's a DNG I just grabbed of an evenly lit source with the lens off, no Resolve involved:

http://bit.ly/22QnfQ2

For those individuals experiencing the problem, please let us know what blackmagic tells you. My guess is it's going to have to be an RMA from what I'm seeing this is not something that is going to be fixable with a firmware update.

My camera is out at the Fremont repair site now, having the #2 card slot fixed. I called my contact out there to have the techs take a look at my sensor. I did not notice any weirdness when I was shooting last weekend and just took a look at my DNGs, which look clean, but certainly not tested in a controlled environment. However, it's out there and I am hoping that the techs can do some tests, as my camera was from the first lot to B&H. I will let peeps in the know if I learn anything. In the meantime, keep in mind that BMD is not going to respond rapidly to this issue that people are noticing. They will conduct controlled and lot sampled testing to try and pinpoint the problem(s); and, I can't imagine that any of us would want them to do otherwise.

[Kristian Lam]

Hi guys,

We’re going through the posts and looking into this.

In the meantime, I would appreciate it if you contact our support office to provide more details, such as the type of lens used, aperture settings and if possible, provide DNGs shot without any lens attached. Please make sure that you don’t clip anything and have it white balanced.

You should also make sure to check the RAW DNG files in an application such Resolve or Photoshop. We’ve noticed that Preview in OS X generates a very weird over constrasty image.

Thanks Kristian, I messaged the support email letting them know I’m available to provide further tests/info/DNG files regarding the issue.

I will test mine tomorrow and will let you guys know.

http://imgur.com/a/mZaE5

Posted in the other thread as well. I tested my new 4.6k MINI today and see no magenta cast anywhere.

Tested

RAW
prores
prores w/no lens

This is reassuring. I hope most people are getting sensors like this. And for others there should be RMA. Waiting to see Benton's tests.

Well, I done goofed. I was running a slightly older version of V12 and using the BMD film gamma instead of the 4.6 gamma. After using it, the results were much easier to attain. A little work with the power windows and offset and I was able to eliminate the lens vignetting. Just another example of how important working with Resolve is to this camera, which is not a surprise:

This first Parade is from your image which you balanced to :-

"WB set to grey card in the RAW tab--TEMP: 4620 TINT: -5.20"

erica.jpg (823.54 KiB) Viewed 29833 times

ericascope.jpg (52.19 KiB) Viewed 29833 times

Well, I done goofed. I was running a slightly older version of V12 and using the BMD film gamma instead of the 4.6 gamma. After using it, the results were much easier to attain. A little work with the power windows and offset and I was able to eliminate the lens vignetting. Just another example of how important working with Resolve is to this camera, which is not a surprise:

This second one is from the image above and to me it leaps off the screen and wouldn't call it eliminated at all :-

ericb.jpg (640.16 KiB) Viewed 29829 times

ericbscope.jpg (52.07 KiB) Viewed 29829 times

http://imgur.com/a/mZaE5

Posted in the other thread as well. I tested my new 4.6k MINI today and see no magenta cast anywhere.

Tested

RAW
prores
prores w/no lens

Not sure if you are taking the mickey or not but your 'white' No Lens image actually has no data and in the other two there is a definite drop off on the left and right :-

nolens.png (21.51 KiB) Viewed 29348 times

prores444.jpg (108.85 KiB) Viewed 29348 times

prores444parade.jpg (52.24 KiB) Viewed 29346 times

Still have not recieved my copy yet... All this delay and silence and then you see this.. it does leave a bitter taste in the mouth..

Figure BMD will get on top of this relatively quickly.

So we spend lots of time discussing the tech and art of cameras, screen etc.

If you are actually keen to play a game to see how your eyes perform, try this and put yourself under pressure by only giving yourself 90 secs per row....see how you go :-

http://www.colormunki.com/game/huetest_kiosk

That was pretty freaking cool. Scored a Zero which is best?

I've always said I could see in a trillion colors. LOL Actually my monitor being a plasma did all the heavy lifting for me.
Thanks for that link. That was first try and took me about two minutes each row.

The attachment Screen Shot 2016-03-30 at 3.00.42 PM.png is no longer available

Yeah, now I understand....you and are I gunna debate color till the cows come home...same (;->

Perfect.JPG (118.21 KiB) Viewed 29801 times

p.s. Have given up on the idea of 'screen capture' of images on my PC as I have no idea what the cause of the hotter color, so am just screen capturing the Resolve Parade. Is a heck of a lot more useful anyway.

For some perspective....

Here's a DNG from another camera. This is my Leica M-E Type 220. I just this evening shot this without a lens on the camera. It shoots DNG as well. Even though this is a current Leica camera this sensor is an old sensor design, the CCD sensor that is made by TrueSense / Kodak, the same as the digital Bolex.

I wasn't sure what to expect.

Leica M-E

L1005515.jpg (543.37 KiB) Viewed 29737 times

So what's interesting to me here, is that there's a fall off into the corners AND there's a magenta cast at the bottom but a green tinge across the top. These are two separate issues, that maybe compound to look the same = magenta cast


Here's the original file.
https://drive.google.com/a/johnbrawley. ... sp=sharing

Thing is that I've been using this camera to shoot GREAT photos. Now I'm not saying that there aren't some issues that need to be worked though by Blackmagic, but even at it's worst I think you'd be hard pressed to pick this on corrected footage. My example above kind of surprised me. Because I've been using it for about 18 months to take some great photos ! Sorry to spam the thread with irrelevant photos, but the point I make is that with the fault above, I captured the photos below, and I'd say it's pretty hard to pick it.

JB

L1000185 by John Brawley, on Flickr

L1001726 by John Brawley, on Flickr

L1002378 by John Brawley, on Flickr

L1000517 by John Brawley, on Flickr

L1001720 by John Brawley, on Flickr

L1000535 by John Brawley, on Flickr

L1000574 by John Brawley, on Flickr

L1002154 by John Brawley, on Flickr

L1001487 by John Brawley, on Flickr

John, great pictures! Loving them. But in this case it's not something I discovered by just pointing the camera to a white wall. I discovered this issue by seeing it in real life scenarios, that's why I took the test. Otherwise I probably didn't knew either.

John, great pictures! Loving them. But in this case it's not something I discovered by just pointing the camera to a white wall. I discovered this issue by seeing it in real life scenarios, that's why I took the test. Otherwise I probably didn't knew either.

For sure and I don't mean to discount your issue at all. If you or anyone is having issue with it then you should be going straight to Blackmagic service to discuss. But I'm seeing others posting images that have less of a severity of this issue. The point I make is that most cameras have some kind of degree of this issue. It's not exclusive to Blackmagic.

I think most cameras will fall off in the corners like this. I think some cameras have shading issues. My very expensive Leica M-E, long renowned for their image and IQ has got a similar problem. In fact it's just as bad because it's green AND magenta. It's about DEGREES of how much this problem actually affects work.

I'm not trying to downplay what might be happening. I'm not trying to dismiss at all anyone who thinks they have an issue. But I invite you to try the same test with other cameras. You might be surprised at what you find.

JB

My very expensive Leica M-E, long renowned for their image and IQ...

Marketing hype.

...has got a similar problem.

A flat sensor plane should not have a cast (unless introduced by lens optics). I fix color cast all the time when shooting product in C1 with LCC's. Your Leica example is worse than my Phase1 (w/ a much larger sensor) on a full shift (yes flat plane still I know, but it's about the directional light rays) Fuji680 photo.

In fact it's just as bad because it's green AND magenta.

Add green and make it all green. Add magenta and make it all magenta.


EDIT: Unless letting bare light rays hit the sensor (at such an angle) without optics is causing an non uniform cast (in John's example).

You guys seem to be overreacting to me. If not, then there should be recalls/RMAs on the red dragon because it's got a SERIOUS green problem. Looks at his face here, it looks terrible. Ryan's URSA Mini at least makes his face look healthy and good. Even the Arri he looks a bit dead and lifeless straight from camera. But you can see if you added some color to the Arri it would probably be fairly close to the ursa, or if you pulled some color from the URSA the same. The dragon doesn't look good here. But I haven't heard of any RMA's for the very serious green issue it seems to have.



http://s9.postimg.org/qj1gzf8dr/uvdva.jpg

Also, the Red is 6K but looks like far lower detail than the ursa. The Arri is just as detailed as the dragon, what is up with that? If someone has a good way to sharpen the dragon and can show it looking as detailed as the ursa mini, I'd love to learn something new. I would have thought I would see moire or aliasing in his beard on the ursa, but nothing. It just looks detailed and good. What's also eye opening, is image wise there are things I actually prefer in it over the Alexa/Dragon. Thanks Ryan for the comparison.

My very expensive Leica M-E, long renowned for their image and IQ...

Marketing hype.

This guy. Must be taking a break from exploiting third world cultures to further his career. Or maybe just in between another totally over done photoshop job that makes phase one images look like cartoons.

Also curious if this might be a manufacturing/QA issue. Are there any 4.6K owners NOT seeing magenta cast? Care to post a pic of a white wall or something?

No cast here at all.

You guys seem to be overreacting to me. If not, then there should be recalls/RMAs on the red dragon because it's got a SERIOUS green problem. Looks at his face here, it looks terrible. Ryan's URSA Mini at least makes his face look healthy and good. Even the Arri he looks a bit dead and lifeless straight from camera. But you can see if you added some color to the Arri it would probably be fairly close to the ursa, or if you pulled some color from the URSA the same. The dragon doesn't look good here. But I haven't heard of any RMA's for the very serious green issue it seems to have.



http://s9.postimg.org/qj1gzf8dr/uvdva.jpg

Also, the Red is 6K but looks like far lower detail than the ursa. The Arri is just as detailed as the dragon, what is up with that? If someone has a good way to sharpen the dragon and can show it looking as detailed as the ursa mini, I'd love to learn something new. I would have thought I would see moire or aliasing in his beard on the ursa, but nothing. It just looks detailed and good. What's also eye opening, is image wise there are things I actually prefer in it over the Alexa/Dragon. Thanks Ryan for the comparison.

It's something else then just a green or magenta cast.

It's difficult to correct because it's an uneven tint, you need to correct this with powerwindows.

This guy.

Different guy. Pretty sure THAT guy has never done product work. But Christ you made me LOL. When it comes to cartoon Photoshop work I thought Blair Bunting had the market cornered.

http://www.colormunki.com/game/huetest_kiosk

That was fun!

coloriq-1.jpg (55.37 KiB) Viewed 29572 times

coloriq-2.jpg (35.66 KiB) Viewed 29572 times

Andrew, you and Timothy must'a cheated. lol But, seriously, I completely agree with the Munsell Hue Test's assessment of my #totes biased hue discrimination.

My example above kind of surprised me. Because I've been using it for about 18 months to take some great photos ! Sorry to spam the thread with irrelevant photos, but the point I make is that with the fault above, I captured the photos below, and I'd say it's pretty hard to pick it.

Would you say that this was a part of the Leica's charm? Perhaps an aspect of its character? Something that gives it a bit of a personality? And like working with people, camera's sometimes can be divas and need a lot of attention, others don't mind being left alone, but each require a different approach to how you handle them, how you communicate with them and how you work with them? So, where some see "flaws", others see "personality?" Sure, sometimes there are flaws that make someone not worth dealing with, sometimes those can be corrected. Othertimes, it's perfectly fine and, in fact, is part of the joy of working with...a camera...or a person?

I did a test with a Rokinon 35mm pointing at a GTI Graphiclite lightbox cropping the center.
I am getting uneven color. Greener in the center, magenta toward the edges. I took the DNG into Photoshop and made two frames. One frame I did a custom white balance in the center and the other I did a white balance near the top right corner. There is vignetting, but that is normal on virtually all lenses. That I'm ok with, but the lack of color uniformity is not thrilling news Until I test it with real subject matter, it's hard to say how this will show up. It doesn't appear as severe as some of the other tests and at this point it does look correctable in post without too much fuss. But time will tell.

Just trying to help. The images out of the camera look fine to me. The dropoff to the right is just the shading of the scene, I did not set up a perfectly lit environment - just a quick test. Some people here seem as if they want something to be wrong with the camera.

Also curious if this might be a manufacturing/QA issue. Are there any 4.6K owners NOT seeing magenta cast? Care to post a pic of a white wall or something?

No cast here at all.

Hi Frank, Would you be so kind and please post a photo of a flat white evenly lit surface so we can have a reference of what a no cast image from the 4.6k can and should look like? Some of us are having a tough time achieving that. Thanks!

I've run into this issue on footage shot with wider standard speeds on a dragon sensor. It was extremely difficult to correct, never really was fully corrected.
I'm going to include some links but John is right, it's a sensor color shading issue.
My understanding is that as the angle of incidence of light relative to the photosite microlens changes it introduces crosstalk and in a bayer patern CMOS sensor with higher red sensitivity uneven red or magenta shading emerges on the frames corners and edges.
This is a known problem in sensor design and camera and sensor manufacturers address this in the sensor design stage.
Back to the older standard speed and the leica grabs of the sensor without a lens, both illustrate less direct light and show the issue at its most extreme.
I won't speculate on a possible fix as that is outside of my knowledge base but as a colorist and DP I would say that every camera has issues to some extent and a luma and color sensor character that affects the look, starting with the right camera for the job and manipulating that image to fit the final look is what experience and taste are all about. There really is no perfect standard.
Given what I'm seeing here I would definetly run my own tests before shooting with the Ursa Mini 4.6. Workarounds would include not letting skin move through the color shaded areas or not panning through flat or neutral areas where the color will show. As far as fixing this with power windows etc it's very difficult as the area is not uniform and 3d objects are moving through.

Looks like I can't post links so Google these terms for the info...

Color Shading on a Digital Sensor

Crosstalk, color tint and shading correction for small pixel size image sensor

For some perspective....

Here's a DNG from another camera. This is my Leica M-E Type 220. I just this evening shot this without a lens on the camera. It shoots DNG as well. Even though this is a current Leica camera this sensor is an old sensor design, the CCD sensor that is made by TrueSense / Kodak, the same as the digital Bolex.

I wasn't sure what to expect.

L1005515.jpg

So what's interesting to me here, is that there's a fall off into the corners AND there's a magenta cast at the bottom but a green tinge across the top. These are two separate issues, that maybe compound to look the same = magenta cast


Here's the original file.
https://drive.google.com/a/johnbrawley. ... sp=sharing

Thing is that I've been using this camera to shoot GREAT photos. Now I'm not saying that there aren't some issues that need to be worked though by Blackmagic, but even at it's worst I think you'd be hard pressed to pick this on corrected footage. My example above kind of surprised me. Because I've been using it for about 18 months to take some great photos ! Sorry to spam the thread with irrelevant photos, but the point I make is that with the fault above, I captured the photos below, and I'd say it's pretty hard to pick it.

JB

L1000185 by John Brawley, on Flickr

L1001726 by John Brawley, on Flickr

L1002378 by John Brawley, on Flickr

L1000517 by John Brawley, on Flickr

L1001720 by John Brawley, on Flickr

L1000535 by John Brawley, on Flickr

L1000574 by John Brawley, on Flickr

L1002154 by John Brawley, on Flickr

L1001487 by John Brawley, on Flickr

John, may I ask you which lens you did use for picture no. L1000574?

Jeff

But Frank as a beta tester may have the preproduction camera. Might not be conclusive about what is happening with the production cameras. It's also possible some production or beta are fine and some are not. People posting should specify if the camera is a preproduction or production camera if they know.

BMD really sent out so many cameras to beta testers, some of the testers who received their cameras later may even have production cameras. BMD knows.


Sent from my iPad using Tapatalk

John, great pictures! Loving them. But in this case it's not something I discovered by just pointing the camera to a white wall. I discovered this issue by seeing it in real life scenarios, that's why I took the test. Otherwise I probably didn't knew either.

For sure and I don't mean to discount your issue at all. If you or anyone is having issue with it then you should be going straight to Blackmagic service to discuss. But I'm seeing others posting images that have less of a severity of this issue. The point I make is that most cameras have some kind of degree of this issue. It's not exclusive to Blackmagic.

I think most cameras will fall off in the corners like this. I think some cameras have shading issues. My very expensive Leica M-E, long renowned for their image and IQ has got a similar problem. In fact it's just as bad because it's green AND magenta. It's about DEGREES of how much this problem actually affects work.

I'm not trying to downplay what might be happening. I'm not trying to dismiss at all anyone who thinks they have an issue. But I invite you to try the same test with other cameras. You might be surprised at what you find.

JB

...Like I posted before, the C100MKII has this issue, but not that strong like in some examples here in this thread. I would never accept a sensor-image, that looks like the one you posted here - even for photography. In video footage it becomes much harder to get rid of such a cast. But i am sure, BM will solve the problem. They probably need a better production end controlling...