URSA Mini 4.6k The Original™ Magenta Cast Issue

here is my quick and dirty test.

What F stop did you shoot this test?

I can still see magenta on the right side...

Happy Birthday Benton!!

Same on Rick's tests...

Just did a quick test. Judge yourself, but keep in mind that this is an extreme situation that nobody would shot.


Settings: ProRes 422 2160p50 90° ISO200 - Sigma 18-35mm f1.8 @ 35mm

at f16



at f8

Just did a quick test. Decide yourself for the result.
Settings: ProRes 422 2160p50 90° ISO200 - Sigma 18-35mm f1.8 @ 35mm

at f16



at f8

Yep, definitely an improvement.

"Extreme situation... nobody would shoot" u being ironic I hope

Joshua,

I would like to know if BMD is going to continue to work towards eliminating this as it is still evident in the footage that this problem still exist even after the new firmware 3.3

Perhaps it is an improvement but the fact is that this magenta issue was not eliminated entirely.

the fact that there are cameras out there that do not exhibit this at all is completely unfair to all the individuals who have units that are "faulty" .

I myself opted not to return my second unit with the hopes that BMD will be able to fix this and I am not a happy camper.


Eli.

My prior quick test had no colour correction. These are the same raw open gate shots at T8 ND 0.4 (one hour before sunset, north sky light overcast followed by south by south-east sky) with Saturation at 100% which may be considered extreme:

Sat 100% 3.3

Sat 100 3dot3 2016-06-14 at 5.36.24 AM.png (999.13 KiB) Viewed 9730 times

Sat 100% 3.3

Sat 100 3dot3 2016-06-14 at 5.37.35 AM.png (986.81 KiB) Viewed 9730 times

That looks pretty bad to me to be honest.

I find it hard to believe and I sincerely hope to be wrong, that this thing can be fixed via firmware.
Too many variables, too different from camera to camera.
All that without considering the fact that there're supposed to be good clean cameras out there.

I hope to be wrong... it's been a month already since I sent my camera back.

the fact that there are cameras out there that do not exhibit this at all is completely unfair to all the individuals who have units that are "faulty" .

I have not seen any examples of footage shot with a Sigma or Zeiss at F8 or higher without the issue. Show me this is a "fact".

Thanks for the update, but according to the screenshots posted here: it's still there.

Guess I'll have to wait for month again before buying one, if I still want to :/

Damn, I really want to buy and enjoy this one, but seriously I won't pay that amount of money for uncertain results.

Saturation 100% appears to be too strong, but here are the same shots at Saturation 75%. Best? Better? Good?

Sat 70 3dot3

Sat 70 3dot3 2016-06-14 at 5.59.28 AM.png (966.29 KiB) Viewed 9752 times

Sat 70 3dot3

Sat 70 3dot3 2016-06-14 at 6.00.08 AM.png (954.02 KiB) Viewed 9752 times

Saturation 100% appears to be too strong, but here are the same shots at Saturation 75%. Best? Better? Good?

It's obviously still there, mainly on the top.

And finally, let’s set Saturation 70 and Contrast 1.3. Looks reasonable to my eye. But when you push harder, not likely acceptable. The parade and waveform are included in each quick test posted. Scroll down or click on the image if you don’t see them.

Although the same adjustment is applied, each shot of the sky was in a very different direction, so it appears that each shot needs to be tweaked rather than apply the same grade across both shots. White balance was set at 5600K in camera and 6000K in the Camera Raw tab. These two images looked fine setting the project white balance to 11000. I didn’t think to meter the colour temperature last evening.

Sat 70 Con 1.3 3dot3

Sat 70 Con 1.3 3dot3 2016-06-14 at 6.14.59 AM.png (992.55 KiB) Viewed 9724 times

Sat 70 Con 1.3 3dot3

Sat 70 Con 1.3 3dot3 2016-06-14 at 6.15.41 AM.png (966.65 KiB) Viewed 9724 times

[Joshua Helling]

Hey Eli,

This is indeed just the first step. We are continuing to work on improvements. This update should be helping a lot with the symmetrical magenta shading, but not so much with some of the others.

As we've mentioned before there were a few different issues that were getting lumped into one big one (part of what made it hard to identify). So we'll keep working on improvements.

We'd really like to see the feedback on this update.

The new camera will be affected with this magenta issue, or can you fix the new camera ?

Hey Eli,

This is indeed just the first step. We are continuing to work on improvements. This update should be helping a lot with the symmetrical magenta shading, but not so much with some of the others.

As we've mentioned before there were a few different issues that were getting lumped into one big one (part of what made it hard to identify). So we'll keep working on improvements.

We'd really like to see the feedback on this update.

Hi Joshua,

I'm really digging this new policy of being very responsive! I think we can all see the positive effect it has had on the tone here (and on BMCUser). It's also great that this new firmware has been released so quickly.

I had mistakenly thought that this update would attempt to solve the issue entirely, and my concern now is not knowing how long this will take to fully work out. I know you've been sharing what you can, but can you tell us what the few different issues were and maybe share what the engineers are planning to do about them? Even if they don't have a fully formed plan yet, just a rough timeline would be amazingly reassuring.

This is the most expensive camera purchase I've made (with my own money), and I'm sure you can understand it makes me nervous if I don't have any reassurances that this will be all worked out in a reasonable amount of time (I use f/8 and f/11 for daytime establishing shots and landscapes pretty often).

Thanks!

[Joshua Helling]

The new camera will be affected with this magenta issue, or can you fix the new camera ?

Well, any of the cameras can get this firmware, but obviously new cameras will have the new firmware. This will be the same for future updates as well. As a general, we make sure that any product that ships out, ships with the most recently installed firmwares.

The new camera will be affected with this magenta issue, or can you fix the new camera ?

Well, any of the cameras can get this firmware, but obviously new cameras will have the new firmware. This will be the same for future updates as well. As a general, we make sure that any product that ships out, ships with the most recently installed firmwares.

My question was not good, apologies. What i ask is : do you intend to make hardware fix on the new camera to avoid this Magenta issue ?

Was that frame-shaped object coming from the lower-right corner made from magenta material, or was it the sensor again?

yes i do see some magenta cast on the bottom there. although there is a purple book on my table there that could be reflecting the color onto the wall.

here is my quick and dirty test.

What F stop did you shoot this test?

I'm at f7 or somewhere around there on my zeiss 21mm.

[Joshua Helling]

Hi Joshua,

I'm really digging this new policy of being very responsive! I think we can all see the positive effect it has had on the tone here (and on BMCUser). It's also great that this new firmware has been released so quickly.

I had mistakenly thought that this update would attempt to solve the issue entirely, and my concern now is not knowing how long this will take to fully work out. I know you've been sharing what you can, but can you tell us what the few different issues were and maybe share what the engineers are planning to do about them? Even if they don't have a fully formed plan yet, just a rough timeline would be amazingly reassuring.

This is the most expensive camera purchase I've made (with my own money), and I'm sure you can understand it makes me nervous if I don't have any reassurances that this will be all worked out in a reasonable amount of time (I use f/8 and f/11 for daytime establishing shots and landscapes pretty often).

Thanks!

You're right. I've tried to keep the details a little light until we had a better idea of what we were going to need to do on our side in order to get this completely addressed.

Today's release, is meant to address the vignetting magenta. Or what what we're calling the symmetrical magenta. Obviously this update had many other improvements as well but those aren't really the focus.

Another issue is the asymmetrical magenta cast. This is where certain sides or areas will have more of a magenta cast than the others. This is what we're looking at right now. I can't really talk about what the plan is right now as we're still working on it.

As for rough timeline. That's tougher. It's going to be sooner than you'd probably expect, but it won't be tomorrow. But again, we're still working out the details which of course I will share with you the second my leash is off.

So yeah, this update was definitely not the "drop the mic" update, where champagne and rose petals fall from the ceiling, all issues fixed. But it is a really solid first step. We'll keep making it better.

So yeah, this update was definitely not the "drop the mic" update, where champagne and rose petals fall from the ceiling, all issues fixed. But it is a really solid first step. We'll keep making it better.

Hey Joshua,

thank you for your open words. In my opinion the solid first step was, to take the community concerns serious. The discussions and rumors reached a level that was really redicoulous and wasn't good for BMD and the Users.

Greetings Joerg

[Joshua Helling]

The new camera will be affected with this magenta issue, or can you fix the new camera ?

Well, any of the cameras can get this firmware, but obviously new cameras will have the new firmware. This will be the same for future updates as well. As a general, we make sure that any product that ships out, ships with the most recently installed firmwares.

My question was not good, apologies. What i ask is : do you intend to make hardware fix on the new camera to avoid this Magenta issue ?

Ah! No problem. I understand what you mean now. I'm leery to answer a question put like this, because it would be very easy to take any answer I give as an admission, or indication of some larger problem. This is a potential trap and my inclination is to avoid it.

However, if everyone promises to read my answers and try to avoid, sky is falling conclusions. I will try to answer as best as I can.

READ!!!!: We don't think there needs to be a hardware change. Let me say that again. We don't think we need a hardware change. That said, if there did need one we would change it in future cameras.

So to directly answer your question. No.

To indirectly answer it, we know there is still work to be done, but we are confident in our ability to improve the issue without any sort of retooling. Additionally I will let you all know what those next steps are as soon as I am able to.

I hope that answer helps.

[Joshua Helling]

Hey Joshua,

thank you for your open words. In my opinion the solid first step was, to take the community concerns serious. The discussions and rumors reached a level that was really redicoulous and wasn't good for BMD and the Users.

Greetings Joerg

Thanks Joerg!

I agree that the discussions started to turn very toxic for the community. We were absolutely looking into the issue (because we do read these forums). We definitely could have been quicker to jump in. This issue was particularly difficult because the issues kinda overlap and it took us time to sort through it all in order to have any "meaty" (i'm gonna trademark that) information to give you.

But alas, here we are.

The new camera will be affected with this magenta issue, or can you fix the new camera ?

Well, any of the cameras can get this firmware, but obviously new cameras will have the new firmware. This will be the same for future updates as well. As a general, we make sure that any product that ships out, ships with the most recently installed firmwares.

My question was not good, apologies. What i ask is : do you intend to make hardware fix on the new camera to avoid this Magenta issue ?

EDIT: answered. Thanks Joshua.

Joshua, much appreciate the efforts of the team working on this. The current correction is interesting in that it has gone a long way to removing the magenta colour, but I'd like to comment on the last screenshots I posted above. To me, you can still see there is something not quite right as if the rightmost 10% of the frame from top to bottom has a different tone than the adjacent part of the frame. It's like there's evidence or a footprint left behind even though the magenta colour is not evident. Very strange and I'm sure not easy to analyze in terms of correcting that faint but apparent vertical band on the right. Not very precise in my description, but that's my feedback on the first step forward to correct the problem.


Sent from my iPad using Tapatalk

Rick, if you have a raw frame to share, I'd love to pull it into Resolve to view. It would be much easier to evaluate that way. Thanks!

I like the "drop the mic" line
LOL.

Thanks for being candid and honest.

Rick, if you have a raw frame to share, I'd love to pull it into Resolve to view. It would be much easier to evaluate that way. Thanks!

Here are seven DNGs from the quick clips I shot last night. You’re interested in those files that begin with "RickLang_3.3.”

https://www.dropbox.com/l/scl/tHvkei1Pwj7I8jh1hC9yfs

Okay, here’s a stab at grading this image and I like the way it looks on my iMac 5K Retina fullscreen from the original 4608x2592 raw. i know you can’t see that resolution from this tiny upload. Don’t know how you feel about it, but it made me smile that BMD is getting close to a solution. Actually decided the tree branches would look better with no change to contrast, but this fine tuning has to stop!

Temp 5200K Con 1.2 Sat 70% Tint -10 Boost 20 3dot3

Graded 3dot3 2016-06-14 at 10.58.57 AM.png (972.26 KiB) Viewed 9358 times

[Joshua Helling]

This goes more into the asymmetrical magenta cast, which is what we're working on now! This one is a little harder because this varies between cameras. But rest assured, we are working on this.

Thanks Joshua!


Sent from my iPhone using Tapatalk

The newer one has virtually no magenta bias or magenta corners. But I must say that this most resent camera is looking outstanding!

Benton, how new is your newer camera, when did you order it and from what dealer?

I did a test with the two UM4.6k cameras that I luckily have on hand, both have the new 3.3 firmware installed. The slate in the photos say either "Magenta Camera" or "Non-Magenta Camera" depending on the camera used for that particular frame. The Magenta camera (serial # 2911458) I've had for several months now and is itself a replacement to the first camera that I originally received. The "Non-Magenta Camera" Is a replacement to the second camera, which also had FPN, it's serial # is: 2996335. One camera will be going back to BMD.

Both cameras were set to ISO 200, FULL RAW, Shutter angle 11.25 and a WB of 5600k. No filters were used and a f/T 8 was set on a Nikon 28mm AIS and a Rokinon 35mm Cine DS.

Here is a link to the .DNG's from the two different UM4.6 cameras. https://www.dropbox.com/sh/a5l2g3hfzyv4 ... lPiTa?dl=0



The jpegs seen below were processed in Capture One Pro 9 as they load in without any correction applied. When white balance is applied (as sampled from the white slate), the "Magenta Camera" frames shift dramatically much closer to neutral, where the "Non-Magenta Camera" frames shift only a tiny bit, meaning that it's already close to a perfect WB balance already. But even the "Non-Magenta Camera" is not perfect and still shows some non-uniform color in those same areas, but I think
the differences are pretty livable and could be selectively adjusted much easier than the often severe magenta seen in the other camera. Also, shooting at wider apertures should also help to minimize the non-uniform color.

Does this answer the question of wether there are good or bad UM4.6 cameras out there?

Blackmagic URSA Mini_1_2016-06-14_1146_C0091_000054.jpg (768.02 KiB) Viewed 9301 times

Blackmagic URSA Mini_1_2016-06-14_1150_C0021_000095.jpg (698.59 KiB) Viewed 9301 times

My two camera testing setup:

IMG_2576.JPG (246.41 KiB) Viewed 9285 times

Third time’s the charm! Looks quite good and may be getting better with the next firmware updates. Thanks for the update, Benton. You must be much happier with the results on the new camera. Also the serial numbers you referenced are all higher than mine which is far better than the Magenta Camera so whatever the problem, it doesn’t seem to be closely related to serial numbers, like a single bad batch of sensors.

Does this answer the question of wether there are good or bad UM4.6 cameras out there?

Blackmagic URSA Mini_1_2016-06-14_1146_C0091_000054.jpg

Blackmagic URSA Mini_1_2016-06-14_1150_C0021_000095.jpg

Why is the top image not white balanced? You say you balanced to the slate but it looks nothing like the bottom slate.

When I open your two images in photoshop I get this. That portion of the road almost looks static and as if its slightly framed out in the second one, but they're not that different from each other. Although in the second image the stuff in the background is more out of focus, so looks like a different aperture setting?





This "new" image at the bottom actually has more magenta on the bottom left than the top one. Neither one is clearly better than the other.

Steven, I think Benton mentioned the photos posted had no correction applied. The comment about white balancing was information only, not reflecting what was posted. Your correction shows the large shift to white balance the Magenta Camera even though both cameras have 3.3 firmware installed.

Steven, I think Benton mentioned the photos posted had no correction applied. The comment about white balancing was information only, not reflecting what was posted. Your correction shows the large shift to white balance the Magenta Camera even though both cameras have 3.3 firmware installed.

The "correction" was just white balance, the temperature was adjusted more than the tint. White balance is something you should do with every image. It's like basic primaries in Resolve.

I thought the point was showing uneven magenta, we already know cameras vary in color temperature but that's easy to solve. The real issue is uneven magenta.

Does this answer the question of wether there are good or bad UM4.6 cameras out there?

Blackmagic URSA Mini_1_2016-06-14_1146_C0091_000054.jpg

Blackmagic URSA Mini_1_2016-06-14_1150_C0021_000095.jpg

Why is the top image not white balanced? You say you balanced to the slate but it looks nothing like the bottom slate.

When I open your two images in photoshop I get this. That portion of the road almost looks static and as if its slightly framed out in the second one, but they're not that different from each other. Although in the second image the stuff in the background is more out of focus, so looks like a different aperture setting?

I said the images posted here are NOT white balanced and were processed as they loaded in. I said that WHEN I did a WB on the slate the Magenta camera shifted dramatically, which accurately tells me that the WB was way off to begin with. I think my focus was off in one of them when I changed the lens. Yes, when you do a white balance on the magenta camera, it gets pretty close to the other, but it is still better in the "Non-magenta" camera especially when that also has it's WB tuned. I think any camera should deliver a clean WB out of the camera when set on the proper color temperature. If your shooting on ProRes with everything baked in and need to hand the files over to someone else, the last thing you need is to have to do an additional WB before delivering it.

I agree. The 3.3 firmware concentrated on the magenta corners, but future change(s) will address the uneven magenta according to Joshua. Benton and I both have some uneven magenta although my camera isn’t too bad with a normal good exposure, but my images fall apart unevenly when pushed beyond the adjustments I cited earlier.

Shot with the new 3.3 firmware. Two frames with a Sigma 50mm Art lens, at F8 ISO1600 for the white ceiling, and F16 ISO 200 for the sky.

These have both been pushed in the grade to show the problem, but it does really show the problem.

Doesn't look so bad on the sky. Have to really push the grade a ridiculous ton to get bad magenta in the cloud. But the white ceiling is awful. What do you guys think? ISO related?

50mm Art F8 ISO 1600 DNG.jpg (179.43 KiB) Viewed 9235 times

50mm Art F16 ISO 200 DNG.jpg (134.94 KiB) Viewed 9235 times

James, you’ve changed two items that possibly affect the magenta behaviour: the iris setting and the ISO setting. You might get an answer if you keep the iris setting constant and play with the shutter angle as you change from ISO 200 to 400 to 800 to 1600. Do you think you could reshoot? If not, I could test changing ISO and shutter angle (so changing ND isn’t a factor). All my tests are always ISO 800 with 180 degree shutter angle. I need to walk on the wild side!

It's possible it's not ISO related, but that the F16 setting introduces some diffraction which might reduce the problem. I think the key here is the famous F8 setting is really still making huge problems even with the new firmware.

I'll try repeat the tests more thoroughly when I get the chance to, changing only the shutter angle. Probably have to go from 360º to 11º or something

To reduce the magenta cast does F2.8 perform better than F8 ?

Marc, it appears that the cast is only a problem beginning at T8 on most every camera that has a problem. So any stop below T8 may be fine. Certainly T5.6 or T4 and wider look clean on my camera depending upon the shot at times. Others may have a better or worse experience.

Yes, when you do a white balance on the magenta camera, it gets pretty close to the other

Yes, and as many here including yourself have posted, the issue is not an even cast. So when you post "Does this answer the question of wether there are good or bad UM4.6 cameras out there?" it doesn't. Because you yourself have said before that an even cast is not a problem since it's easily solved.

but it is still better in the "Non-magenta" camera

I disagree there. The new one still has magenta on the right in basically the same place but also has it worse on the left bottom side. Overall it's about the same, if not worse because of the left side. To be fair I say so only based on the dngs you have posted.

Yes, when you do a white balance on the magenta camera, it gets pretty close to the other

Yes, and as many here including yourself have posted, the issue is not an even cast. So when you post "Does this answer the question of wether there are good or bad UM4.6 cameras out there?" it doesn't. Because you yourself have said before that an even cast is not a problem since it's easily solved.

but it is still better in the "Non-magenta" camera

I disagree there. The new one still has magenta on the right in basically the same place but also has it worse on the left bottom side. Overall it's about the same, if not worse because of the left side. To be fair I say so only based on the dngs you have posted.

You're right Steven, the issue is the uneven cast. But at the same time there is clearly a large sampling of UM4.6k cameras that have a strong magenta bias out of the camera and this new one at least gets the WB way closer than any of the others I have had. It's far from perfect as I have pointed out, but I think it's workable in post. So In my estimation having one camera delivered that gives a pretty good ballpark white balance against one that is FAR from the mark, does qualify as one being good and the other bad. Both these cameras should look very close to each other and they don't.
We will have to agree to disagree on wether the Non-Magenta camera is better or not. I have played with the white balance in both versions with the RAW files within Resolve and I'm able to get an overall much cleaner balance with the "Non-Magenta Camera" frames.

*Addendum.. After spending at least an hour on correcting both the "Magenta Camera" 28mm frames and the "Non-Magenta Camera" frames, I have concluded that the Non-Magenta Camera produces an image that is far superior in balance and subtlety that is much easier to fully correct than it's "Magenta Camera" counterpart.

If I knew Resolve better, I could probably achieve a better result, but the uneven magenta cast is fairly tame here I think:

Blackmagic URSA Mini_1_2016-06-14_1150_C0021_086807.jpg (575.01 KiB) Viewed 8869 times

(Sorry about the focus!)

Steven, Do you have a UM4.6 camera that is free from uneven color? I'd love to see some frames if you do.

Would be great if BMD share WHY is this problem occurring and come clear. I'm a long time Reduser, been through some problems with their cameras, but every time something huge like this happened they always came up with an explanation on why is it happening. This is very important for building trust with your customer base imo. Not just fixing, but explaining.

Quick 40 second Interview today for an internal Facebook piece. This camera produces beautiful images, and looking forward to all the magenta being sorted out. Firmware 3.3.



Sent from my iPhone using Tapatalk

Quick 40 second Interview today for an internal Facebook piece. This camera produces beautiful images, and looking forward to all the magenta being sorted out. Firmware 3.3.



Sent from my iPhone using Tapatalk

Sent from my iPhone using Tapatalk

Thanks Benton for all the time you put into testing the cameras!

That is a pretty big difference in white balance/tint between two cameras of the same make basically. Of course in no camera every sensor is exactly the same, but from years of multi camera productions (where of course you always have to apply a little bit of correction between cameras) I have never ever seen anything that drastic.

But it's still kind of unfortunate to have not one single official explanation as for why this occurs. The only thing so far is the rumor of the "not fully dissipated sensor protection" - but what to make of that, who knows. It's really sad that there are only rumors and no official explanations.