URSA Mini Pro Survival Guide for film newbie

When I'm the colorist and editor, I don't mind the flat look either, but otherwise I would rather have it either baked in or auto-loaded so that I can take advantage of in-camera color grading for ProRes ad well as for RAW.

The fpn in the 120fps footage is most likely due to under exposure. It's easy to forget that 120 fps eith a 180 degree shutter yields a 1/240th integration time, and we tend to get used to the 1/48th that we get at 24fps.




Sent from my SM-G930V using Tapatalk

Let's not forget resolution here. Downscaling from 4.6K will hide quite a bit of FPN as opposed to a 2K window for 120 fps. After all, FPN happens at photocell level.

The fpn in the 120fps footage is most likely due to under exposure. It's easy to forget that 120 fps eith a 180 degree shutter yields a 1/240th integration time, and we tend to get used to the 1/48th that we get at 24fps.

Sent from my SM-G930V using Tapatalk

Let's not forget resolution here. Downscaling from 4.6K will hide quite a bit of FPN as opposed to a 2K window for 120 fps. After all, FPN happens at photocell level.

Both great points. I am trying to think back to whether I also may have had an ND filter engaged... in addition to the 1/240 shutter... easily might have been user error.

I do think what I've seen may be something to have BM check out because it's really clear to see the FPN in certain shots. I'd hate to have to avoid underexposing by 1-2 stops because of fear... I'd love to have the artistic choice. Granted I would totally take the beating if I was screwing around more than 4 stops under... I am waiting to hear back from one of the BM techs I spoke to two weeks ago regarding the raw clips I sent him.

I do think what I've seen may be something to have BM check out because it's really clear to see the FPN in certain shots. I'd hate to have to avoid underexposing by 1-2 stops because of fear... I'd love to have the artistic choice.

What you need to do is learn what the camera's range is, so that you can use it most effectively. Also learn to use your exposure evaluation tools, including false color and if you have one, a light meter. Also learn the Zone System; it's great tool for learning to understand exposure and ratios. I still use the Zone System when evaluating exposure; it's not really anything all that revolutionary, but learning it is a way to short cut the learning curve of understanding exposure and lighting ratios.

I'd hate to have to avoid underexposing by 1-2 stops because of fear... I'd love to have the artistic choice. Granted I would totally take the beating if I was screwing around more than 4 stops under... I am waiting to hear back from one of the BM techs I spoke to two weeks ago regarding the raw clips I sent him.

In general, BMD cameras do not tolerate underexposure, but they do respond well up to 4 stops of overexposure. Personally, I would not recommend underexposing them under any circumstance if your goal is to avoid noise (especially fixed pattern noise). I tend to shoot a stop or more over specifically for that reason. In most cases, I expose under a LUT that brings the image down about a stop and that allows me to just go by eye. Over time I'm sure you'll find your own sweet spot for the camera that balances noise and exposure.

[Tim Schumann]

The name of the LUT you use is baked into the metadata of the file.

Is that the case for both raw and ProRes recording?

Yes it is recorded to the metadata of both ProRes and RAW files.

Yes it is recorded to the metadata of both ProRes and RAW files.

Great! Besides Resolve, what post apps are able to read and display that metadata?
Few of my clients use Resolve (even though I do encourage them to use it for making dailies/proxies).

I tend to shoot a stop or more over specifically for that reason. In most cases, I expose under a LUT that brings the image down about a stop and that allows me to just go by eye. Over time I'm sure you'll find your own sweet spot for the camera that balances noise and exposure.

Is this a LUT you created, or a standard film LUT?

What you need to do is learn what the camera's range is, so that you can use it most effectively. Also learn to use your exposure evaluation tools, including false color and if you have one, a light meter. Also learn the Zone System; it's great tool for learning to understand exposure and ratios. I still use the Zone System when evaluating exposure; it's not really anything all that revolutionary, but learning it is a way to short cut the learning curve of understanding exposure and lighting ratios.

I agree totally in learning the camera's range... that is what is most exciting!

I use false color, and I am familiar with the Zone System as well. I think what I was trying to communicate in my previous post was less about my ability to use the tools to properly expose but more-so about a situation that I may find myself in that might restrict my ability to underexpose if I thought there was some value to that versus overexposing.

I tend to shoot a stop or more over specifically for that reason. In most cases, I expose under a LUT that brings the image down about a stop and that allows me to just go by eye. Over time I'm sure you'll find your own sweet spot for the camera that balances noise and exposure.

Is this a LUT you created, or a standard film LUT?

Some I created and others have been made by Kholi. I never use the standard BMD LUT because it clips the highlights quite harshly and doesn't work well at all for any scene that contains more than about 9 stops of dynamic range. Also, I tend to avoid anything described as a "film" LUT. None of the ones I've seen floating around online were designed for the source gamma and color space recorded by the Ursa Mini.

Here's a link to one I made that matches the ACES rendering of BMD 4.6K Film v3 to REC709:
https://drive.google.com/open?id=0B1bnfGXKHqpcR2lKMmUzcTNzdVU
I use that when I plan to grade in ACES.

Lately, I've been using Kholi's Comet Color LUTs more often. They are lovely, they help keep the noise down and are definitely worth the money:
https://www.facebook.com/KhoColor/

Also, an FYI if you use FilmConvert: Their transform for the 4.6K sensor uses the original film v1 color, not the v3 that's been what the firmware on the camera has run since nearly a year ago. The result is that the starting point is never accurate. I've contacted the Film Convert folks about it, but they don't seem to care enough to bother updating the camera pack for the 4.6K sensor to v3.

Giving Kholi's LUT's a shot. As always, thanks for the great insight Jamie!

[Tim Schumann]

Yes it is recorded to the metadata of both ProRes and RAW files.

Great! Besides Resolve, what post apps are able to read and display that metadata?
Few of my clients use Resolve (even though I do encourage them to use it for making dailies/proxies).

Any post apps are able to read and display the metadata... It is a question of whether they choose to add this functionality or not.

We are happy to provide them with the information on how our files are tagged.

On the subject of learning the camera's acceptable range:

I've gone ahead and done a series of unscientific tests of multiple exposure settings in LED indoor lit rooms (dark), fluorescent lit office rooms (normal), and outdoors (bright). In regards to the noise floor of my unit, I've found that using false color makes it incredibly easy to see. If I had to make a general summarizing statement about the results, I could stand firmly behind the following statement:

My Ursa exhibits vertical fixed noise within the following parameters:
- Visible on fold-out screen and monitor connected via SDI
- Visible without pressing record
- Visible at all false colors below middle grey (green)
- Very visible in the blue section
- Visible in the purple (pink?) section until stopping down and crushing the blacks in camera.... then the noise goes away completely
- Doesn't matter what iso, aperture, frame rate, resolution, codec, or shutter speed, the noise always displays based on the false color range of purple, blue, and black

It's great that the noise isn't visible in the range where dark skin should be exposed (green), but my honest question is: should the fixed noise be visible in the range immediately below middle grey (dark gray/black)? This is the range above "Near Black Detail Loss" so I imagine that a whole lot of information should be usable in this range, but would be subject to fixed noise if my camera sensor is deemed "normal."

Help me out

Screen Shot 2017-05-24 at 10.24.22 AM.png (134.54 KiB) Viewed 7519 times

I have a thread showing how I mounted this to my MōVI Pro. Not hard at all.


Sent from my iPhone using Tapatalk

One quick question : Is there any indication on the built in display when a viewer LUT is turned on? Other than cycling through the menu (or function button) is there any way to tell by looking at the display?

So I was able to mount the Ursa on the Ronin with the addition of these items:

- CineMilled Ronin Arm Extensions
- CineMilled Pan Arm Extension for DJI Ronin (Long, Gen 2)

I also purchased the Extended Tilt Arms but decided to return them, as I didn't need it to make the Ronin operational.

I am currently using the gimbal and camera in a setup with the Ronin Ring Grip with a monitor attached to the top of the ring, and gimbal feet on the bottom.

Ursa + Ronin.png (482.3 KiB) Viewed 7381 times

One quick question : Is there any indication on the built in display when a viewer LUT is turned on? Other than cycling through the menu (or function button) is there any way to tell by looking at the display?

As far as I've seen, there is no on screen indicator. I've got the on/off toggle set to one of the function buttons with the other one set to false color. Most of the time it's pretty easy to see when the LUT is engaged based on the black levels in the image, but not always. In some cases I have to toggle the function button to know for certain whether it's engaged or not.

So I was able to mount the Ursa on the Ronin

Looks great! Which model Ronin is it?

It's the original Ronin. Didn't want to wait for pricing on the R2, even though it looks pretty cool.

I am using one of Kholi's contrast Lut's and it is sometimes hard to tell without cycling. Not a big deal, just curious about it.

I'm looking to buy the KomputerBay 3700x 256gb cards as well, which aren't officially supported/approved by BlackMagic. Have you done any testing on them? How do they perform?

Asher I have yet to have any issues whatsoever with the 3700x cards. No dropped frames.

Note that my camera is most often set to UHD (3840x2160), 60p, ProRes 422 or 422LT. Not sure if any issues would derive from constant RAW recording to these cards.

Allen, if you can find a free hour, you could test them running raw up to 60 fps single card and dual card recording. I'd let each test record for 2-3 minutes and turn On the Stop Recording with Dropped Frames option so you can see how long it went before the dropped frames occurs.


Sent from my iPhone using Tapatalk

The Komputerbay 3700x 256 card will do dual card 4.6K 60fps uncompressed raw. I own 2. In single card mode it will drop frames at 11 seconds. In my tests it does 24fps 4.6K uncompressed for me in single card without dropping frames but I generally don't rely on it for actual use. I have filled cards with 4.6K 3:1 24 fps and 60 fps without issues.

Has anyone tried to use the URSA Mini with DEFY G5? Too much weight? I recently picked up a used G5 and the parameters in the firmware seem to be set incorrectly even for use with a Micro Cinema Camera. I tried contacting the support email for their company but have yet to hear back.

Are there any G5 users who might be willing to share their parameters or any experience changing the parameters?

Hey, I own 2x G5's and am in the process of selling one of them.

The support from Defy is virtually non-existent at the point, unfortunately. If you have a working unit you can connect to the software and pretty much tune the gimbal to your hearts content. I've done this for about 3 years now and have loved using the G5 as to me it is much smoother than any DJI gimbal out there.

I haven't tried (nor do I think I will try) the Ursa on the G5 as the rated max weight is between 6-7lbs.

If you've gotten the Ursa on your G5 and the only issue is the power to the motors, I can possibly talk you through the tuning software... The motors are really powerful so it may be able to work. Let me know.

Hi Allen,

I downloaded the software; simple bgc32 v2.63. It says the firmware in the defy is up to date, the last owner must have updated it, but not loaded the profile. It seems well balanced before I turn it on, then once it is on it moves the camera to a different position. I am using a dslr at this point just to try and get it working. Is there a parameter file someplace that I can download?

There are so many options to adjust in the software, I was trying to look for a place to start. The PID controller seems to make a difference.

There is a parameter file.... I am away from my studio so I'd have to check to see how mine is configured...

From what I remember, you have to use the simple bgc version that matches the firmware on your gimbal, and that would depend on when the gimbal was manufactured. For example the version that works with my gimbal is 2.41 but nothing after that will work. To figure out which software to run you can email Defy and also, they can give you a profile to load to match your camera's weight. I've gotten one from them for an a7s and for a 60D.

You are correct that the PID section is the most important section. I've been able to get the best results just focusing on the P and D settings. I typically set the P of each motor (one at a time) as high as I can get them without the motor oscillating or overloading, then back off by between 5-10. That way the power is kicking in pretty high whenever the gimbal detects any drift and it results in a "locked in" movement. You could also back those off for a more "loose" movment.

The I setting is for how quickly the gimbal will turn when you input force so I typically leave that at the default setting. Raising D will help remove minor rumblings and minor input errors. I try to set this as high as possible without the gimbal freaking out. Start here if the gimbal is erratic.

Also another big thing I've experienced but never read anywhere is when I've got the gimbal on a tuning stand or c-stand the settings will vibrate at lower values than when handheld, so remember to test your settings with the gimbal in your hands before finalizing your settings.

Thanks Allen, that's very helpful. I didn't get a response from their email support so I have gotten pretty far adjusting settings.

The name of the LUT you use is baked into the metadata of the file.

Hi Tim. I've searched high and low through the manual, but I can't find a way to reveal this baked in LUT metadata in Resolve. Can you please explain how to see this info in the media pool? There appears to be no option available in the metadata viewer that reads the LUT applied in camera. Any help here would be much appreciated.

metadata_options.png (63.12 KiB) Viewed 6859 times

[Tim Schumann]

• Open Resolve 12.5 or 14 Beta 4
• Go to media window
• Drag clip into media pool
• Select clip
• Ensure metadata window is showing on the right hand side by clicking 'Metadata'
• Just under that top right there is a dropdown with an up arrow and a down arrow with a hyphon in between them. That lets you select which type of metadata you want to see. Hit that and select 'Camera' or 'All Groups'
• Scroll down to the field that says 'LUT Used' and your LUT should be listed here.

Now I see it. Thank you Tim!

One more question: Is there a way to get that "LUT used" as a column in the column layout?
It would be useful to be able to sort a set of clips by the LUT used on set so that I can quickly make a separate timeline for each set and add the correct LUT to a timeline node.

It's great that the noise isn't visible in the range where dark skin should be exposed (green), but my honest question is: should the fixed noise be visible in the range immediately below middle grey (dark gray/black)? This is the range above "Near Black Detail Loss" so I imagine that a whole lot of information should be usable in this range, but would be subject to fixed noise if my camera sensor is deemed "normal."

@Allen Ross - Delighted to see this broken down so clearly. I would love an answer to this too. I know the UMP isn't as low light camera but I'm seeing a lot of FPN on my new model. I will try overexposing with a LUT to compensate for now but I'm a little worried about just how bad it is.

Anything to report re: FPN issues since this post?