Page 1 of 1

URSA G2 Fixed Pattern Noise

PostPosted: Fri Nov 08, 2019 9:27 pm
by winterwarrior
Hello!

I am a new owner of a URSA G2 and I've been quite excited to get one. I knew in the past that BMD has had problems with fixed pattern noise but had it fixed and when doing research before purchasing, I didn't hear anybody talk about having FPN problems.

When shooting some low light tests, I knew that 1600 ISO would be the most I could push the camera and still get a pretty clean image. When I looked at the footage, I found horrible FPN in the shadows. Even at 800 ISO, it's noticeable when viewing the BMD Film footage but is hidden when graded.

I've attached two screen shots that showcase the FPN. Look at the top right of the image. The darker one is 1600 ISO and the more apparent is 3200 ISO (I know 3200 is unusable, I've just attached it for comparison.).

This is not caused by the lighting on the building, I can see the line pattern in other footage I've shot in different locations.

Is this normal? Or did I get a botched sensor?

FPN3200.jpg
3200 ISO
FPN3200.jpg (530.25 KiB) Viewed 540 times


FPN1600.jpg
ISO 1600
FPN1600.jpg (533.16 KiB) Viewed 540 times

Re: URSA G2 Fixed Pattern Noise

PostPosted: Fri Nov 08, 2019 10:23 pm
by timbutt2
First question: have you performed a sensor calibration?

If not, then perform a sensor calibration and see if that improves anything.


Sent from my iPhone using Tapatalk

Re: URSA G2 Fixed Pattern Noise

PostPosted: Fri Nov 08, 2019 10:38 pm
by winterwarrior
timbutt2 wrote:First question: have you performed a sensor calibration?

If not, then perform a sensor calibration and see if that improves anything.


Sent from my iPhone using Tapatalk


I have done a sensor calibration. These were taken after a calibration.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Fri Nov 08, 2019 11:49 pm
by Ellory Yu
Do not underexpose with any of the BMD cameras. You can shoot low light but always with correct exposure using native ISO (ie. faster glass, lighting accordingly, etc).

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sat Nov 09, 2019 7:06 pm
by Jamie LeJeune
winterwarrior wrote: I have done a sensor calibration. These were taken after a calibration.
I would re-run the calibration and test again. That doesn't look like fixed pattern noise to me. It looks like what you get when the camera isn't calibrated after a firmware update (and you should always recalibrate after any and all firmware updates).
In any case, you can't judge noise by the log signal. You need to be transforming the signal to your display color space and gamma (most likely REC709 at 2.4 or 2.2) before making any judgement about noise. Log gamma was not designed for display, it was designed for efficient data storage.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sat Nov 09, 2019 8:37 pm
by Oyvind Fiksdal
Ellory Yu wrote:Do not underexpose with any of the BMD cameras. You can shoot low light but always with correct exposure using native ISO (ie. faster glass, lighting accordingly, etc).


Please enlighten us of this correct exposure. We are all ears.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 11:29 am
by Uli Plank
It's very simple: use false color. If areas are indicated as underexposed, i.e. dark purple, they will look bad.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 12:25 pm
by Oyvind Fiksdal
Uli Plank wrote:It's very simple: use false color. If areas are indicated as underexposed, i.e. dark purple, they will look bad.


Are you suggesting that this solve the OP’s problem?

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 2:37 pm
by John Brawley
winterwarrior wrote:This is not caused by the lighting on the building, I can see the line pattern in other footage I've shot in different locations.

Is this normal? Or did I get a botched sensor?



I don't think this is normal.

FPN tends to be vertical for starters. You have horizontal noise.

I would get in touch with BMD or your vendor. If this is a very recent buy I'd try and just do a straight swap for a new body.

I've shot a lot of G2 and I've never seen that kind of noise.

JB

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 5:34 pm
by winterwarrior
Jamie LeJeune wrote:
winterwarrior wrote: I have done a sensor calibration. These were taken after a calibration.
I would re-run the calibration and test again. That doesn't look like fixed pattern noise to me. It looks like what you get when the camera isn't calibrated after a firmware update (and you should always recalibrate after any and all firmware updates).
In any case, you can't judge noise by the log signal. You need to be transforming the signal to your display color space and gamma (most likely REC709 at 2.4 or 2.2) before making any judgement about noise. Log gamma was not designed for display, it was designed for efficient data storage.



I have run the calibration again and I am still getting the noise.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 5:36 pm
by winterwarrior
John Brawley wrote:
I don't think this is normal.

FPN tends to be vertical for starters. You have horizontal noise.

I would get in touch with BMD or your vendor. If this is a very recent buy I'd try and just do a straight swap for a new body.

I've shot a lot of G2 and I've never seen that kind of noise.

JB


I got into touch with the local vender but haven't heard back yet. I'll follow up on Monday. I also got into touch with BMD and they asked for a .braw file and am waiting for them to reply. I got the camera last Wednesday so it's very new.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 5:37 pm
by winterwarrior
Jamie LeJeune wrote:
winterwarrior wrote: I have done a sensor calibration. These were taken after a calibration.
I would re-run the calibration and test again. That doesn't look like fixed pattern noise to me. It looks like what you get when the camera isn't calibrated after a firmware update (and you should always recalibrate after any and all firmware updates).
In any case, you can't judge noise by the log signal. You need to be transforming the signal to your display color space and gamma (most likely REC709 at 2.4 or 2.2) before making any judgement about noise. Log gamma was not designed for display, it was designed for efficient data storage.



Also, even after converting LOG to Rec709, I still see the noise. I posted LOG screen shots just to show the noise pattern better.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 9:11 pm
by Michael Moore
FPN is still a general issue for Ursa mini pro G2 cameras? BM change the cameras with this problems?

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 10:09 pm
by Pete Tomkies
Is that image as is from the camera or has it been brightened up at all in post?

I have had similar FPN with my original BMCC but as stated above it isn't visible after grading.

Re: URSA G2 Fixed Pattern Noise

PostPosted: Sun Nov 10, 2019 10:18 pm
by winterwarrior
Pete Tomkies wrote:Is that image as is from the camera or has it been brightened up at all in post?

I have had similar FPN with my original BMCC but as stated above it isn't visible after grading.


The two images are straight out of camera. At 800 ISO, the FPN isn’t visible after grading. At 1600 ISO, it’s still visible and at 3200 ISO it’s extremely visible after grading.