Blackmagic Forum

Hi,
I just bought a second-hand URSA mini 4.6k and went out to shoot some test scenes. However, I found the fixed pattern noise is really really horrible. I Is that the camera already old enough to consider as a broken/bad camera?

it is what it is..... ive sent mine in, support rated it as "within the standards" what im seeing in my UMini looks like yours in ISO1600 window low light.

Wilson,

is this grab in log that your showing or have you tried to grade/lift the blacks? If this is still in log once you take your blacks down to a normalized setting I think you will mostly see this disappear.

Was Black Calibration done?

Are you shootig at ISO 1600 with a long, slow zoom, and then lifting your exposure in post? There’s gonna be noise. Stop saying FPN. This is just noise because the footage is horribly underexposed. This is why they make plugins like Neat Video. But you need faster glass. Ursa Mini is not a low light camera like an A7S. It’s a normal camera that takes exquisite images when properly composed and lit.

This is nothing new.

Where is the FPN?

thomas bruegger wrote:it is what it is..... ive sent mine in, support rated it as "within the standards" what im seeing in my UMini looks like yours in ISO1600 window low light.

I think those standards allow a pretty wide range. Just yesterday I tested a brand new Ursa Mini Pro (calibrated with latest firmware) against an original non-pro Ursa Mini 4.6K. Same settings, same lens, same light and charts for each camera. Noise on the Pro exhibited significantly more of a fixed pattern in comparison to original 4.6K. I'm sending comparison files into BMD today. Will see what they say.

thomas bruegger wrote:it is what it is..... ive sent mine in, support rated it as "within the standards" what im seeing in my UMini looks like yours in ISO1600 window low light.

Oooops...
Just sent them DNGs for further investigation. I’m worrying now

Ryan Hamblin wrote:Wilson,

is this grab in log that your showing or have you tried to grade/lift the blacks? If this is still in log once you take your blacks down to a normalized setting I think you will mostly see this disappear.

Jason R. Johnston wrote:Are you shootig at ISO 1600 with a long, slow zoom, and then lifting your exposure in post? There’s gonna be noise. Stop saying FPN. This is just noise because the footage is horribly underexposed. This is why they make plugins like Neat Video. But you need faster glass. Ursa Mini is not a low light camera like an A7S. It’s a normal camera that takes exquisite images when properly composed and lit.

This is nothing new.

It’s untouched ProRes. The fact is I also own Panasonic HMC-150 and Sony Z5P which both about 10years old. I personally consider them as rubbish quality. But both that two camera perform better in low light compare to my URSA. That’s why I said it’s ridiculous.

JasonFinnigan wrote:Was Black Calibration done?

Yep

Wilson, the Ursa is not a “Video”Camera, but rather a Cinematography camera and the sensor needs a minimum level of light for best results. When your scene is properly lit, the camera will give excellent results. But a “low light” camera it is not. If low light on location shooting is what you do, then you,should look at the Panasonic EVA1 instead, or get a camera designed for low light shooting.
Cheers

Denny Smith wrote:Wilson, the Ursa is not a “Video”Camera, but rather a Cinematography camera and the sensor needs a minimum level of light for best results. When your scene is properly lit, the camera will give excellent results. But a “low light” camera it is not. If low light on location shooting is what you do, then you,should look at the Panasonic EVA1 instead, or get a camera designed for low light shooting.
Cheers

Yes, I understand this very much. What I am trying to do is to use this on stage/studio multi cam live feed. And the image I uploaded above is a stage scene. I saw many success cases which cause me to doubt my camera is performing abnormally. Here is one with similar light conidition:

Wilson Luniz wrote:

Denny Smith wrote:Wilson, the Ursa is not a “Video”Camera, but rather a Cinematography camera and the sensor needs a minimum level of light for best results. When your scene is properly lit, the camera will give excellent results. But a “low light” camera it is not. If low light on location shooting is what you do, then you,should look at the Panasonic EVA1 instead, or get a camera designed for low light shooting.
Cheers

Yes, I understand this very much. What I am trying to do is to use this on stage/studio multi cam live feed. And the image I uploaded above is a stage scene. I saw many success cases which cause me to doubt my camera is performing abnormally. Here is one with similar light conidition:

We've filmed parts of a few outdoor festivals using the Ursa mini 4k, BMPC 4k to use for promo commercials, wrap up video etc with great success. However the Screen grabs you posted are much darker than an typical large scale festival/concert. They are usually very bright. The Ultra music fesitval video you posted is much closer to the light levels we've dealt with at festivals, you can't do that with a few Par cans.

This sort of content would never be displayed to the end user as 'film' which is what you have posted above. 'Film' is intended for providing the most information possible in post so that it can be color graded, not intended as a display or delivery format.

In fact as discussed here the Ultra Music Festival uses URSA Mini and URSA Mini Pro cameras to produce their content. The URSA Mini is more than capable of what you are describing but you need to look at your equipment and the process you are using rather than labelling our cameras as ridiculous and claiming that your unit has FPN.

The cameras you are comparing it to that you have mentioned are 8-bit video cameras with small sensors and inbuilt lenses that can only shoot 1080 as a maximum resolution... and you are comparing them to a 4.6K camera that, has interchangeable lenses, can shoot RAW or ProRes and you are looking at your images in 'Film' mode. You are not comparing apples with apples.

As Jason F pointed out these are very capable cameras when used correctly.
If you are often shooting in dark environments then as Jason J mentioned you may want to look into using some faster lenses with a low f-stop.

If you want to produce video content in a live environment then you need to either set your camera to 'Video' dynamic range, get someone to help you create a 3D LUT in Resolve that gives you the gamma settings that you are after, or if you are using it in conjunction with an ATEM switcher in a live environment you can use ATEM's Camera Control software to adjust black levels and contrast to achieve the look you are after.

Tim Schumann wrote:This sort of content would never be displayed to the end user as 'film' which is what you have posted above. 'Film' is intended for providing the most information possible in post so that it can be color graded, not intended as a display or delivery format.

In fact as discussed here the Ultra Music Festival uses URSA Mini and URSA Mini Pro cameras to produce their content. The URSA Mini is more than capable of what you are describing but you need to look at your equipment and the process you are using rather than labelling our cameras as ridiculous and claiming that your unit has FPN.

The cameras you are comparing it to that you have mentioned are 8-bit video cameras with small sensors and inbuilt lenses that can only shoot 1080 as a maximum resolution... and you are comparing them to a 4.6K camera that, has interchangeable lenses, can shoot RAW or ProRes and you are looking at your images in 'Film' mode. You are not comparing apples with apples.

As Jason F pointed out these are very capable cameras when used correctly.
If you are often shooting in dark environments then as Jason J mentioned you may want to look into using some faster lenses with a low f-stop.

If you want to produce video content in a live environment then you need to either set your camera to 'Video' dynamic range, get someone to help you create a 3D LUT in Resolve that gives you the gamma settings that you are after, or if you are using it in conjunction with an ATEM switcher in a live environment you can use ATEM's Camera Control software to adjust black levels and contrast to achieve the look you are after.

I knew this VERY MUCH. The fact is I CAN'T get a good result after using any LUT or Davinci Resolve(If you need I can post some grading image and how I grade). I post a untouched image because everybody can try to import the image into any video editor and try to grade it yourself. And my conclusion is, I'm not satisfied even after grading. I know a 8-bit 420 camera is heavily pre-process inside the machine. But if I can't get satisfaction after post processing with URSA, then why I need URSA? Please, I'm not newbie. But maybe you're more experienced. So please let me know what post processing I can do to get rid of these problems?

Wilson,

have you tried letting bit more light on your camera sensor and if you needed darker more contrasty look, bringing it bit down in post and maybe shooting at ISO 400 if it is ProRes?

Its like pulling teeth.
Under the "Dynamic range" setting in the main menu did you shoot on "Film" or "Video" ?

Darko Djerich wrote:Wilson,

have you tried letting bit more light on your camera sensor and if you needed darker more contrasty look, bringing it bit down in post and maybe shooting at ISO 400 if it is ProRes?

Yes, I did. Bringing it down or use lower ISO will certainly reduce noise. However this will result darker image that is not quite usable.

Stephen Press wrote:Its like pulling teeth.
Under the "Dynamic range" setting in the main menu did you shoot on "Film" or "Video" ?

Yes, the image I uploaded above is untouched "Film". And I just tried to apply "URSA mini 4.6k to Rec709" and "URSA mini 4.6k to Video" LUT in Davinci Resolve. It is even worse because the color noise looks more conspicuous.

Wilson Luniz wrote:

Darko Djerich wrote:Wilson,

have you tried letting bit more light on your camera sensor and if you needed darker more contrasty look, bringing it bit down in post and maybe shooting at ISO 400 if it is ProRes?

Yes, I did. Bringing it down or use lower ISO will certainly reduce noise. However this will result darker image that is not quite usable.

Based on your responses here and image shown, it appears that your lenses are either not set to right exposure or not fast enough to deal with that scene.

What lenses were you using Wilson? And what is the max F/stop for them?

I bet you had "DETAIL SHARPENING" set to on when recording this.

And you had a very low light situation and/or slow lens when filming in ISO1600 and 360° shutter resulted in such dark images.

I have attached a frame from a privat concert I filmed in film log, ISO800, and 360° on a F2.8 lens and a windowed sensor.
If I hadn't luckily brought with me a Dedolight DLED7D with 90W there wouldn't be much to see because they had only dim roomlight in the castle's ballroom.

Most professional lit concerts I have filmed in my life can be shot at ISO400/800 with 180° shutter and F4-5.6

Johannes Jonsson wrote:What lenses were you using Wilson? And what is the max F/stop for them?

A Fujinon 8-120 b4 lens with 2x extender(I forgot the exact model, get you back later). The F/stop limited to f/4.0 after applying 2x extender(and I'm using f/4.0).

Robert Niessner wrote:I bet you had "DETAIL SHARPENING" set to on when recording this.

And you had a very low light situation and/or slow lens when filming in ISO1600 and 360° shutter resulted in such dark images.

I have attached a frame from a privat concert I filmed in film log, ISO800, and 360° on a F2.8 lens and a windowed sensor.
If I hadn't luckily brought with me a Dedolight DLED7D with 90W there wouldn't be much to see because they had only dim roomlight in the castle's ballroom.

Most professional lit concerts I have filmed in my life can be shot at ISO400/800 with 180° shutter and F4-5.6

Nope, I didn't turned it on. I'm familiar with the menu more then my girl friend!(even when I haven't bought it!!).

Hmm...I'm just picking a scene with pure black background as a reference. Actually some other scenes suffer from the same problem(just not that such conspicuous). See my next post.

And I enlarge your image again and again. Looks way better compare to mine. I'm highly suspect my camera is a special case.

Another reference, see attachment

Another reference, with Rec709 LUT

Wilson Luniz wrote:

Johannes Jonsson wrote:What lenses were you using Wilson? And what is the max F/stop for them?

A Fujinon 8-120 b4 lens with 2x extender(I forgot the exact model, get you back later). The F/stop limited to f/4.0 after applying 2x extender(and I'm using f/4.0).

That lens could big a big portion of your problems... but I don't think it's your only problem...

I'm trying to explain more about my issue here: there's three separate issues but both affect image and result noisy:
1. Coloured vertical lines
2. FPN
3. Cross Hatching
And I'm trying to write more point by point:
1. Coloured vertical lines
This was officially confirmed that my camera is damaged from BMD support. This issue affects even images are bright(still quite visible while around 50% exposure). You can see the attachment, some pink and green lines appear on the screen even shoot with body cap.
2. FPN
There are some color noise in around 5-6% exposure(Film range) with ISO1600. They will go away when more light comes into the camera(or you can say, they become hardly to notice). Capture from SDI/Record RAW/Record ProRes/Calibrate sensor/Downgrade or upgrade firmware(3.0,4.3.1,4.4) won't help.
3. Cross Hatching
I'm starting to suspect my camera also suffer from cross hatching after I saw this post(viewtopic.php?f=2&t=51936). The big difference between cross hatching and FPN is: FPN is color noise and cross hatching is luminance. As you can see there're many many white/grey cross pattern appear on the first untouched ProRes image. It becomes conspicuous when you scale down the image.

I've decided to return this second hand camera to the seller since it's confirmed as damaged machine. But I still want to hear more because I still want to buy another URSA mini.

JasonFinnigan wrote:

Wilson Luniz wrote:

Johannes Jonsson wrote:What lenses were you using Wilson? And what is the max F/stop for them?

A Fujinon 8-120 b4 lens with 2x extender(I forgot the exact model, get you back later). The F/stop limited to f/4.0 after applying 2x extender(and I'm using f/4.0).

That lens could big a big portion of your problems... but I don't think it's your only problem...

Maybe, but I also shoot some scene using a 28-70 f2.8 canon lens. I also got problem with that. But I can't post any frame from it due to privacy problem. I agree to your sentence: " I don't think it's your only problem..."

Most B4 lenses open up to around f/1.8 so by using the doubler you are throwing away a couple of stops of light there as well. It's not always possible to change position but in a situation like this where you are struggling for light it would have been very beneficial to move the camera forward and not be forced to use the doubler.

Also worth checking if your B4 lenses cover at 2K 16:9 on our sensor with the B4 mount. Any HD B4 lens should cover that image area. Using 2K instead of HD, especially in cases like this where you don't have enough light, should help you as it will be oversampling the image slightly then allowing you to scale it down for 1920x1080 for delivery which helps reduce perceived noise.

I'd like to add that while that Fujinon B4 lens opens up to f/1.7 at the wide end it does not hold its f-stop throughout the zoom range.

Just did a short test with the Sony EX3 I've lying around and its Fujinon 5.8-81.2mm f/1.9. At the narrow end setting the aperture to f/1.9 does nothing - the image has the same brightness as at f/2.8.

So I guess your Fujinon 8-120mm f/1.7 in reality is a f/2.4 @ 120mm and with the doubler losing around 2.5 stops you will get something like f/5.6

I did some further tests in windowed sensor mode, ISO1600, and 360° shutter.

I'd say the shot with high detail sharpening looks exactly like your sample shot of your initial post:

Robert Niessner wrote:I did some further tests in windowed sensor mode, ISO1600, and 360° shutter.

I'd say the shot with high detail sharpening looks exactly like your sample shot of your initial post:

Hmmm...mine looks worse...
SAD but true.

Robert Niessner wrote:I'd like to add that while that Fujinon B4 lens opens up to f/1.7 at the wide end it does not hold its f-stop throughout the zoom range.

Just did a short test with the Sony EX3 I've lying around and its Fujinon 5.8-81.2mm f/1.9. At the narrow end setting the aperture to f/1.9 does nothing - the image has the same brightness as at f/2.8.

So I guess your Fujinon 8-120mm f/1.7 in reality is a f/2.4 @ 120mm and with the doubler losing around 2.5 stops you will get something like f/5.6

Tim Schumann wrote:Most B4 lenses open up to around f/1.8 so by using the doubler you are throwing away a couple of stops of light there as well. It's not always possible to change position but in a situation like this where you are struggling for light it would have been very beneficial to move the camera forward and not be forced to use the doubler.

Also worth checking if your B4 lenses cover at 2K 16:9 on our sensor with the B4 mount. Any HD B4 lens should cover that image area. Using 2K instead of HD, especially in cases like this where you don't have enough light, should help you as it will be oversampling the image slightly then allowing you to scale it down for 1920x1080 for delivery which helps reduce perceived noise.

You are both correct. The doubler has diaphragms inside which limits the maximum f-number. On my lens I found f/2.8-f/4.0 looks almost same. So I said f/4.0 is real maximum f-number on my lens.

1. To fix cross-hatching pattern change X/Y position to 0.5px or use this fix https://ursa-mini-fpn-grid-pattern-nois ... 18/fix-46k
2. It seems you try to recover information from shadows by simple boosting Expose or Gain alone, which is wrong. Try to use physically correct Boost Expose and Lower Gain method described here viewtopic.php?f=21&t=65149 and with more examples on bmcuser thread with same name.
To adjust Exposure in ProRes footage same as in RAW add two Color Space Transform nodes, and one more node in the middle.
Set first node input gamma Timeline Gamma in , Linear out
Set another node input gamma Linear in , Timeline Gamma out.
Adjust Gain in the middle node. This produce exact result as Exposure slider in RAW controls.

Dmitry Shijan wrote:1. To fix cross-hatching pattern change X/Y position to 0.5px or use this fix https://ursa-mini-fpn-grid-pattern-nois ... 18/fix-46k
2. It seems you try to recover information from shadows by simple boosting Expose or Gain alone, which is wrong. Try to use physically correct Boost Expose and Lower Gain method described here viewtopic.php?f=21&t=65149 and with more examples on bmcuser thread with same name.
To adjust Exposure in ProRes footage same as in RAW add two Color Space Transform nodes, and one more node in the middle.
Set first node input gamma Timeline Gamma in , Linear out
Set another node input gamma Linear in , Timeline Gamma out.
Adjust Gain in the middle node. This produce exact result as Exposure slider in RAW controls.

w0w~~! You're genius!! You make me understand what is Camera RAW. I used to use lumetri menu in Adobe PR, which result a lot different from RAW. But now I understand. BTW, I wrote this topic because I'm trying to figure if my camera is damaged/under abnormal condition/abnormal performance(which means a brand new ursa mini won't be like mine). And i had proven it.

Wilson Luniz wrote:

Dmitry Shijan wrote:1. To fix cross-hatching pattern change X/Y position to 0.5px or use this fix https://ursa-mini-fpn-grid-pattern-nois ... 18/fix-46k
2. It seems you try to recover information from shadows by simple boosting Expose or Gain alone, which is wrong. Try to use physically correct Boost Expose and Lower Gain method described here viewtopic.php?f=21&t=65149 and with more examples on bmcuser thread with same name.
To adjust Exposure in ProRes footage same as in RAW add two Color Space Transform nodes, and one more node in the middle.
Set first node input gamma Timeline Gamma in , Linear out
Set another node input gamma Linear in , Timeline Gamma out.
Adjust Gain in the middle node. This produce exact result as Exposure slider in RAW controls.

w0w~~! You're genius!! You make me understand what is Camera RAW. I used to use lumetri menu in Adobe PR, which result a lot different from RAW. But now I understand. BTW, I wrote this topic because I'm trying to figure if my camera is damaged/under abnormal condition/abnormal performance(which means a brand new ursa mini won't be like mine). And i had proven it.

Every camera will be noisy under abnormally exposed conditions and have all sorts of artefacts especially under high ISO settings.
They are made to film scenes within in Dynamic range stated and if example you got bad noisy footage with lots of artefacts in low light situation under low ISO setting (Prores) within DR of that camera, then that would be
legitimate argument.

It is like you wish your truck to tow 5 tonnes load when it is only rated for 1 tonne then you complain that is overheating and not handling well.
Film 4-5 stops under in ISO 200 without sharpening image in camera and show us some samples.

Darko Djerich wrote:

Wilson Luniz wrote:

Dmitry Shijan wrote:1. To fix cross-hatching pattern change X/Y position to 0.5px or use this fix https://ursa-mini-fpn-grid-pattern-nois ... 18/fix-46k
2. It seems you try to recover information from shadows by simple boosting Expose or Gain alone, which is wrong. Try to use physically correct Boost Expose and Lower Gain method described here viewtopic.php?f=21&t=65149 and with more examples on bmcuser thread with same name.
To adjust Exposure in ProRes footage same as in RAW add two Color Space Transform nodes, and one more node in the middle.
Set first node input gamma Timeline Gamma in , Linear out
Set another node input gamma Linear in , Timeline Gamma out.
Adjust Gain in the middle node. This produce exact result as Exposure slider in RAW controls.

w0w~~! You're genius!! You make me understand what is Camera RAW. I used to use lumetri menu in Adobe PR, which result a lot different from RAW. But now I understand. BTW, I wrote this topic because I'm trying to figure if my camera is damaged/under abnormal condition/abnormal performance(which means a brand new ursa mini won't be like mine). And i had proven it.

Every camera will be noisy under abnormally exposed conditions and have all sorts of artefacts especially under high ISO settings.
They are made to film scenes within in Dynamic range stated and if example you got bad noisy footage with lots of artefacts in low light situation under low ISO setting (Prores) within DR of that camera, then that would be
legitimate argument.

It is like you wish your truck to tow 5 tonnes load when it is only rated for 1 tonne then you complain that is overheating and not handling well.
Film 4-5 stops under in ISO 200 without sharpening image in camera and show us some samples.

We're arguing different things. I guess this is because my bad english. However, my camera was officially confirmed as damaged by BMD's technician.

thomas bruegger wrote:it is what it is..... ive sent mine in, support rated it as "within the standards" what im seeing in my UMini looks like yours in ISO1600 window low light.

Jamie LeJeune wrote:I think those standards allow a pretty wide range. Just yesterday I tested a brand new Ursa Mini Pro (calibrated with latest firmware) against an original non-pro Ursa Mini 4.6K. Same settings, same lens, same light and charts for each camera. Noise on the Pro exhibited significantly more of a fixed pattern in comparison to original 4.6K. I'm sending comparison files into BMD today. Will see what they say.

Update: My Ursa Mini Pro went in to BMD's Fremont, California office for RMA. It took multiple weeks to get it back because, according to their support staff, their lab equipment wasn't working and needed to be repaired. Just tested the returned camera and there is zero improvement and it remains worse than my original 4.6K Ursa Mini. Even though it's been over 30 days since purchase (due to the RMA delay) B&H, bless them, are allowing me to send it back to them for a replacement unit. This is third time I've had to return a defective BMD camera to a reseller. So, rinse, repeat, around I go again.