FPN on BMCC 2,5K

Hello,

I have been the very happy owner of a BMPCC now since September and thought I'd add to my arsenal the BMCC 2,5k MFT camera. I received the camera a few days ago and during testing I noticed that it was showing quite noticeable fixed pattern noise. Normal you say? Well, I shot the BMPCC and BMCC head to head with the same settings etc and the pattern noise is not noticeable at all in BMPCC until perhaps 3 stops below optimal exposure. On the BMCC the pattern is noticeable in mid tones on out of focus areas with just a slight underexposure, some times with optimal exposure.

I guess my question is; when are you guys seeing the fixed pattern noise on your cameras (BMCC)? I was expecting if anything the BMCC to be a little better than the BMPCC based on higher resolution and same pixel pitch?

Can FPN issues improve in time / with use?

Anyone else compared the two cameras side by side to give me an idea if this is normal (sour grape)?

Thanks,
E.

I don't get any FPN on my bmcc ef mount at all. I do get a grid pattern on the pocket but if it's properly lit it's not an issue.


Sent from my iPhone using Tapatalk

I don't get any FPN on my bmcc ef mount at all. I do get a grid pattern on the pocket but if it's properly lit it's not an issue.


Sent from my iPhone using Tapatalk

Yea I get the grid pattern to when I push my pocket a couple of stops below where it should be…

Thanks.

Bmcc has no issues like this at all. Mine doesn't anyways. It's a great camera. I'm getting the 4k delivered today, I'm hoping there are no issues with that on there!

Bmcc has no issues like this at all. Mine doesn't anyways. It's a great camera. I'm getting the 4k delivered today, I'm hoping there are no issues with that on there!

Hmm, I'm the only one with this problem?

Good luck with the 4k delivery!

Some BMCC do had a sensor noise pattern evident in underxposed areas. Even the "beta" footage from JB in the past had one or two shots with it. But its not a normal behaviour. You should check it very carefully and if its like you say, you probably should get in touch with BMD and ask for a replacement.

If you could provide us some frame grabs it would be much easier than telling this "by blind".

Hope it helps.

Yup I have it on my BMCC, but hard to notice. Vertical lines or patterns tend to happen occasionally with high ISO, or in dark areas. Didn't bother me too much, but I'd love a black calibration tool... I can't get rid of it in post easily (at all most of the time).

Hello,

I have been the very happy owner of a BMPCC now since September and thought I'd add to my arsenal the BMCC 2,5k MFT camera. I received the camera a few days ago and during testing I noticed that it was showing quite noticeable fixed pattern noise. Normal you say? Well, I shot the BMPCC and BMCC head to head with the same settings etc and the pattern noise is not noticeable at all in BMPCC until perhaps 3 stops below optimal exposure. On the BMCC the pattern is noticeable in mid tones on out of focus areas with just a slight underexposure, some times with optimal exposure.

I guess my question is; when are you guys seeing the fixed pattern noise on your cameras (BMCC)? I was expecting if anything the BMCC to be a little better than the BMPCC based on higher resolution and same pixel pitch?

Can FPN issues improve in time / with use?

Anyone else compared the two cameras side by side to give me an idea if this is normal (sour grape)?

Thanks,
E.

In the early days, the only thing people were complaining that I can remember is when shooting with very low available light, there was a horizontal line that split the screen, leading people to think that the sensor is made from two halves joined together. Other than that I can't remember BMCC FPN complaints.

Jules

In the early days, the only thing people were complaining that I can remember is when shooting with very low available light, there was a horizontal line that split the screen, leading people to think that the sensor is made from two halves joined together. Other than that I can't remember BMCC FPN complaints.

Jules[/quote]

I do get that band too, but only if I completely starve the sensor for light and pull it back up. The FPN usually hits hard at about one to two stops below optimal to the right exposure. My pocket handles the same beautifully.

So if I'm reading you guys right, you don't see any fpn at all only the banding and only if you pouch the sensor below 2-3 stops?

I'll post some frames when I get back home later today.

Thanks for the replies!
E.

I haven't done so update your firmware. I got my BMCC back in February and it had FPN. I decided to update the firmware to 1.62 even though there were no real fixes listed for the BMCC. FPN is gone and has not returned since. It may have nothing to do with the firmware but may as well give it a shot.

I have two BMCC's and one had FPN quite strongly, could even see it with a lens cap on... BM replaced it and since all is fine... I have three Pocket Cameras none of them seem to have the problem as the BMCC had, so I would suggest to exchange it.


Sent from my iPhone using Tapatalk

I haven't done so update your firmware. I got my BMCC back in February and it had FPN. I decided to update the firmware to 1.62 even though there were no real fixes listed for the BMCC. FPN is gone and has not returned since. It may have nothing to do with the firmware but may as well give it a shot.

Already up to date but no improvement.

I don't see any FPN with the cap on. Also the noise I get from black frames when pushing it is nice and uniform. The FPN isn't in the shadows its in the out of focus areas, mid tone sections and looks like you have dirty glass with bars in front of the lens.

I tried sharing the images but my album doesn't like dngs, so I'm giving up on that.

The plan is to make a head to head comparison with a friends camera to see how off the mark it is and take it from there. But if nothing strange happens to my camera it seems I'll have to return it to BMD for a replacement.

I'd like to thank everyone that contributed and helped me sort this out. Thanks!

Could you share a DNG or prores clip - untouched - which demonstrates this FPN?

Could you share a DNG or prores clip - untouched - which demonstrates this FPN?

Hi Tom et al..

It has taken my a few tries, but I think this might finally be working.
Here is a prores clip. Its converted to LT so it wouldn't take forever. The DNGs are not obvious unless i underexpose 2-3 stops and then pull it up in post. The clip above is just a little under.

http://www.filemail.com/d/glxhidetqpqyhng

The file is only up for 7 days (its the only solution that has worked so far).

Look at the upper left corner as the camera moves you see the FPN.

Thanks for looking into it guys!

Could you share a DNG or prores clip - untouched - which demonstrates this FPN?

Hi Tom et al..

It has taken my a few tries, but I think this might finally be working.
Here is a prores clip. Its converted to LT so it wouldn't take forever. The DNGs are not obvious unless i underexpose 2-3 stops and then pull it up in post. The clip above is just a little under.

http://www.filemail.com/d/glxhidetqpqyhng

The file is only up for 7 days (its the only solution that has worked so far).

Look at the upper left corner as the camera moves you see the FPN.

Thanks for looking into it guys!

Hi, Thank you for sharing, but processed footage is not very helpful in terms of gauging the severity or probable cause of the problem. A single DNG is only 4-5 MB - that would be the best way to show the problem.

In reality, it is perfectly normal to expect some kind of noise and artefacting when shooting underexposed - even my 5D Mkii, 550D and Nikon D80 all have the same FPN when underexposed.

However - considering I have previously been able to expose for 12800 on the bmcc, it is still worth showing us an untouched DNG rather than transcoded prores.

Could you share a DNG or prores clip - untouched - which demonstrates this FPN?

In reality, it is perfectly normal to expect some kind of noise and artefacting when shooting underexposed - even my 5D Mkii, 550D and Nikon D80 all have the same FPN when underexposed.

However - considering I have previously been able to expose for 12800 on the bmcc, it is still worth showing us an untouched DNG rather than transcoded prores.

Hi Tom,

I completely agree you'll usually see some sort of noise when pushing underexposed material, but too me at least this is different from noise I have seen in my slrs and pocket camera. And it strikes very easily with only slight underexposure. I even see it on the screen when the light is a little low and I'm moving the camera. For tripod work it is a little less noticeable.

It is very visible in moving images thus the previous clip, a single dng and its a little less visible but this one that I gauge to be about 3 stops under shows it quite well in the upper left corner. Do you think I should expect better performance than this. From what you have shown us in your clips I would say yes?

http://www.filemail.com/d/ttagnqnddvlgmpu

Thanks,
Espen.

No that looks perfectly acceptable to me, exposing for ASA 6400 on a camera which shoots natively at 800 without any noise reduction - expect some noise and FPN.

also, 3 stops under is not really "slight" - it is quite significant - especially on a camera which thrives in exposing to the right.


Here is a quick example I just did using my 5D Mkii,
Underexposed by 3 stops - 100% crop
Shot at asa 800 - pushed 3 stops in post.

http://tommajerski.com/publicimages/5d-mkii.jpg



Noise the noise and FPN.

From what I can see, your results are perfectly normal and expected. If you are underexposing your image, you will need to use Noise reduction.

No that looks perfectly acceptable to me, exposing for ASA 6400 on a camera which shoots natively at 800 without any noise reduction - expect some noise and FPN.

also, 3 stops under is not really "slight" - it is quite significant - especially on a camera which thrives in exposing to the right.

From what I can see, your results are perfectly normal and expected. If you are underexposing your image, you will need to use Noise reduction.

No, 3 stops isn't slight, but I'm able to see it on moving footage around 1 stop below and its annoying. I did a shootout between the pocket and BMCC, on the 3 stops under shot, I can't see any FPN on the pocket. Variable noise yes, but no FPN.
So basically my pocket does better than my BMCC. Same lens, same light, same pixel pitch - the BMCC should be better, its not.

I'm not sure the MK2 is a fair comparison, smaller pixel sites, compression etc..

Filemail is restricting me any more uploads for another 12 hours, but I would be happy to show you the dng when its available.

When you say acceptable, does that mean the same as yours?

Any chance this FPN gets better with use?

Hopefully I'll get some more ansvers when I compare mine to another camera on tuesday.

Thanks for your time Tom!

No that looks perfectly acceptable to me, exposing for ASA 6400 on a camera which shoots natively at 800 without any noise reduction - expect some noise and FPN.

also, 3 stops under is not really "slight" - it is quite significant - especially on a camera which thrives in exposing to the right.

From what I can see, your results are perfectly normal and expected. If you are underexposing your image, you will need to use Noise reduction.

No, 3 stops isn't slight, but I'm able to see it on moving footage around 1 stop below and its annoying. I did a shootout between the pocket and BMCC, on the 3 stops under shot, I can't see any FPN on the pocket. Variable noise yes, but no FPN.
So basically my pocket does better than my BMCC. Same lens, same light, same pixel pitch - the BMCC should be better, its not.

I'm not sure the MK2 is a fair comparison, smaller pixel sites, compression etc..

Filemail is restricting me any more uploads for another 12 hours, but I would be happy to show you the dng when its available.

When you say acceptable, does that mean the same as yours?

Any chance this FPN gets better with use?

Hopefully I'll get some more ansvers when I compare mine to another camera on tuesday.

Thanks for your time Tom!

That 5d Mkii shot is from a raw file, not compressed video. It shows far more noise and FPN than in your dng - it is an example of how the exact same effect is seen in other cameras when underexposing in the same way. So there is no compression.

The frame you shared and video look to be exactly what I would expect to see with mine.

Underexposing by 3 stops is not a good idea for any shoot, but even so - even moderate NR can pretty much eliminate it anyway.

Can you share a DNG underexposed in the same way from your pocket camera?
When exposing correctly -can you see any FPN?

No that looks perfectly acceptable to me, exposing for ASA 6400 on a camera which shoots natively at 800 without any noise reduction - expect some noise and FPN.

also, 3 stops under is not really "slight" - it is quite significant - especially on a camera which thrives in exposing to the right.


Here is a quick example I just did using my 5D Mkii,
Underexposed by 3 stops - 100% crop
Shot at asa 800 - pushed 3 stops in post.

http://tommajerski.com/publicimages/5d-mkii.jpg



Noise the noise and FPN.

From what I can see, your results are perfectly normal and expected. If you are underexposing your image, you will need to use Noise reduction.

It's more about bars, lines and such. Sometimes on 800 iso or even lower (bmcc). I think every digital sensor has that problem, but the software handles it more or less.

No that looks perfectly acceptable to me, exposing for ASA 6400 on a camera which shoots natively at 800 without any noise reduction - expect some noise and FPN.

also, 3 stops under is not really "slight" - it is quite significant - especially on a camera which thrives in exposing to the right.


Here is a quick example I just did using my 5D Mkii,
Underexposed by 3 stops - 100% crop
Shot at asa 800 - pushed 3 stops in post.

http://tommajerski.com/publicimages/5d-mkii.jpg



Noise the noise and FPN.

From what I can see, your results are perfectly normal and expected. If you are underexposing your image, you will need to use Noise reduction.

It's more about bars, lines and such. Sometimes on 800 iso or even lower (bmcc). I think every digital sensor has that problem, but the software handles it more or less.

That image is from my 5d - it is an example of the "lines and bars" in another camera. I am trying to demonstrate that fpn is perfectly normal and expected when underexposing by several stops by showing it on another non-bmd camera. The 4k cam is clearly problematic as it is far more visible even when exposing correctly. But on the bmcc and in the dng provided by the OP - fpn is only visible when not exposing correctly - which is normal.

That 5d Mkii shot is from a raw file, not compressed video. It shows far more noise and FPN than in your dng - it is an example of how the exact same effect is seen in other cameras when underexposing in the same way. So there is no compression.

The frame you shared and video look to be exactly what I would expect to see with mine.

Underexposing by 3 stops is not a good idea for any shoot, but even so - even moderate NR can pretty much eliminate it anyway.

Can you share a DNG underexposed in the same way from your pocket camera?
When exposing correctly -can you see any FPN?

OK, if its raw then its all good. However the noise I'm seeing here, and I can see the mk2 noise on both my cameras when pushed enough, is different from what I'm seeing at mild underexposure. Looks like dirty glass.

Here is the pocket DNG: http://we.tl/QwMliMPWAT

But again this is most apparent on moving footage. If I can find a way to share it I'll try to make a moving comparison of the two.

Thanks,
Espen.

Another little funny thing is if I shoot with the cap on and boost the signals in post, the BMCC exhibits a cleaner noise distribution than the pocket. Also the typical mid horizontal line is visible.

I just don't get why the semitransparent lines are so easy to see on the BMCC and completely lost on the BMPCC on the same exposure..

I loaded the DNG you posted into Resolve.

When viewing the DNG File outside of Resolve I noted that it looks like the exposure was pulled very far up to the point of overexposure. It definitely an awful looking result.

However, I loaded the file in, set it to BMD Film and noticed that the Color Temperature was set to Daylight Levels (5556 as interpreted by Resolve) whilst shooting under (what I assume) is an Incandescent light source - I reset this to 3200k and applied Captainhook's basic BMD Film LUT and adjusted the colors a tiny bit. There is some noise but you could easily get rid of it if you'd like (I do not have a full version of Resolve unfortunately so no NR applied ).

The Color temperature change reveals that the image is at a lower exposure but not necessarily underexposed (at least by the definition I work off of).

The clip I loaded showing an FNP Pattern gives me very little information, What Lens/Stop/Filter are you using (there's also a black blob in the upper left corner, should maybe clean your lens).

Here's the DNG you posted earlier regarding being 3 stops underexposed uploaded as an Uncompressed TIFF (Viewing BMD DNGs outside of Resolve tends to yield wacky results):

https://www.dropbox.com/sh/jmf2fqoq80h7s0a/ajAtYOqNE-/bmpcc3stopunderTest2_1.1.2.tif

I'd love to have an original DNG from the video posted earlier showing the vertical noise pattern (I'd also suggest to clean your lens as I saw a mysterious blob on the upper left corner).

How are you viewing your footage? directly in an NLE or through Resolve?

Any information will help us help you.

No that looks perfectly acceptable to me, exposing for ASA 6400 on a camera which shoots natively at 800 without any noise reduction - expect some noise and FPN.

also, 3 stops under is not really "slight" - it is quite significant - especially on a camera which thrives in exposing to the right.


Here is a quick example I just did using my 5D Mkii,
Underexposed by 3 stops - 100% crop
Shot at asa 800 - pushed 3 stops in post.

http://tommajerski.com/publicimages/5d-mkii.jpg



Noise the noise and FPN.

From what I can see, your results are perfectly normal and expected. If you are underexposing your image, you will need to use Noise reduction.

It's more about bars, lines and such. Sometimes on 800 iso or even lower (bmcc). I think every digital sensor has that problem, but the software handles it more or less.

That image is from my 5d - it is an example of the "lines and bars" in another camera. I am trying to demonstrate that fpn is perfectly normal and expected when underexposing by several stops by showing it on another non-bmd camera. The 4k cam is clearly problematic as it is far more visible even when exposing correctly. But on the bmcc and in the dng provided by the OP - fpn is only visible when not exposing correctly - which is normal.

Yes, I understood, but I think you may not. Bars, literally. Your 5d is not a good example of this problem. Our cameras may have different sensors. I won't search for frames to show you, but it's true.

https://www.dropbox.com/sh/jmf2fqoq80h7s0a/ajAtYOqNE-/bmpcc3stopunderTest2_1.1.2.tif

I'd love to have an original DNG from the video posted earlier showing the vertical noise pattern (I'd also suggest to clean your lens as I saw a mysterious blob on the upper left corner).

How are you viewing your footage? directly in an NLE or through Resolve?

Any information will help us help you.

I do both. Resolve and adobe Camera RAW. I do get a lot of noise when pulling up the dngs but I'm expecting that noise, what is the problem is a vertical transparent looking bars that become visible when panning or moving the camera. They are almost impossible to see on a dng, thus I think we are talking about different things.

The lens is dirty, yes.. just put it on without checking after yesterdays shoot.

Can you guys see bars on your cameras screen when in low (not dark) light?