DEAD PIXELS

Well, for me my dngs have very low light, but u are right, depends from every person.

For sure the one with the street has very low light. One thing that maybe you have noticed, and i´m very impressed with that camera is the sky colors. I didin´t see with my eye all the colors from the grey sky.

I will put now more dngs if you want to play xD

In lowlight I can see these stucked pixels like in the linked video (download the original file and you will see them easily, because they don`t move - instead of the other noise). With a denoiser I can get rid of them pretty fast - but I thought they would get mapped out by the camera software - so maybe I have a faulty cam? This one was shot with Iso 1600 and pushed only a little bit in post - but I get them also with morge light. I tried to turn the cam on and of - but the pixels are always visible.
Should I contact the support or do you get same issues?




Password: bmcc

In lowlight I can see these stucked pixels like in the linked video (download the original file and you will see them easily, because they don`t move - instead of the other noise). With a denoiser I can get rid of them pretty fast - but I thought they would get mapped out by the camera software - so maybe I have a faulty cam? This one was shot with Iso 1600 and pushed only a little bit in post - but I get them also with morge light. I tried to turn the cam on and of - but the pixels are always visible.
Should I contact the support or do you get same issues?




Password: bmcc

I'm not sure where I'm looking to see the stuck pixel, but...
This is NOT "low light". This is no light. You underexposed by way too many stops and are trying to pull back what just isn't there. You can't expect to pull back clean shadows when almost the entire image is severely underexposed.

Did you download the file? In the middle of the picture and also on the right edge there are a lot of these. I`m watching it on a 17er 1080p macbookpro with brightness set to full.

Of course it is underexposed and there is no way to get a noisefree picture. But there are parts of the picture which are ok, like the windows. Sometimes I need these kind of shoots - and I never, ever had such stucked pixels with other cams. The dark parts of the picture should have noise but no stucked pixels. Or am I wrong?

The other day I did filming during the night trough the carwindow - there where a lot of lights from other cars, but in the dark aeras there where these stucked pixels again - also with ISO800. The problem is, that they are very visible - because they don`t move, like noise does.

John Brawley said in another thread, that there should not be stucked pixels - because they will get mapped out by the cam.

Despite of this I really love this cam - it has fat, deep colors and looks really nice even in prores...

That´s right. Those pixels must be mapped put and the BMCC can do that. There just seems to be a bug with some cameras.

Did you download the file? In the middle of the picture and also on the right edge there are a lot of these. I`m watching it on a 17er 1080p macbookpro with brightness set to full.

Of course it is underexposed and there is no way to get a noisefree picture. But there are parts of the picture which are ok, like the windows. Sometimes I need these kind of shoots - and I never, ever had such stucked pixels with other cams. The dark parts of the picture should have noise but no stucked pixels. Or am I wrong?

The other day I did filming during the night trough the carwindow - there where a lot of lights from other cars, but in the dark aeras there where these stucked pixels again - also with ISO800. The problem is, that they are very visible - because they don`t move, like noise does.

John Brawley said in another thread, that there should not be stucked pixels - because they will get mapped out by the cam.

Despite of this I really love this cam - it has fat, deep colors and looks really nice even in prores...

I downloaded it and can see them, they are hot pixels indeed, however, in raw mode on my 5d Mkii, when very underexposed like this and pushed I can see hot pixels also.

I doubt very much you can see them in properly exposed shots.

The solution - Shoot with more light.

I did a lot of underexposed shots like this with the 7D, Gh2 and HVX200 - never had these kind of pixels but a lot of noise and I`m totaly ok with noise (I love noise - especially the bmcc noise is beautifull).

I hope that this is only my cam and it can befixed. Are there any BMCC owners which can`t reproduce situations in lowlight with stucked pixels? This would be interesting.

I did not see any stucked pixels in the picture which was exposed only with a lighter...

I did a lot of underexposed shots like this with the 7D, Gh2 and HVX200 - never had these kind of pixels but a lot of noise and I`m totaly ok with noise (I love noise - especially the bmcc noise is beautifull).

I hope that this is only my cam and it can befixed. Are there any BMCC owners which can`t reproduce situations in lowlight with stucked pixels? This would be interesting.

I did not see any stucked pixels in the picture which was exposed only with a lighter...

Were you shooting video or RAW stills with those cameras (7D, GH2, etc)?

If you were shooting video you have to remember that the camera is down scaling and line skipping to get the image down to 1080p from the much higher stills resolution of the sensor. It is also then compressing the footage. So you would never see the hotpixels as they have been scales smaller and then the compression would have blurred them.

The video I posted was also shot in prores not in raw. Why did John Brawley said that there is a function to map out stucked/hot pixels - when they are still visible in the footage shot with my bmcc?

There´s definitely some weirdness going on. Let me quoute Albert from the previous page concerning hot pixels:
"No, i mean that i saw them once and they never came back."

At the same time we have low and no light examples like the lighter video or Frank Glencairns low light test. No hot pixels anywhere. And John Brawley confirmed that the camera should map it out but his beta cam didn´t do it on every startup.

If just somebody from blackmagic would comment in this thread...

Not necessarily a connected issue, but I keep noticing single frames in my RAW and ProRes footage that exhibit single white dots. The only possible explanation I can think of is that the pixel remapping function isn't working properly and randomly makes hot pixels appear for a single frame, though I am sure my camera isn't the only one that does this. I have seen it in many BMCC videos posted so far. Has BMD commented on this issue yet?

I love that what I supposedly "say" is reported third hand and then others demand to know why I "said" this !

Every camera digital camera has dead or stuck pixels.

Dead pixels change over time. You can have more and more appearing as the months and years go by.

They are often caused by cosmic radiation. For example, if you do a lot of air travel the sensor can develop more dead pixels than usual because of the increased exposure.

The QA of the sensor maker mightn't be good. Maybe you have a bad sensor.

Every digital camera has a routine that locates and turns off these dead pixels. It's not done by the user or if it is, it's usually part of another process called "black shading" or "black balance"

Sometimes, for many reasons, dead pixels will still show up after this. This can happen for many reasons. Sometimes you have so many dead pixels that there's no memory left to map them all out. Sometimes the camera is so overheated, the dead pixels still show up.

I remember Sony users arguing with Sony about how many dead pixels were acceptable on their 100K Beta cams, back when you had to return it to service to have the dead pixel mapped out. From memory a brand new out of the box Sony could have 6 dead pixels and still be considered acceptable.

The point is that dead pixels are normal and to be expected.

They shouldn't be impacting your normally exposed images. If you're seeing them on your footage then this is a sign of a camera fault.

When you turn the BMCC on, it loads it's dead pixel routine / map and you shouldn't see the dead pixels.

In the eary days of the BMCC, if dead pixels showed up, you could often restart the camera and it would then re-laod the map and fix it. This was a bug with very early firmware which I don't believe is there any more.

jb

Not a BMD service engineer, just a DP who's seen a lot of dead pixels)

Hello JB, thank you for commenting - I don`t know someone else in the forums who spoke about the pixelmapping thing - so I reffered to you. Please excuse if I misinterpreted you.

Everytime when I turn on my cam with the lenscap on - raw or prores, ISO set to 1600 with 180degree shutter - I can clearly see these stuck pixels, when using the zoom in feature from the bmcc.
When I turn down to ISO 800, some pixels disappear - but there are still some visible blue pixels there.

I don`t want to send my cam back to the support - because I love the picture so much - thats why I`m trying to get to know if it`s a fault or not.

Thanks to everyone in the forum - it`s so nice to sit in Germany at midnight and get some feedback from somewhere in the world...

I don`t want to send my cam back to the support - because I love the picture so much - thats why I`m trying to get to know if it`s a fault or not.

From all those who have reported having warranty returns they have got replacement cameras really quickly.

If I was you I'd get in touch with BM support.

Hello JB, thank you for commenting - I don`t know someone else in the forums who spoke about the pixelmapping thing - so I reffered to you. Please excuse if I misinterpreted you.

Everytime when I turn on my cam with the lenscap on - raw or prores, ISO set to 1600 with 180degree shutter - I can clearly see these stuck pixels, when using the zoom in feature from the bmcc.
When I turn down to ISO 800, some pixels disappear - but there are still some visible blue pixels there.

I don`t want to send my cam back to the support - because I love the picture so much - thats why I`m trying to get to know if it`s a fault or not.

Thanks to everyone in the forum - it`s so nice to sit in Germany at midnight and get some feedback from somewhere in the world...

Hi.

At 1600 with a 180 deg shutter, with the blacks sitting at a normal level, I wouldn't be expecting to see dead pixels. You will see dead pixels eventually if you're pushing the image a few stops up from black, but your pictures won't look any good anyway by then

I'm afraid you'll have to part with your camera if you want this addressed. Perhaps get in touch with the online service guys. They will probably want to see some DNG's so be ready to send some and see what they say.

jb

Everytime when I turn on my cam with the lenscap on - raw or prores, ISO set to 1600 with 180degree shutter - I can clearly see these stuck pixels, when using the zoom in feature from the bmcc.
When I turn down to ISO 800, some pixels disappear - but there are still some visible blue pixels there.

I will have a look if this is the case on my camera, as I also noticed some stuck pixels on the back screen when shooting in low light a while ago.

The single white dots in single frames that appear from time to time are visible in regularly lit situations as well though. Does anyone know more about this issue?

Thanks again for replying to everyone - I think I will contact the bm-support and let you know how it turns out. Good night!

Thanks again for replying to everyone - I think I will contact the bm-support and let you know how it turns out. Good night!

So what happened?

What's new ?
Answer from BM support ?
New camera with no problem ?

Did you download the file? In the middle of the picture and also on the right edge there are a lot of these. I`m watching it on a 17er 1080p macbookpro with brightness set to full.

Of course it is underexposed and there is no way to get a noisefree picture. But there are parts of the picture which are ok, like the windows. Sometimes I need these kind of shoots - and I never, ever had such stucked pixels with other cams. The dark parts of the picture should have noise but no stucked pixels. Or am I wrong?

The other day I did filming during the night trough the carwindow - there where a lot of lights from other cars, but in the dark aeras there where these stucked pixels again - also with ISO800. The problem is, that they are very visible - because they don`t move, like noise does.

John Brawley said in another thread, that there should not be stucked pixels - because they will get mapped out by the cam.

Despite of this I really love this cam - it has fat, deep colors and looks really nice even in prores...

I downloaded it and can see them, they are hot pixels indeed, however, in raw mode on my 5d Mkii, when very underexposed like this and pushed I can see hot pixels also.

I doubt very much you can see them in properly exposed shots.

The solution - Shoot with more light.

Hi Tom, can you upload DNG at 1600asa 360° with the cap on just to see how that work on your camera.
I don't say that is " normal condition light " , its just to compare.

Thx

Let's see if I can join the club

I got my BMCC mFT today and of coarse I was looking for something wrong right away

I didn't had any ssd (didn't arrive yet) so I had to look at the screen only. I noticed some hot pixels (mostly blue and less red) when underexposing or with closed cap. Some of them seemed to be fixed, some not. Then I updated to 1.4.1 and the pixels got 5 times more!!! Then after a while the pixels disappeared except some blue which I still can see with the zoom option on the screen when on dark background.

Then I took a 2.5 hdd to get at least some seconds of footage. Checked it on my computer and couldn't find the blue pixels, but instead I found red hot pixels on another place
I guess I'm lost... here is the footage... (files are pretty short)

1. Red pixel(s):

https://dl.dropboxusercontent.com/u/305 ... 000008.jpg
DNG: https://dl.dropboxusercontent.com/u/305 ... 000008.dng

With closed cap: https://dl.dropboxusercontent.com/u/305 ... 000006.jpg
DNG: https://dl.dropboxusercontent.com/u/305 ... 000006.dng

2. Then I saw some banding in dark areas

Shot at evening:
https://dl.dropboxusercontent.com/u/305 ... 000011.jpg
DNG: https://dl.dropboxusercontent.com/u/305 ... 000011.dng

Prores: https://dl.dropboxusercontent.com/u/305 ... _C0000.mov

closed cap: https://dl.dropboxusercontent.com/u/305 ... _C0000.mov
https://dl.dropboxusercontent.com/u/305 ... _C0001.mov


What do you thing? Is there any issue with the camera or it's just paranoia?
Is there any better way to test the camera? I saw some low light footage on vimeo and it was pretty clean form banding and artefacts, so unless pushed in post, I expect the dark underexposed areas to remain banding free.

The thing is that soon I'll need the camera for a two month shooting in a remote area in the Himalayas . And I want to get a replacement ASAP if anything is not ok with this camera... Any opinion is much appreciated!

John Brawley already mentioned the same, he had to turn it off and on again sometimes to activate the mapping function.

This information is from a long time ago, back when I had a prototype camera. I've not seen it "forget" to load the dead pixels since before V1 firmware.

jb

Was wondering the same... Is this something I should live with or return the camera. Kinda new to the dead,stuck, live pixel world. I posted my example photos in another post.
http://forum.blackmagicdesign.com/viewtopic.php?f=2&t=11255#p74410
Taj

Considering that topic and many others that I've read in the past.

I can now see that horizontal split line, which you consider normal. The problem is - it wasn't there until today. I've done some tests, in my room in a low light situation, using about 6 70watt bulbs pointed in the white ceiling as a diffuser. Using 1800ASA, 360 degree angle, and a 1.8 lense. The line was still there slightly. The zebra even showed up at my testing object. I don't consider that as a low light situation. And due to the fact that line only appeared today - THAT'S NOT NORMAL sensor behavior. It's definitely has to do something with false camera operating, for example - short circuit from the charger or USB port.

My fourth replacement camera was RMA'd by BMD and confirmed faulty after several months of working fine - then one day we did a comparison test between prores, dnx and raw - something happened and then we had lots of single frame hot pixels, several dead pixels, serious banding on any area that wasn't over exposed and the sensor split became extra prominent.

Exactly what feeds my thoughts even more.

Then we can all conclude that this camera will "FAIL" after a few months of usage (on average), and that we the consumer will "PAY" for any and all RMA, Shipping and repairs and then???? Maybe not have a fixed camera after all?

Then explain to me how this camera can be used on any professional or Semi professional
shoot that has a budget and schedule?

And FYI: The best (Most Reliable) and affordable 480GB SSD drive you could use is now discontinued, what else are we going to deal with with this "Affordable RAW Camera" Yes Affordable is becoming a JOKE, or GIMMICKY! at the very least.

We still cannot conclude that now. Camera has flaws, but maybe that one was there from the very start. And my eye isn't THAT catchy, so I've only noticed now.

We need a fresh out of box camera to confirm.

John Brawley already mentioned the same, he had to turn it off and on again sometimes to activate the mapping function.

This information is from a long time ago, back when I had a prototype camera. I've not seen it "forget" to load the dead pixels since before V1 firmware.

jb

I wrote this seven month ago^^

I really don't think that "dead pixel mapping" thing is working on my BMPCC...



What do you guys think?

I really don't think that "dead pixel mapping" thing is working on my BMPCC...



What do you guys think?

You should raise a support request with BMD and get their opinion on the footage.

jb

I really don't think that "dead pixel mapping" thing is working on my BMPCC...



What do you guys think?

You should raise a support request with BMD and get their opinion on the footage.

jb

Thanks John, that's the plan and the reason I made the video. I just thought I would see what the opinion of other users is since this is my first Blackmagic camera. I'm not sure if this is just the nature of a camera that doesn't do any oversampling to speak of, or if this is a problem that has potential to get worse...

Like I said in the video description they only really come out badly when I really push the grade and heavily crush the blacks. But it seems like there really are a lot of them... I've never really noticed this on any other camera I've used.

Dead pixels are always there on any camera.

You will always see dead pixels in certain situations.

Camera / sensor heating certainly will make them show up more. Higher ISO's and a pushed grade as you've discovered will also do it.

You didn't mention what your exposure setting where but I can certainly make this happen on any BMD camera once I shoot with a 360deg shutter at 1600 ISO in low contrast with a flat field and a slightly pushed grade.

Of course ideally they shouldn't be there. As you yourself have said though, you've had to really push the shot and artificially sharpen it to get them to show.

Like the "half sensor" issue and fixed pattern noise you sometimes also see in this camera, because it's a RAW camera, BMD don't do any under the hood processing to tidy up the image. That means you tend to get everything including the faults. Most of these show up once you're in low light AND you then push the image a moderate amount like you have. Although the camera is good in low light, this is one of the reasons I've been known to say it's not a great low light camera. This is when these kinds of "faults" show up the most.

jb

Yeah I definitely get that they're there on every camera. It was more there number that was concerning me rather than their existence. I guess maybe I've just never seen so many because other cameras were probably squeezing a lot of the bad pixels out when the images were scaled down to 1080 from a much higher resolution sensor. That and maybe a lot of other cameras have some kind of internal processing that deals with them. And the heavier compression on say a Canon DSLR probably kills a lot of it inadvertently.

But also the other thing that concerned me is that I've used the same kind of heavy handed grading on the footage you shared from the original BMCC, and while I most definitely did see some I didn't run into nearly as many hot pixels. Those images were also de-noised and sharpened with Neat Video and graded heavily with Colorista in After Effects. But then again you probably shot them at ISO 200 or 400.



Everything I shot was at 180o shutter and the shot of the bowl was ISO 800. The outside night shot may have been 1600. But I've seen some of this pixel stuff as low as ISO 400. The camera was never on for more than a few minutes at night so heat shouldn't be an issue.

Like I said, if it's normal and to be expected from all of the BMD cameras then I can work with it. I can kill them in After Effects pretty easily at the expense of a little sharpness. I just need to make sure that I don't have an over-abundance of them on my sensor, or that it's a problem that is likely to get worse with time. Hence the inquire.

Thanks for taking the time to respond. And thanks for all the footage you have shared over the past year. It was a lot of fun to play with those Raw files for the first time. I can't wait until we get Raw in the Pocket.

Angus

That and maybe a lot of other cameras have some kind of internal processing that deals with them. And the heavier compression on say a Canon DSLR probably kills a lot of it inadvertently.

I think a lot of compression definitely kills a lot of super fine detail. a lot of people say they like the appearance of the BMCC in ProRes over the 2.5K because of this reasons i think. They find the ProRes micro softness more pleasing on skin detail. h.264 is even worse for that sort of thing.

But also the other thing that concerned me is that I've used the same kind of heavy handed grading on the footage you shared from the original BMCC, and while I most definitely did see some I didn't run into nearly as many hot pixels. Those images were also de-noised and sharpened with Neat Video and graded heavily with Colorista in After Effects. But then again you probably shot them at ISO 200 or 400.

They were shot RAW using my early ETTR principals so no ISO other than native 800. I guess in effect they were rated at ISO 200 but i didn't really expose them with an ISO in mind. More like ISO 800 over exposed a couple of stops.

Everything I shot was at 180o shutter and the shot of the bowl was ISO 800. The outside night shot may have been 1600. But I've seen some of this pixel stuff as low as ISO 400. The camera was never on for more than a few minutes at night so heat shouldn't be an issue.

At 800 you probably shouldn't be seeing that many then I should think. Give BMD a go and see what they say. Be specific about HOW much you're lifting the grade too. They will probably also want to see untouched ProRes files.

jb

I just received the BMCC 2.5k last weekend (august 31st, '13) and like others on this forum, was in love with the image quality, but appalled by the amount of dead/stuck pixels on my LCD screen as WELL as the actual footage when shooting in low/no light.

My problem is that I should be able to shoot black if I want to shoot black. Hypothetically speaking an artistically and narratively justified shot could have a very high light:dark ratio and be lit in very low conditions, and if it's the look I am trying to achieve, I should be able to capture that at 800 iso without a swarm of dead pixels scattered across the whole screen.

I'm not even talking about pushing anything in post. I have 3 DNG files here that I shot outside at night of my cousins in a jacuzzi. They are only being lit by the pool lights from below, which is pretty cool looking in my opinion.

Yes, it is technically under-exposed. But the subjects are visible and I feel like the vibe in these pictures is cool even so. What if that's the image I want to portray the scene?

Yet without pushing these shots in post at all, you can clearly see a myriad of these stuck pixels all over the place, even trailing across their face at times.




I called and emailed Blackmagic support, and they basically told me that they thought I was wrong, and that those weren't dead pixels, they're just grain because it's low light.

When I took three separate images and lined them up in Photoshop, then offset them by a few pixels each and lowered their opacity, you can clearly see the same pattern of stuck pixels side by side. I made a strong case with proof that I was right, and got no more responses from them, which is really disappointing. I just spent $2500 on this camera and I feel like I can't get any legitimate help.

(the three images overlayed to show that they are, in fact non-moving pixels)


Luckily, Melrose Mac didn't hesitate to say they would let me exchange this camera for a new one, but what I'm bothered by is the seemingly large group of users who have this problem, and at this point I can't tell if it's just a characteristic of the camera or a common flaw. For all I know the replacement model I receive tomorrow could have the same problem!

Here are the unaltered DNG files. The subjects are lit decently for the mis-en-scene, plus this was off the cuff, not a pre-planned scene. I'm not going to drag out a light kit to get footage of real life spur of the moment. And it kills me because these shots look killer in terms of clarity, quality light roll off- etc.

I know this is a long post, but I guess my only other related question is that I know it's been said to slightly over-expose when possible. Well how are we supposed to expose "to the right" when our iris is completely automatic and if anything, underexposes to preserve highlights? I'm not sure if it that rule of thumb is assuming we have a manual aperture ring lens (which I don't unfortunately) or if it's implying we adjust the the shutter angle or iso to adjust our exposure (the latter of which then exaggerates dead pixels! )

I'm really hoping getting a new model will solve my problem, because I just can't believe that they would sell a camera that has so many stuck pixels even at 800iso without being pushed on purpose. If these cameras do have a mapping system at start up, the pixels in these shots should surely not be showing up, right?

Thanks in advance for the help/support. Sorry for the length.

Helo, I'm having the same problem on my camera. It is a white dot in the lower left part of the screen and on the footage too. The white dot or dead pixel seems to icrease its size when pulling up iso or opening shutter angle. But its always there even when shooting at 200 iso and 45º of shutter. Can anyone help me with this issue? Have been any "official" response about this by BM? Solutions? We have to always post-pruduct this matter for the rest of our experience with this camera?
If anyone knows anything new about this problem, please let us know. This is an important issue. I can't believe that Blackmagic didn't make an official update about this.

...The difference? he wasnt boosting them several stops in post or lifting the blacks.

Interestingly enough I find that the hot pixels pop out of my lower light footage when crushing the blacks rather than "lifting" them.

Hi, I'm wondering if there is going to be some kind of solution or at least some sort of official answer to this trouble. I can't believe that Blackmagic, after stopping production for almost 6 months, send us cameras with this malfunction. I'm very disappointed with this product. Really a dead pixel on your footage is like a part of the image is missing. The problem is always there and is really difficult to do professional work with this kind of issue. I'm starting to loose clients because of this. Why Blackmagic doesn't give us some solution?