BMPCC - Dusty sensor and Highlights Grid Pattern (texture!)

Honestly, we do not have to try to solve this problem. This is not normal, that's all.
If you pay your equipment 1,000 or $ 20,000, the reasoning is the same.
The fact is that today more than ever, it is us who actually test the hardware. It is a reality, but the point is that everything is in order quickly, people are reactive, that's all that matters.
As I said earlier, I love this little BMPCC, for whom I reserve a lot of beautiful things. For me, its potential is greater than its big sisters. (despite the limitations of its sensor, HD ...)
We bet that this problem of grid pattern will be fixed at the next firmware?

Christophe

Just wanted to add that I've also seen this in mine and I can reproduce it 100% of the time; so I haven't touched the camera since I found this because it's truly a major issue that WILL ruin shots.

I'll try using a variable ND or CPL filter next time to see if that helps any but I'm starting to feel like this was a wasted purchase, this needs to be fixed asap, before I try to sell my BmPCC for a GH3.

Is this happening in the BmCC too? If so I expect this in 4K BmPC as well, and if there isn't a fix by the time the BmPC is available then I'll be canceling my order.

We're not alone in wondering why support Blackmagic is no more reactive.
Yet they have everything to succeed, similar to Red.
Many promises have not been kept, victims of their success ... but today? It finally moves? ...
From my side we still believe. If this important issue of grid pattern is solved with a new firmware, so I remain faithful. I am willing to test the reliability and stability of their products, but I also have a job to do, and our patience has its limits. This is normal, right?

I still do not understand is the willingness of some buyers BMPCC or BMCC, wanting to find a particular explanation for this anomaly ... Hey guys, you have time to do that? Just take the time to prevent Blackmagic support, that's all. This is important. For the rest, it is up to them to respond!

There is one thing I still do not understand is the willingness of some buyers BMPCC or BMCC, wanting to find a particular explanation for a technical problem ... spend hours, days to find a way to solve a technical problem ... Hey guys, you have time to do that? OK, but you can also simply prevent you support, that's all. This is what matters. For the rest, it is to Blackmagic to do the job!

What I note is that at this point, we still do not have any information.
So what do we do in a professional context, inevitably replaced momentarily by another camera, perhaps by a material of another company ...
Action, guys!
Again, this does not prevent my interest, my enthusiasm for the Blackmagic products, Davinci resolve (a total success for me) to BMPCC.

Just hope that everything is finally most linear and reliable for what concerns the communication already.

Christophe

They fixed the white orbs problem and the black dots problem, so i'm confident they are going to fix this purple grid problem as well. They have to, otherwise i'm sure a number of buyers (myself included) will send the cameras back...

It would be nice to know if the "use a variable nd"-solution solves the problem because a) it reduces light (so basically closing the iris should do the same), b) it softens the image or c) it polarizes the light. Any findings?

So I did a few more tests but the grid appeared in them all...

I tried all ASA/ISO speeds, all shutter angles, closing down the iris, attaching a polarizing filter (Marumi DHG Super CPLD) and then a variable ND (Lighcraft Workshop's Fader ND mark II) and the purple grid was still there every time; dialing up the ND or down the exposure reduced the flare and therefore the grid but the grid is still there (localized to the size of the flare). None of these are practical solutions anyway.

The only one work around we have at the moment is blurring/pixel shifting the image in post or avoiding flares, which is a horrible solution lol.

I'm willing to work around a camera's flaws for a good image; black blowouts, blooming highlights, the odd 24/180 shutter angle, dropped frames with every card but this one is just too much and without a decent temporary fix. I understand this was an ambitious project and I'm actually ok paying to be in a live beta for this camera but it would be nice to hear something from BMD and if it can be fixed with a firmware update. I've heard two assumptions so far; one is a debayer issue which hopefully would be fixed with software and the other is that it's a reflection from the sensor's protective layer which would mean it needs a hardware fix.

UPDATE: I just did another test and happy with the results! The Grid does not seem to show up over the HDMI out, so we might be able to record external (I know kinda defeats the form factor advantages) and not worry about the flares. This would also mean it's not a hardware issue with the sensor and should have a firmware fix soon, I'm guessing it's in compression to card with BMD, it's not a debayering issue if it's clean over HDMI right?

Wow thats really weird if its true. The image data being sent out via HDMI and the data being compressed should be identical normally and this thing doesn't look like a compression artefact at all.. and didnt some of you also see it on the BMPCs Screen already when shooting?

Personally I wasnt able to reproduce it.. the only time I see something this grid like is when I point the BMPC without a lens attached into a bright light source (but then its also already visible on the onboard screen/monitor).
And I can definitely vouch for my BMCC being free of this artefact as I've already Shot a lot into light sources/recorded flares with it

So I would guess at least it doesn't happen on all cams/models.. hopefully its fixable without another recalibration...

I agree it's odd. HDMI is 10 bit 422 so I assume it's like their ProRes compression but maybe something else is happening before the LCD and card.

I first saw it when shooting someone standing near a window with a low sun shining in but now I shoot it into my desk lamp (Halogen bulb) and slowly pan it away, eventually the majority of the screen turns purple with a fine dot grid; I see this on my BmPCC LCD and footage recorded to card but I do not see purple or the grid on my Field Monitor (IKAN D5).

It could be that my monitor just hides it (just remembered it's a 720p panel) so I urge everyone else to try it too. I might try another test tomorrow with a 1K and the BmPCC hooked up to my TV.

I don't have a RAW capable card yet (waiting for xmas...), but i'd like to see if the pattern is also visible in raw recordings - this would eliminate ProRes encoding and even debayering as the culprits. But if it really isn't visible in HDMI out, this is really strange.

Did another test with two different 1080p TVs (42" and 23"), I still did not see any grid on the TV and clearly saw it on the camera's LCD.

BUT (stick with me it gets odd here) I also did not see the grid on the TV when playing back the clip through the camera. I did see the grid during playback on the camera's LCD and on my laptop and on the TV through my laptop; only playing back the clip through the camera's HDMI did the footage look clean.

I assumed the HDMI was getting the footage before this grid was added but it looks like the HDMI is just hiding it. That's OK for a workaround but that means anything ciuld be causing this; it could be a Bayer issue or the protective glass or compression or something else. For now am external recorder is probably there safest way to use the BmPCC until BMD figures out a fix. (Maybe the HDMI is using a potential fix right now).

Would be nice to know if BMD's QA is on this.

Most probably you cannot see the grid pattern due to the lack of rescaling which emphasizes the pattern on native fullhd displays.

Most probably you cannot see the grid pattern due to the lack of rescaling which emphasizes the pattern on native fullhd displays.

If a native fullhd display emphasizes the pattern then wouldn't it be more noticeable on the TV, plus I see in on my laptop (720p) and not on my field monitor (also 720p).

Has anyone else tried this? I need to know if I'm crazy lol and can anyone try this with an external recorder?

No, a natural 1080p display doesn't emphasize the grid. It would be exaggerated by scaling down to an uneven fracture, like 51% oder 35% though.

I meant that the rescaling emphasizes the pattern. Sorry if that sentence was not understandable.

I can see this on almost any shot, sometimes its more visible, sometimes I just get to see it when I push the material around in davinci. I don´t mind fixing this by resampling as a workaround until a firmware fix is coming. The only question for me is: Is this a general known problem with all cameras or should I sent my camera back? So a statement here would be really useful.

No, do not send your BMPCC. I had a message from the support: the problem has been detected and recognized as such. A new firmware will fix the problem soon.
So this is good news.

Christophe

Hello,

I just wanted to say that this same issue is happening on the Sony F5 and F55, 19x and 30x more expensive cameras... happening both in RAW and XAVC. Caused by flares.

check out the forum at community.sony.com, and the thread titled F55 Sensor Grid Artifacts. Images on page 1, page 3, page 5, and a link to a XAVC video example on Vimeo on page 15.

Interesting find. So we know that this is no singular problem with blackmagic's sensors, but a common problem, just like the orb problem was found in other cameras before. And from the explanations given, it seems this problem could be fixed in firmware - so let's just wait for BMD to do it, like they did with black suns.

Is this going to be fixed soon?

Not within the next 3 months

Hello,

I just wanted to say that this same issue is happening on the Sony F5 and F55, 19x and 30x more expensive cameras... happening both in RAW and XAVC. Caused by flares.

check out the forum at community.sony.com, and the thread titled F55 Sensor Grid Artifacts. Images on page 1, page 3, page 5, and a link to a XAVC video example on Vimeo on page 15.

This only happening in a rescaled preview window for F5 and F55 right? Not in the actual full res footage.

This only happening in a rescaled preview window for F5 and F55 right? Not in the actual full res footage.

If it is visible in rescaled preview, it has to be there in full res as well - yet on an other scale. I don't know about F5 and F55, but i know first hand from the Pocket cam: the pattern is there, it is a regular, pixel sized pattern within the full image. You can only see it on close inspection or magnification, and it doesn't look very regular at first glance. But simple rescaling will amplify the effect and transform the pixel-sized pattern to a larger checkerboard pattern, its structure size depending on the scaling ration; thus you will get a much more prominent pattern.

Has BMD said anything in regards to this issue over the last little while?

The interesting thing is that in all the examples shown here, the shot is vastly over-exposed. Perhaps with correct exposure it will not be evident.
Geoff

I really don't understand this comment, as if overexposure is a bad thing in cinematography or a technical fact that you get these errors with overexposure.

Here is my example shot today(and the reason I stumbled upon this thread), this shot should be clean and is definetly not overexposed overall.
The purple flare in the middle shows the pattern.

10th of February BM uploaded 1.6.1
The release notes say only stability improvements nothing about the pattern in highlights.
I hope we will have a fix soon..

bmpcc_pattern.jpg (509.72 KiB) Viewed 9772 times

1231.JPG (62.83 KiB) Viewed 9772 times

Blackmagic accept an exchange of my BMPCC ... but with my new model, the problem remains!
I have a feeling that this problem affects all models distributed on the market!
Blackmagic is not interested in the default bound to the sensor?
The problem can not be solved?
For months I note of complaints about this and no firmware solves nothing.
OK guys, take out all the years on the market of new cameras, but controlled and accomplished cameras! ... And above all, whatever happens, keep your customers informed! ...
The BMPCC has everything to mark his time ... but we must get rid of her little cancer ... Okay so for this "grid pattern" what do we do?

This is the only remaining issue with my camera, too. It wasn't mentioned in Kristian's post about the upcoming firmware update...

http://forum.blackmagicdesign.com/viewtopic.php?f=2&t=21179

Maybe the new debayering method that he mentions will help with this issue?

The grid flare pattern is a major PITA; and thusly the only real solution i've comeacross is to act like i'm shooting back in the 1950s/40s where lens flares were frowned upon by studio execs, same with underexposure a bit since it won't play well at the drive in.
I still need to test, but haven't had reason to, whether or not the purple grid stills comes up when shooting raw on the pocket-- all my pocket shoot save for one or two have been pro res affairs.

thusly the only real solution i've comeacross is to act like i'm shooting back in the 1950s/40s where lens flares were frowned upon by studio execs

I heard that Emilio Fernandez once punched his cinematographer in the face for getting a lens flare.

Ahhh, how times have changed.

He was also the only director to ever stab a critic, and to shoot an actor for getting his lines wrong. Even Bunuel was scared shitless of him.

Sorry for the OT

My kind of guy

After the last update, the grid pattern has she gone? No.
This is the ultimate point of Blackmagic problem.
I ask people with the BMPCC, and I see that everyone has this problem. So I ask again these 2 questions: all BMPCC are affected?
After all this time, why BMPCC unable to solve the problem?

To the best of my knowledge, all Pocket cams suffer from the "purple grid in highlights" problem, although the problem also seems to depend on the lens used - as far as i can tell, faster lenses are more prone to cause purple grid. To me this looks like the green photosites in the bayer pattern don't respond to high exposure in the same way as the red and blue ones - i just don't know wether this is a technical problem of the sensor (calibration? any one remember...?) or a faulty debayering algorithm; if it's the latter, the problem should not be present in raw recordings.

I just bought a BMPCC and shot a music video, at night, with tons of flares. The grid pattern is in every shot. Clearly I was unaware of this issue prior. Any new word on getting this fixed via firmware?

Can I ask what lens and adapter you used?

I've only ever had it with my Rokinon's on a dumb ebay EF-MFT adapter, never with my 12-35mm Lumix MFT lens - no matter what I point it at... My assumption was it's stray light bouncing around inside the camera.

Last time it showed up I was thinking of lining the inside of my adapter with black felt, thinking the large image circle of EF lenses may be bouncing off things it shouldn't?

I don't think it is the adapter; but could the a combination of anti-reflective coatings inside the lens. I've had it on Nikkors, Super Speeds, and Ultra Primes, and Optar Illuminas, but never on Cookes-- yet.

It is probably a calibration/sensor/firmware thing since I don't recall this happening to my camera prior to sending it in to get rid of the white orbs-- which also tells you how long I've had the thing.

Regarding the pattern: People are blind??? I even post a frame of the GRID pattern, and I made the test again 10min ago and I assume that it's related with highlight exposure (maybe reflection with the "glass" protector and the sensor!!! because that only happens when you get a light source lateral to the lens / sensor.... Well... What to say....

You'll see the pattern much much more intensive when shooting infrared.

Don't know why but I have an assumption: it's an artefact of the debayering. Whats different in IR is, that all channels got the same information. Normally the luminance of green is greater.
So that matches with you purple. The pattern is visible more intensive when there is less green in the picture.
It's not on the green channel only. It's on all channels. But it appeares when green luminance is low. At least that my working hypothesis.

It's hard to see on youtube. But here

at 0:24 look at the shadows on left house. You see a kind of strange rectangular noise. This is a combination of the pattern, some sharpening, and video compression. Normal noise looks differnetly. I cannot load pictures directly in the forum.

or a faulty debayering algorithm; if it's the latter, the problem should not be present in raw recordings.

It is in RAW and also in Adobe Camera Raw.

That's bad news, as it lessens my hope for firmware fix.

Hi there, folks.

I just bought a BMPCC and while editing some test (thank God it was a test) shots I noticed the grid pattern. First I thought it could be the cheap circular ND I was using, but after some exhaustive google search I found that it's a camera issue.

So, first thing to do was updating the firmware, what I did yesterday night.

So, here comes the mystery: Before updating the firmware, I did some test shots. Since I could see the small grid pattern only with the Metabones Speedbooster attached (no lens) I did a clip in this condition. Then, another without the metabones, what also shows the pattern. Right after that, I updated the firmware and did the same test shots - the pattern still there. BUT... when I attached my Nikkor 50mm this morning and did the very same shot from my window, that revealed the grid (same light indeed, sunny day at 10am) no grid there.

All this explanation is to ask you 2 questions: Have you guys experienced an improvement on this issue after the last firmware update?

And I've read somewhere (not sure if it was on this topic) that it would be happening when the battery is low, and replacing the battery would stop that. Anyone noticed that?

Once again, I feel quite disappointed indeed. Quite disappointed. I felt with my Nikon D7100 once - a camera that have a nice image, but when you pull down the contrast on the color profile looking for a "cinestyle like" look for a better grading on post, a ridiculously ugly banding noise makes your footage a crap. Now this, with a camera named "cinema camera". Ok, we all know it's a marketing exaggeration, but having equipments with such issues on the market for people to buy and put the food on the table, is something we should not allow as professionals and customers. Just saying.

Cheers, people.

Ricardo Lisboa - Sao Paulo / Brasil

Has anyone figured out a fix to this problem yet? Either in post or in camera? Here are two screenshots: in the first you can barely make out the grid in the flare on the right. In the second, I panned right and purple maze madness ensued.

What's the AE fix for this? Can we seriously not shoot into the sun with this camera? Any help, BM?

BMPCC RAW
Lumix 14-42 MFT (so no adapter)
IR and ND filters on the lens

Has anyone figured out a fix to this problem yet?

In post shift the clip 0.5px left and 0.5px down. Grid will disappear but you'll still have the pink flare. BM are aware of it and are probably looking at a fix.

BM are aware of it and are probably looking at a fix.

The keyword being "probably". It took months to get them to even acknowledge the problem, and since then there hasn't been a single official word about fixing it...

OK, the .5px/.5px shift gets rid of the grid, but softens the whole image somewhat. Plus, there's the purple (and still noisy) flare. I wish BM would say something more definitive about it.


Sent from my iPhone using Tapatalk

[CaptainHook]

Hi, what happens when you render this out and view at 1:1?
Do you still see it if you set the viewer in AE at 100%/full resolution?

Hi CaptainHook - Output to ProRes 422HQ, the problem remains, though on my monitor the pattern is really small. Zoomed in AE, what looks like a grid in Resolve looks like a maze:

I'm getting this too. I haven't even gotten to the NLE step with what we were shooting and was just doing a onelight in Filmconvert. It was my first time using the BMPCC in production along side a BMCC. The pocket has the issue and the BMCC does not. I should also note that I was using a Speedbooster (Active EF) and was shooting prores 422.

PS - I just tried to add an attachment so everyone could see and am getting an error: "Sorry, the board attachment quota has been reached."

One more user of the GridPatternBMPCC !

i got it once shooting with hard side light and towards the light.