BMCC out of phace

So we know that it's a bug, confirmed through BM. Now i'd like to know if this is a bug in all CC 2.5k (which some folks didn't notice until now), or only in some few cameras? Is this maybe even the "quality issue" that stopped CC shipments last year?

?? No, that issue is very well documented and publicized Mac.

Doesn't seem to affect my camera, so no, it doesn't affect all BMCCs.

https://www.dropbox.com/s/cwytetaxipbp0 ... %20red.rar

Here are a few images if someone wants to compare and to give his opinion.
You can find here images shoot at 5600K and 5000K.
1. 7D RAW NATURAL PROFILE
2. BMCC CinemaDNG
3. Tiff images exported from BMCC DNXHD Video and Film log.

looks good to me....

Well, that settles it I guess. We were all just imagining the red jacket looked orange. Don't do the brown acid.

Interesting...when trying to balance for a neutral background...I get 3950 for the 7D footage.



When doing so for the BMCC, I get 4350.



If I try to use the same 3950 on the BMCC footage, it looks ridiculously cool.



While all cameras are different, so are lenses...the white balance shouldn't be that different. Not really. Were you using the same lens on both?

BTW, why is your 7D footage out of focus?

https://www.dropbox.com/s/cwytetaxipbp0fv/Orange%20is%20the%20new%20red.rar

Here are a few images if someone wants to compare and to give his opinion.
You can find here images shoot at 5600K and 5000K.
1. 7D RAW NATURAL PROFILE
2. BMCC CinemaDNG
3. Tiff images exported from BMCC DNXHD Video and Film log.

I see it in the DNxhd.

Hi Sean,
I used the same lens at f4 just at different focal distance, 35mm vs 50mm.
I'm getting also different neutral color values. In the lightroom I'm getting for 7d 4050 tint 7 and for BMCC it's 4400 tint 11

7D is not in focus because I was enough lazy not to focus it

Hi Sean,
I used the same lens at f4 just at different focal distance, 35mm vs 50mm.
I'm getting also different neutral color values. In the lightroom I'm getting for 7d 4050 tint 7 and for BMCC it's 4400 tint 11

7D is not in focus because I was enough lazy not to focus it

So you're getting roughly the same degree of separation as I am. About 400 Kelvin. I really do think that's more than should be expected for using the same lens and the same exposure...weird.

I did another test. This time WB is tungsten (3200k) and I used 35mm Zeiss Distagon at F2.

Findings:
1. RGB values of the "coca cola" red are different between 7d Raw and all profiles of BMCC.
2. RGB values of the red in 7D raw are (after white balance on white as neutral point):
R:100 (depends on luminance and saturation, it may go up and down). G: 0.00 (always). B: 0.00 (always). It means that red is red and zed is dead (sounds good).
3. Values of the red in BMCC raw images are not clean. In Resolve 9, with custom WB point, the "coca cola" red contains the "nice" portion of blue and green . R:120 , G:030, B: 030. It means that the red is not pure red.
4. Values of the red in DNXHD and PRORES are also not clean. In the resolve I'm getting R:130, G:030, B:030
5. In the Lightroom the picture is a bit different. Values of the red for 7d raw remains clean: R: 43, G: 00, B: 00 . Values of the red with CinemaDNG are not clean but if I change the process in camera calibration option from 2012 (current) to 2010 I'm getting the clean red with : r: 43, g: 00, b:00.
6. In the Lightroom I'm not able to get pure red neither with DNXHD nor with PRORES, doesn't matter if it was video or film log.
7. White Balance points are different between 7d Raw and Cinema DNG. For 7d is 2650K tint 1. For BMCC DNG is 2800.
8. Red on Sony F3 s-log is the clean red: R: 43, G: 00, B: 00 (the shot was different with different lens and location so it's not relevant)

Conclusion (for the conditions that I shot):

The correct color values with BMCC is possible to get only (at least at the conditions that I shot) with CinemaDNG and not in Resolve 9. Which means that another program is needed for correction (lightroom in my case) and it costs time and money.
DNXHD and PRORES are not usable for any serious job. I couldn't get clean red.

I know that some people here will jump to protect their babies like parents that will jump to protect their spoiled child (love is blind) and will claim that it's wonderful camera (which it is but with a lot of disadvantages) that costs "only" 3000$. Keep in mind that some people in this world have to work 3000 days in order to make 3000$ and in fact that's a lot of money for the camera that can't reproduce color correctly. I hope that BMD are working to fix the bug and to fix it fast or at least
someone will explain me how to get an image color correctly without using another programs except resolve and with DNxHD.

Best,
Leon

https://www.dropbox.com/s/rdgj0z0ksl7ny ... m%20f2.rar

Yeah, but that will give you better skintones.

That's the reason why the BMC has cinematic skintones while most other cameras (besides the Alexa) have typical video skintones.

And since I shoot way more skintones than Coke bottles.....

When I shoot, I'm not interested to reproduce the actual colors, I want a certain look and feel, that supports my story.
So I'm far from "protecting my baby", it is a intentional pro choice of tools,that are able to deliver that look. So I hope BM is not gonna mess with the color science - I like it the way it is, since it makes totally sense for a CINEMA camera. Horses for courses. Reality is the very last thing you want in cinema, since it's a guarantied recipe to ruin your film.

If you do color critical pack-shots for commercials, take the right tool for that.
Besides the whole point is mute anyway, cause in the whole signal chain, the colors get messed up anyway till final devilry and some "vivid magic wonder color" setting in the viewers TV is gonna ruin the whole attempt anyway.

What is better skintone is the completely individual opinion. If you like a skinton you get but all other colors are wrong it's your preference and this is not what should we except from a serious camera. If all of your colors are correct, it's not problem to get any skintone you like in CC and if an actress is pale ask for some make-up and correct a lights. Here it's not the case. It's the bug. Colors are not correct. It's contaminated. I'm just asking to red be red and not red with blue and green. Is it too much? Don't think so.

What is better skintone is the completely individual opinion.

I wasn't talking about "better" skintone but cinematic vs videoish skintone.

And what the "right" color for Coke is, isn't exactly clear, since you mix up all color spaces.

Walking arount with a spectrophometer one can measure Pantone 185 or Pantone 485 or RGB 235, 45, 46 or RGB 218, 37, 29 or CMYK 0, 100, 100, 0 and more. That’s a result of inconsistent industrial and differing production standards. According to a record at the United States Patent and Trademark Office there is a ‘Coke Red’ defined as Pantone 484.

http://brandcreation.wordpress.com/tag/coca-cola/

So, if that bottle was printed with CMYK and you are recording with (S/W - filtered) raw - that gets debayered and than transformed to YUV - than you try to read RGB values out of this to make a point, is pretty awkward in my book. If the bottle was printed with Pantone color, you have to measure the bottle first with a good colorimeter to get a reference.

But if you download the Logo from the Coca Cola press center, take in Photoshop and measure the RGB values, you get R228 G30 B44. http://www.coca-colacompany.com/press-c ... assic-logo

And now what?

Common Frank, everyone one here agrees that the red is wrong and goes more to orange with DNXHD and PRORES and you can get it only with the Raw and not in Resolve. People are not color blind, at least most of us. Now print the red that you get in a good lab and shot it, you will get different values. I can't understand why you trying to protect BMD even when they by themselves confirmed that there is something wrong with the color. Can't understand also the connection between cinematic, video and the wrong red values....Douglas Sirk movies don't look to me as video and the red is red and not orange red. Basic premise of any camera since it was invented is to take the fingerprints of the reality.
If you are happy with orange-reds go for it. For me and for the most of cinematographers that's wrong and it's warning not to use the camera in DNXHD or PRORES compression until it will be fixed.

Maybe we shouldn't base this discussion on Coke bottles, but on little red riding hood? If i was to shoot the 1001st repetition of grimms fairy tales, would i get into trouble with this camera?

Definitely YES (until someone show me different) and it seems also in the RAW mode and not just compressed....soon will send something more to show you.

These are images of the red sheet created in the photoshop. Values are: R-255, G-0, B-0
My monitor is calibrated with spyder3 to 5000K, Gamma 2.2.
Images are taken with the same lens at 5000K , f5.6

Photoshop RED


7d RAW at 5000K


BMCC RAW the same image at 5000K.

Basic premise of any camera since it was invented is to take the fingerprints of the reality.

I would not agree with that at all. If you take film as the "gold standard" every film stock has its own character, and people have always chosen different film stocks because of their preference for a particular stock's "look". I don't think the output of any film stock or digital camera is (or is intended to be) a perfect representation of reality.

In addition, if you are seeing a "pure" red from any Bayer sensor camera, that will be down to the manufacturer's chosen post-processing. The nature of a Bayer CFA means a pure saturated colour is not possible in the raw data.

BMD have chosen the colour matrix they use for their own reasons, which none of us are privy to. It is perfectly reasonable to lobby them to provide an alternative matrix which you may prefer, but bear in mind that no matrix will be perfect. It will, like anything, be a balance of different compromises.

Hi Nick,
Thanks for the post. I been talking in a more theoretical way regarding invention of the machine which reproduced reality automatically. You are right, every stock have advantages and disadvantages. FUJI for UK green colors as Ken Loach love it or Kodak VIsion 3 for beautiful yellows as they love it in California. But with film stocks you always have option to change them or to apply different processes and with most of them you will get your portion of RGBY. Here I got one strange stock that I can't change nor use it for the most of the projects. If a director wants red to be red it seems that this tool is not good enough compared to every other camera. BMD should warn they costumers that their camera cannot film red color but giving you an orange instead. If they didn't clarify it at the very beginning then they should fix it as soon as possible. Alexa, Red, F3, all of them will give it to you.

I don't know.

Just for the sake of it, I just took a quick shot of the screen (all "Film" mode) - DNG looks perfectly fine to me, ProRes and DNxHD is off, but that's to be expected, since it's Quicktime which is a PITA on a PC since donkeys years..

Just loaded the material in Resolve applied the BMC to 709 LUT and hit "A"

Grrr.. just got this: Sorry, the board attachment quota has been reached.

Here you go (right click to open in full size):



Here are the original grabs from the timeline, if someone is interested:
https://drive.google.com/folderview?id= ... sp=sharing

I'm still confused (and sorry that i have to ask those basic questions): Are those issues confined to shooting with "video" setting, or "film" as well?

I've preordered a Pocket CC (cancelled my CC 2.5k preorder shortly after NAB), which supposedly features the same sensor and color science, so it might inherit the same color aberrations; yet with Pocket CC shooting (compressed) raw won't be an option for me.

I don't have done any baked in "video" tests, since that would render the use of a raw camera obsolete for me. But even if you need to work that way for some reason and bake in a 709 LUT, the codec has enough flesh, to push the image to the look you like.

I have different experience. You can probably get what you want at the end but after a lot of masking and not sure that it will be perfect at the end, especially with compressed formats. And yes, it affects film and video log.

And yes, it affects film and video log.

Thank you for clarification! For my kind of (indie) work i'll always use ProRes "Film" setting (on Pocket CC), so i hoped this wouldn't be an issue. Yet i'll have to do a lot of postprocessing any way, and luckily i don't have to fullfill customers wishes, only those of our own crew, so i'm confident we'll manage. Chances are that colors won't be very "natural" in our projects any way

Doesn't filmconvert nowadays have a BMCC profile included? Has anyone of the "affected" (or perhaps we all are, though I haven't noticed any orangy reds yet with my Prores film mode recordings) tried if their LUT/filtering is able to "fix" the reds?

Doesn't filmconvert nowadays have a BMCC profile included? Has anyone of the "affected" (or perhaps we all are, though I haven't noticed any orangy reds yet with my Prores film mode recordings) tried if their LUT/filtering is able to "fix" the reds?

I will check it these days but in general I think that BMD should fix it or to give us more realistic matrix in addition to what we got. My iphone is able to reproduce more realistic colors so it's just question of their decision.

I will check it these days but in general I think that BMD should fix it or to give us more realistic matrix in addition to what we got.

Totally agreeing to that, was just curious if filmconvert could help (which would then in turn sort of mean that they created their LUT/matrix with an "orangy red" BMCC as well) ...

I have different experience. You can probably get what you want at the end but after a lot of masking and not sure that it will be perfect at the end, especially with compressed formats. And yes, it affects film and video log.

You don't need any masking for that. Just a simple hue vs hue vector does the trick.

I wonder, why my BMC had no problems to reproduce your red field test.

You don't need any masking for that. Just a simple hue vs hue vector does the trick.

I wonder, why my BMC had no problems to reproduce your red field test.

I don't know Frank, I'm wondering too. I didn't start the topic, there are a bunch of people who complains about the same. I just did some tests and wrote the results.

Doesn't filmconvert nowadays have a BMCC profile included? Has anyone of the "affected" (or perhaps we all are, though I haven't noticed any orangy reds yet with my Prores film mode recordings) tried if their LUT/filtering is able to "fix" the reds?

Soeren, I applied filmconvert luts and it does make things look better but I try it just with solid red not with a real life shot.

Julius,
Did you get any response from BMD about the problem?

Best,
Leon

+1! I am also waiting for the info. Wonderful camera, but this is simply a problem that needs to be addressed!

Cheers,
Milen

I had a real world example of this the other day. We were shooting a clown - with big bright red hair, red nose, and red stripe pants. The nose and the pants were very orange, but the hair was absolutely and positively orange on the ProRes files. Fun! Fun!

I had to shoot a (red!) building of one of our clients. Most of the shoot was done in ProRes video because we had to hand the clips over immediately after. For this specific shot of their building I shot it in ProRes film, video and raw.

I took these clips into Resolve. In raw, everything is fine and more or less how I perceived the colour in real life. In ProRes video it's off, kinda orange. In ProRes film, if I apply the BMC LUT, it's orange again. If I use the Arri log to Rec709 Out, it's fine.

Again, in raw: if I just tweak the colour a bit and use the Autocolour it's absolutely fine. If I put the clip on BMD film and apply the BMC LUT, then it's all orange again.

Should be fixable, I'd say.

Sounds like the in-camera LUT was designed the same way as the BMC-LUT for Resolve. Glad to hear that another LUT does the trick!

Hi guys,

ACES color space does the trick also, check my posts above. I think it does better job than the arri lut.

Cheers,

Milen

Yup, but that doesn't help the ProRes Video mode either. If only BM could fix / change the BMCC LUT or let us use custom LUT's in camera...

I can also confirm as well as Frank, that I do not have any issue's reproducing sufficient red's (I can't say true cause I haven't actually tested RGB numbers but it looks pretty red to me). Had some issue's first week with Video mode as I was run & gunning a project and wanted to see how far the baked 709 could be pushed in post compared to the Film or (Log) setting. I found that loading into davinchi or adobe works just fine and both the Alexa and BMC Log to 709 LUT create a rather ideal red and skin tone starter for grading, sometimes even to much, though that probably is a characteristic of the limited w/b options within the camera itself, unless of course when shooting in RAW.
I have done some basic RAW test's and found I get an even more defined Red when using Lightroom or Adobe Raw in any form because I have complete meta control in PP. If that made sense at all.....

@Leon

No, I have not heard anything yet. The last time I talked to them requesting a return, they convinced me that it does not matter if they give me a new camera because they know there is color issues and that they also know it is affecting most of the cameras. They hope to get the problem fixed by way of firmware.

There are some lucky people who have these camera without exhibiting a problem. What I don't understand is why they keep arguing giving implication that all cameras have no issue, therefore; the problem lays with the operator.

If you have a working camera, go make a movie or something with it and let those of us who still need our lemon cameras get fixed so we can use them to full potential.

I'm not an engineer but I think if custom white balance was given as an option, maybe, just maybe the problem may be fixed. Color/colour is very important in this business (especially in corporate world)and it just have to be right and timing is everything.

I guarantee Black Magic made a choice to favor skin tones off the bat. Skin tones in my eyes are the most important color in a scene with humans. So, they chose to tone the reds down to give nice tan creamy skin tones. The rest can be corrected in post.

I guarantee Black Magic made a choice to favor skin tones off the bat. Skin tones in my eyes are the most important color in a scene with humans. So, they chose to tone the reds down to give nice tan creamy skin tones. The rest can be corrected in post.

This is a fallacy argument. Whose skin tone was this based on? In a logical world of diversity, this does not hold water. We all have different skin tones and only true reproduction of color/colour in my argument makes sense.

There is no single skintone you can nail by numbers, but there are certain characteristics of skintones.
And yeah, it's the most important part of any color science IMHO.

I take a camera that puts a bias on skinton any day, over faithful color reproduction.

[CaptainHook]

This is a fallacy argument. Whose skin tone was this based on? In a logical world of diversity, this does not hold water. We all have different skin tones and only true reproduction of color/colour in my argument makes sense.

There is no single skintone you can nail by numbers, but there are certain characteristics of skintones.
And yeah, it's the most important part of any color science IMHO.

I take a camera that puts a bias on skinton any day, over faithful color reproduction.

+1

http://www.divergentmedia.com/blog/full ... fleshtones

An interesting look above. Skintone of course varies, but not that much in hue.

I guarantee Black Magic made a choice to favor skin tones off the bat. Skin tones in my eyes are the most important color in a scene with humans. So, they chose to tone the reds down to give nice tan creamy skin tones. The rest can be corrected in post.

It can be corrected in post if you shot RAW. The rest is questionable and lot of us sometimes just want to deliver a footage as it is because many times client don't want to loose time and money for CC. , BMD provided such option with DNXHD and PRORES video log but it doesn't work and even it's not always correctable. I didn't saw nowhere that BMCC is only close-up skintone camera and for the rest use another camera. Mr. X likes skintones of BMCC it's his opinion, I like it too but I need my roses red and my client wants his stage actress dress purely red and not orange. I'm shooting landscapes, concerts, plays, architecture etc. and skintone is just one very small aspect of whole picture and almost always is correctable. It seems to me funny to consider wrong color phase as something good. It's camera bug and I already know people who gave up shooting with BMCC due to those issues.

having the exact same thing here.
regarding color and skin tone, let us not forget that skin tone is colors.

i read people comparing audio monitoring vs consumer and the color problem here, this is total nonsense.
A video camera could maybe be compared to a microphone but not some monitoring.
and if a mic is doing the same to frequencies then it's not a correctly balanced mic. that's it.

honestly this shouldn't be like that, at least if we can have a valid white balance function on the bmcc.
I'm waiting for the pocket one and i really hope it won't be the same.

@Leon

No, I have not heard anything yet. The last time I talked to them requesting a return, they convinced me that it does not matter if they give me a new camera because they know there is color issues and that they also know it is affecting most of the cameras. They hope to get the problem fixed by way of firmware.

Wait, what?

BM is acknowledging they have a color issue with *most* of the cameras they've shipped and are working on a fix? Is this correct?

Just out of curiosity, do you have that in writing, like in an email or something?

I can also confirm as well as Frank, that I do not have any issue's reproducing sufficient red's (I can't say true cause I haven't actually tested RGB numbers but it looks pretty red to me). Had some issue's first week with Video mode as I was run & gunning a project and wanted to see how far the baked 709 could be pushed in post compared to the Film or (Log) setting. I found that loading into davinchi or adobe works just fine and both the Alexa and BMC Log to 709 LUT create a rather ideal red and skin tone starter for grading, sometimes even to much, though that probably is a characteristic of the limited w/b options within the camera itself, unless of course when shooting in RAW.
I have done some basic RAW test's and found I get an even more defined Red when using Lightroom or Adobe Raw in any form because I have complete meta control in PP. If that made sense at all.....

Hi Max,
could you post some tests, please. So far I haven't seen a single one that doesn't show the problem. Honestly I didn't quite get the one from Frank, but maybe it's just me. Remember guys when we all saw the magnificent footage from JB the red was not orange. This was one of the main reasons I liked the camera. The other reason was the ability to shoot Prores and DNxHD. In the current situation compressed modes are just not an option for something involving red color in it. Do I regret for buying the BMCC? Not a single bit! Is there a problem with the reproduction of red? Absolutely! Do we have a statement from BMD? Not really! Why is that? Nobody knows!

In the current situation compressed modes are just not an option for something involving red color in it.

That's not the case if you shoot in film mode. It is the matrix used in the log to video LUT which can cause the orange shift. This is the reason I have added the option to switch that matrix off in the latest beta of my FCP X LUT plugin (that option is not in the watermarked demo version yet).

The screen grabs below show the difference between the default BMD matrix, and switching the matrix off and using the saturation slider instead. Using FCP X's colour panel you could refine it still further.


Note: This was done using a ProRes file created in Resolve from John Brawley's DMG files, to simulate shooting ProRes film.

Nick has it right here.

I think what people aren't happy with it actually only a LUT, specifically BMD's "VIDEO" LUT, which is quite separate to the way the camera can actually record colours, especially when you use RAW or ProRes FILM.

jb

Glad to hear it's not a hardware malfunction, but a software design decision, that could be "fixed" in a firmware update. Actually it would be fun if we could create and install different LUTs to the camera ourselves, so that - in theory - one could create a LUT to match other cameras on the set. This might speed up work in some situations.

Nick has it right here.

I think what people aren't happy with it actually only a LUT, specifically BMD's "VIDEO" LUT, which is quite separate to the way the camera can actually record colours, especially when you use RAW or ProRes FILM.

jb

Thanks John,
I think I was one of the first people who pointed that the LUT is the problem and now when you clarify it too I am almost certain. I can't agree entirely with Nick. Yea it' easy when you are a LUT and workflow ninja and you tweak your own plug-ins but for us ordinary users so far the only solution is creating workarounds like using Alexa LUT and additional CC in Resolve and the results are not very satisfying.

PS Nick can you create a custom LUT for Resolve like the one from Hook? It will be very nice!

Cheers,
Milen

So far I haven't seen a single one that doesn't show the problem. Honestly I didn't quite get the one from Frank, but maybe it's just me.

It was quite easy.

I scrolled to grinleon's post with the 3 color fields in this thread, than took the BMC and filmed my computer screen with those 3 field in raw, DNX and ProRes. Than throw em in Resolve, hit "Auto", exported and montaged them side by side.

As you can see, the Photoshop field reproduces pretty well, the orange field varies in tone.
I blame it mostly on Quicktime, since the raw version is pretty good.