Description image

BMCC vs. Canon 5D Mark III: Did Blackmagic Just Get a Full-Frame RAW Video Competitor?

cinema5D_5D_Mark_iii_BMCCEven though Blackmagic announced two more cameras at NAB, the original 2.5K BMCC is still shipping in limited quantities. While the company has been working hard to address the situation, an unlikely competitor has come along thanks to Magic Lantern: Canon’s year-old 5D Mark III. It is now the only DSLR to shoot RAW stills and video, and even though the hack is still in the early stages, it’s becoming clear that it will be a force to be reckoned with at this budget level. The guys over at cinema5D have been working with the hack, and they’ve now posted the first test comparing RAW video from the Blackmagic Cinema Camera and the Canon 5D Mark III. The results are interesting, to say the least.

Here is the video, shot using the 24-105mm on the Mark III and the 16-50mm Tokina on the Blackmagic:

Some observations from their site:

What became clear is that both cameras provide very similarly capable RAW files. You can adjust everything and everything is possible. You can get similar details out of the sky and set all you like in post. Color temperature, tint, dynamic range, no problem. The 5D is just as strong as the BMCC, providing maybe a bit more dynamic range.

Where the cameras differ in terms of RAW is one big thing: Noise. The 5D mark III can shoot indoors at ISO 1600 and there’s almost no noise while the Blackmagic starts to get ugly at this point.
The noise from the BMCC is also apparent when raising the blacks in a shot.

Moire and aliasing is another big big issue on Blackmagic and there is very litte of that on the 5D RAW. The cleanness of the shots of the 5D RAW in general is extremely pleasing and jumps at you when you sit in front of the RAW images. Check some of the dng’s yourself in our other post.

The Mark III’s video before this hack was pretty lackluster compared to a lot of the other options for video. The Nikon D800 had a sharper image, more dynamic range, and clean HDMI, and Canon has only just delivered that update recently. Honestly, while I think the camera as an overall package was worth it, if you were to look just at the video capabilities, the Mark II has about 75% of the quality for around half the price. If you’re willing to install Magic Lantern, however, the 5D Mark III just became a powerhouse in its budget range.


I really haven’t had any issues with aliasing on the BMCC, but you can definitely find them in specific circumstances. While the guys weren’t using the same lens on both cameras, the Blackmagic has a bit sharper of an image — but not by much — and that’s really the crazy part about this hack. Just in terms of image quality alone, the BMCC completely destroyed the Mark III just a week ago.

It probably goes without saying that the BMCC’s firmware is made to shoot RAW, and the Magic Lantern hack for RAW video is still in its infancy. The BMCC certainly has advantages to the hacked Mark III, like audio input with RAW, internal ProRes/DNxHD, HD-SDI, playback in-camera, more re-framing options in post, and a free color grading application (DaVinci Resolve). If the hack becomes rock solid and allows for longer record times, and if you’re willing to deal with the workarounds (and strictly image quality is your chief concern), there is a very compelling reason to take another look at the 5D Mark III if you haven’t done so already (and if you own one, you can see the quality for yourself).

Either way, these cameras would definitely complement each other on set, and there are situations where one might be more appropriate over the other.

Head on over to cinema5D for more on the hack and how you can get started with it (though again, waiting until they fully release it is probably the best course of action for most people).

Link: Canon 5D mark III RAW vs. Blackmagic Cinema Camera RAW — cinema5D

Disclosure: Blackmagic is a nofilmschool advertiser.


We’re all here for the same reason: to better ourselves as writers, directors, cinematographers, producers, photographers... whatever our creative pursuit. Criticism is valuable as long as it is constructive, but personal attacks are grounds for deletion; you don't have to agree with us to learn something. We’re all here to help each other, so thank you for adding to the conversation!

Description image 151 COMMENTS

  • lol, 5D is like a caveman tool compared to the new Blackmagic Production Camera, which costs $500 more.

    • With all ML functions enabled I don’t think so and raw capabilities in MKIII are just jaw dropping… No moire and aliasing… Look the first shot comparisson, the road…

      Also 5D is a full frame camera, all your lenses already working, no crop factor… You can buy 128gb kompubay cards for less than $200.

      As told in the article, the code is in its first stages, only dev soft now… In a few weeks we could have lots of MKII folks working in raw in a 5 years old camera! lol

      But yeah 4K BM camera is just awesome, but we’ll have to wait for how long to put our hands on one?

      • Don’t kid yourself…the result you’re seeing on the road is due to the lens, not the camera.

        Love what ML is doing, but this particular test is pretty flawed in it’s execution.

        Wake me up when you can record 5 minute segments at fullHD in RAW. Then it’ll be something I can actually use for work.

        • I’d be surprised if a lens could create aliasing…

          • Being too sharp, a lens certainly can help create aliasing. Surprised?

          • “Help”, yes. “Create”, no. The implication of the above post was essentially that if you put a lens of the exact same quality on the 5D, it would moire. Which it wouldn’t.

    • By a “caveman tool,” do you mean having interchangeable batteries, having a full frame sensor, amazing lowlight performance, the ability to take stills, and not having to purchase expensive SSD’s?

      • Chris Lambert on 05.14.13 @ 6:24PM

        expensive ssd’s? By my reckoning 140min of 4k footage is going to cost as much as much as 75min of CF 1080p RAW 5d footage.

      • He nailed it, Drew. Caveman tool = much better low light/cave situation performance ;-D

      • ….and the 5d3 is in boxes in stores, ready to be picked, the BMPC is in the stratosphere. Point for Canon there.

        • This ML advancement is still in the stratosphere too…. while this is *very* exciting (and I’m not even a Canon user! Only Panasonic + Nikon for me at the moment, but maybe that could change….) it stills has a fair while to go before we can see it widely used in production (however at the rate they’re advancing every day I wouldn’t be surprised if this changes very quickly!).

    • I got my 5d Mark III for $2700. Oh, and it’s also a stills camera – something that the Black Magic Cameras are not.

      • Right. Black Magic Cameras are CINEMA Cameras. Designed to shoot cinematic style/quality images. The 5D is a stills camera with a video function. Different strokes for different folks.

    • In terms of functionality, maybe (though I wouldn’t be so crass as to label it a “caveman tool” even without this hack).
      The nature of the RAW video recording hack limits the usability of the camera from many standpoints, but if you can deal with the shortcomings, (particularly if you already own the camera) it’s still a pretty decent deal. I mean, the lack of audio, increased rolling shutter, weird workflow, lack of supporting video functionality, and all the other issues DSLRs already had are still there. Being a hack, it’s also not likely to be acceptable for use on budgeted, insured productions.
      That said, the RAW video hack is absolutely substantial. In terms of low-light performance, noise, aliasing issues, and sensor size, it clearly beats out the Blackmagic cameras which are the only competitiors for RAW video at this price point. For hobbyists or independent shooters whose needs are accommodated by this camera’s ability, it’ll be a very attractive alternative to other offerings.

    • I guess I’m in the minority of those who think that the ALL-I/IPB video quality that comes standard with the camera is more than “good enough” for the majority of creative applications. I mean other than visual effects, RAW just seems unnecessary for most kinds of work. Is RAW a lot easier to color grad/correct than ProRes,DNxHD, H264, or AVCHD? Sure, but it’s not like it’s impossible to grade those formats and if the shots are done right it won’t need much grading or color correction to begin with.

      Video production enthusiasts/professionals maybe able to spot the slight extra bit of dynamic range or detail, but audience/clients simply wont because their attention is more on content. I kinda think we’ve become at bit spoiled by technical specs and slight differences in IQ with these camera for the past couple of years, when in the bigger picture we’re just splitting hairs.

      Two years ago I couldn’t have imagined a camera that performed as well in low light conditions that the 5DIII does and that’s huge for folks who are out in the field and don’t have control of lighting. I would never even consider putting my GH2 (great as it is) in the lighting conditions that the 5DIII can handle with ease. Hack or no hack 5D Mark 3 is a great camera for both photo AND video.

      • Raw is a little easier to grade than ProRes, but either are much better than anything else. It’s a combination of a lightly compressed (or uncompressed) stream, 10-bit or 12-bit color instead of 8-bit, and dynamic range too.

        The low compression means you don’t see blockiness/softening from H.264 or worse.
        The increased bit depth means you can grade without seeing any banding.
        If you shoot a darker figure person in front of a brighter background, you can rescue the shot with high dynamic range, and you can’t without.

        ProRes is a lot easier to deal with in post than Raw — just copy the files and throw them in a timeline.

    • Frederik O. on 05.14.13 @ 6:58PM

      The 5D Mark 3 is 2750 $ on

  • Chris Lambert on 05.14.13 @ 5:29PM

    Pre ordered the BMPC but now I’m torn, that image, that range, that price, that sensitivity jaw dropping

    if they can write a audio file from the microphone into that mix im going to get very tempted

  • now ML must do this with the C100

  • How does this change If I’m reading this right, the hack turns an 8 bit 5D Mark III into a 14 bit camera…somehow?

    • Somehow? The hack by passes the H.264 encoder and gets you a raw feed off the sensor. So the camera so no longer an 8 bit video camera. however, there is a world of difference from h.264 and raw in terms of post processing. Raw is going to require much more time and memory.

      • Would it be possible to bypass the H.264 encoder but still output something sensible to an external recorder via HDMI? I guess it would still have to be debayered at some point. Would that rule out getting these resolution and dynamic range benefits in a friendly Prores 422 form on an Atomos Samurai?

  • I think the Joe has a love for the 5DM3. If you look at the detail in the background of some of the videos, the 5DM3 falls apart. Very muddy. On the segment with the pathway in the center…you can clearly see that the 5DM3 turns the center back tree green. That is not on par the the BMC.

    • I don’t own either camera, but whatever mud you’re seeing (which I don’t) is post compression. These are both RAW cameras. The green on the tree in the Mark III video looks like a lens flare.

      • I still think the 5D3′s image isn’t quite at the same level as the bmcc, and keep in mind the bmcc is 2.5k so they really arent the same thing

        • And the 5D3 is a 4K-capable sensor (pehaps even more?). Andrew Reid says he managed to record up to 3k on his tests, btw.

          • but how much is stable? gh2 could do a lot of things but it also crashes depending on whats happening in the scene, RAW might be fine but I really don’t think 4k or even 3k are going to happen

    • @Brian, I’m fairly certain that is lens flare… Look at how it flashes in and out with the breeze changing the incoming light.

  • the bmcc is still probably going to win out for video, for all of the usability plusses (and more) mentioned in the article. Plus, when the speed booster for m4/3 comes out, it’ll help with low-light and crop factor. Yet to see how the camera hardware and ML firmware hold up with continuous use.

    however in my opinion the 5d still wins as overall camera, for its ergonomics and stills capabilities.

    • Exactly. The fact that we are even talking about Mark III video vs. BMCC video is enough to show you that Mark III wins the overall. Mark III vs. BMCC stills isn’t even a discussion. So if you weigh stills into this at all the winner is quite clear. If you don’t, it’s at least a discussion now.

  • I’m under the impression that the 5D’s shots are somehow more “dimensional” – am I the only one here?

  • I think this is absolutely fantastic for people who already own the 5diii, or are just thinking about getting into RAW workflows. I think for the more pro users out there who will be making their living off their equipment, the blackmagic cameras will be the better bet – simply for things like professional connections, SSD media, and reliability. (If ML breaks, that’s tough – if your BM camera breaks, there’s an entire company out there to help you out).

    I know for me – I was always comfortable using ML on DSLR’s in professional settings. I think for this hack, though, I’ll be sticking with BM for low budget RAW projects. I feel like this hack is going to shorten the life span of the camera with such heavy use.

    • I feel the same way. If you own a 5D this is great. BMCC/BMCC4k is the more convenient and has ProRes.

      The margin is narrow now, but I think the BMCC is still the more video/film friendly camera.

      This is exciting, but I feel like people may be getting carried away when they say there is no need for the BMCC. Just a little bit carried away.

  • Anyone who thinks about getting a BMCC can seriously think about getting a mk3 instead. Even if it were better, there is one simple fact: you will never get 14 or even 10 bit over the HDMI, so you have to record internally on CF since SD is not an option.
    One of those 64GB Lexar 1000x cards goes for around the same price as a 512GB SSD. Even if those prices should magically drop you will only get 12 minutes on your 64 GB card! That’s it!
    So if you are that one special person who only films 1-minute videos in dark alleyways, the mk3 might just be for you. Anyone else who is doing films forget about it.
    And if you are doing documentary work … 1st the full sensor size will bite you in the rear and 2nd with the BMCC at least you’ve got the option to use ProRes or DNxHD. With the mk3 you have to go back to H.264 with its bad resolution and dynamic range.
    (DITs have no fun juggling 10 cards on set and it’s even less fun needing one just for that reason)

  • Having just spent some time pixel-peeping the DNG frame grabs from over at Cinema5D, I have to say that as exciting a development as this is, I’m still keeping my preorder in for a Blackmagic camera. Even with a cheaper lens on the Blackmagic compared to the 5D for this test, the Blackmagic beat it by a long shot for detail, most noticeable in the leaves and on the face of that rock.

    The only benefit I can see to the 5D is that it’s a really good still camera too. For some people, that’s important enough that they’ll go with that camera. For me, it isn’t.

    The BM4KPC has the advantage of significantly better detail, global shutter, an easier workflow, SSD recording, in-camera ProRes, and ostensibly better stability and reliability, to name a few. It also comes with Resolve, which alone is worth the extra $500 to me.

    Ideally, I’d have a BM4KPC as my main camera, and a 5D mark III RAW as a B-cam and still camera with me on jobs. Until I can afford that, though, I’ll be sticking to just Blackmagic.

    • Another advantage to the 5D would be significantly better lowlight performance. But when is that important? For me, lowlight performance only really matters on documentary type work, and I would never take a 5D RAW to a documentary shoot. ProRes bitrates are a hard enough pill to swallow. Uncompressed RAW from the 5D is pretty unmanageable for anything except very tightly controlled shoots. The argument could be made that you shoot 5D in RAW for tightly controlled shoots, and go back to h.264 for more freeform documentary type projects. But I really hate h.264. I never want to go back to it. If the ML team can enable wider dynamic range Motion JPEG on the 5D that’s already a big step up, but I’d still prefer ProRes or compressed RAW, and take the hit in lowlight performance.

    • I think that’s understating the 5D’s image. BMCC is still wonderful but it has a serious competitor here.

      Some of the extra perceived detail in the BMCC’s 2.5K raw comes from false detail. The 1080p on the 5D is much cleaner, and less noisy.

      A good comparison would be to upscale the 5D’s raw to 2.5K size and compare on a 2.5K display. I expect a bit more sharpness from the BMCC but is it worth it when the trade off is some false detail and more noise?

      2.5K resolution is great and 5D3′s full frame likely to top out at 1080p, but the 1:1 crop mode could yet reach it’s full potential and give us nearly 4K from a Super 35mm sized area with a very fast card. The 5D3′s DMA memory in the camera does 700MB/s! That is faster than any SSD and not a bottleneck.

      By the way, you can get the 5D Mark 3 on for 2750 $.

      • Also, please clarify what “false detail” you’re referring to in the BMCC. I’ve never heard of or seen any false detail in that camera, unlike DSLRs which have in-camera sharpening. What you’re saying is misinformation at best.

    • If you like a tiny(relative to full frame) sensor, moire and aliasing, terrible form-factor, limited internal battery, poor touch control and a hard-to-see lcd screen, the choice is clear, go with the BMC. The slight increase in resolution is moot when there’s moire all over the place. Also, in lowlight the BMC stands zero chance up against the 5D3.

      • Fred, I was talking about the Production Camera, which does not have a tiny sensor relative to “full frame”. It’s Super-35, which is the gold standard of filmmaking. The 5D is the oddball by filmmaking conventions, being incompatible with most cine glass. The BMCC does not have a tiny sensor, either. It’s nicely situated between Super-35 and Super-16, both of which have beautiful aesthetics. Even the Pocket Camera doesn’t have a tiny sensor, in my opinion. Super-16 is a very nice format, and far better for documentary work and certain styles of narrative work, in my opinion, than Super-35 let alone “full frame”. To answer your other points, moire and aliasing are present on the BMCC, but not terrible. The form factor is not suited for handheld work, true. Neither is any other true cinema camera I can think of – Epic, Alexa, F65, etc. The internal battery is not a problem for me, as I use external batteries anyway. The LCD screen is no less hard to see than any DSLR I’ve worked with, and I like to use an EVF anyway. Lowlight is the major upside for the 5D, but again, lowlight to me is only important for documentary work, and for those projects I would never consider uncompressed RAW at 7Mb per frame. ProRes is about as high bitrate as I care to go when shooting documentaries.

        Just to reiterate, this is an exciting development. When I can afford it I’d love to roll with a BM4K as my main camera and a hacked 5D RAW as a b-camera for DOF shots and stills. Until then, I really don’t see this as better than the BM4K, BMCC, or even BMPCC for my purposes.

        • “and for those projects I would never consider uncompressed RAW at 7Mb per frame.”

          7MB is for 3.5K raw frames. It is 4MB per frame for 1080p. Less than Blackmagic.

      • It will be very interesting indeed to try and shoot a documentary about homeless people in a dark alleyway! perfect camera for discreet miseryporn! Rejoice! The 1 minute shoot limit will also bring forth the obligation of multiple takes for every two sentences, which is fantastic, as I like to have plenty of coverage when interviewing! DIT’s will also love me. I mean, it gets kinda boring when you have to flip only two 512 gb cards per shoot….twelve 64gb cards sound like more fun!
        I mean…the essence of cinematography is lighting….and if you cant light…well…great…use what is available. Available light is generally used mostly on documentary style work…but the raw workflow of the 5D III is very very limiting on set for a documentary filmmaker…I can see it working on more controlled studio work, but then again, the BMC seems like the better choice….

        • In today’s market its almost pointless to buy a new camera because several years down the line it will seem redundant. Blackmagic in a way have already done this with their current camera by announcing the 4K cam.
          If you were to buy one or the other you need to ask yourself what you need it for and what you regularly shoot. The 5D mark III in my view is a much smarter choice when it comes to owning a camera because its more versatile than the Blackmagic both in its form factor, low light capabilities and now with RAW video fixes the issues we had with h.264 compression. Not to mention its an excellent photography tool :)
          If i’m shooting a budgeted project i’ll rent something more ‘professional’ but if i’m shooting for myself or if its low budget then the 5D is perfectly adequate (even at factory settings).
          I also see this raw capability being used in documentary situations to capture beautiful b-roll footage.

  • Inspirational camera choices ahead…! But there seems to be an issue
    At 13 seconds the 5D has a major glitch. See this screen capture:
    I saw similar issues in other users tests. Also It owuld be good to know if the Green Glow on the 5D shot in the background trees is a flare or some sort of artifacting.

    • It’s a lens flare. I would expect there to be plenty of glitches right now. This is a pre-Alpha software development.

  • FatRick says, what about overheating. There is some push back from ppl that say the hack is not worth the rosk because of overheating. So to the ppl that have actually shot with the hack, what do you say to this point? Is your camera heating up more than before?

    FatRick say, thank you in advance for your response.

  • This is great for people who already own a 5D mk iii (I do mainly for photos, but also use it a ton for video). I would not recommend anyone to go purchase a MK III based on this hack. There are better options out there then this, but it’s great that we get this extended capability for free. Will I use it a ton? Probably not. But I will use it from time to time when needed. I will still keep my BMPC preorder, and will probably use both cameras when I can.

  • Just FYI: The lens on the BMCC was the 16-50mm not 11-16.
    And concerning overheating I heard that the 5D RAW only uses the images directly from the buffer as debayering is done later on and that should actually be a lot less processor intense than H.264, so result in less heat, but you never know, right? So far it doesnt seem like a problem.

    • Ahh, totally read that wrong. Fixed.

    • FatRick says, thank you Seb I’ve heard the same and so that is why i asked for the ppl who actually have shot with the hack to reply.

      perhaps Jo can investigate this further and write something about it? Because I think for some ppl this is all that is holding us back at this point.


  • There were some serious big blinking artifacts in some of those shots. Particularly the second more close up shot of the guys sparring. If they can fix all that stuff this might be a serious option.

  • Hope you guys know this is just the (ALPHA) Raw……the upgrades only get better from this point. I know it hurts….the feeling all BMCC users are getting at this moment is the same feeling Canon users got when they heard the news BM had a pocket camera that shoots 1080p RAW…….that feeling is priceless.

    • What feeling they should get !!! What is priceless is to see how people ego can make them stupid. Did blackmagic stole your child and it is not even Canon who is giving us the RAW. They did everything to block it, how come the raw feed is high definition while the h265 output is low rez. What blackmagic did is give to the low budget crowd some very high quality camera. Apart from full frame (not everyone wants it and perhaps low light) the 5d3 is not giving much more than the BMC and surely much less than the 4k production camera. Somehow if they get it working on the lower end models in the $ 800/1200 price range, then it will truly hurt the blackmagic cinema cameras.

      • Oh stop crying……did you read the other comments……Some people on here is trying to down play what ML have done with the 5D3 & 2….was just poking and making fun of the situation….by the way your tone on your comment sound just like the people I’m making fun of…..carry on.

  • From ML forum:

    “The best you can with 600D (550D/60D) is: (limited by hardware SD transfer rates)

    - 960х540, upscale it to 1280х720 (24 fps max)
    - 1280х400, tight aspect 3.20:1 (24 fps max)
    In Canon’s 720p – lineskipping 1.66х, so you should upscale to restore correct ratio:
    - 1280×400 upscales to 1280×664 – near to 16:9 (24 fps max)
    - 1280×360 upscale to 1280×600 – 2.13:1 aspect
    - 1280×320 upscale to 1280×533 – 2.40:1 cinemascope aspect

    It is only 1280px wide, but pretty nice looking though”

  • Boy, this is a head scratcher.

    The BMCC is a more pleasant, better colored, more detailed picture. It’s easier on the mind to look at it. If you go over it with a trained eye you may find a few very isolated instances where the MarkIII wins. But overall the BMCC wins. To me the comparison of these two cameras is more a study of sociology than video quality. The BMCC 2.5k is $1000.00 less than the MarkIII—that’s without Magic Lantern hacks, and such, for the MarkIII. And even with any extra costs for the MarkIII the BMCC is still better. Why not just get the BMCC and be done with the tinkering? And what will it be for the MarkIII when the BMCC 4k comes out and costs the same as the MarkIII? There will be comparison videos of the two. And the BMCC 4k will be even better than the BMCC 2.5k. But I have a hunch there will still be people insisting the MarkIII is comparable to the 4k. That’s why I say this is a sociology study for me and not a technicals of video quality study.

    BUT! The 5D MarkIII takes excellent photos. There is absolutely no question about that! But if you’re looking for great video just get the BMCC 2.5k, or coming very soon, the BMCC 4k, and be done with all the upgrading of the MarkIII which still doesn’t make it as good a picture as the BMCC.

    • I think that’s understating the 5D’s image. BMCC is still wonderful but it has a serious competitor here.

      Some of the extra perceived detail in the BMCC’s 2.5K raw comes from false detail. The 1080p on the 5D is much cleaner, and less noisy.

      A good comparison would be to upscale the 5D’s raw to 2.5K size and compare on a 2.5K display. I expect a bit more sharpness from the BMCC but is it worth it when the trade off is some false detail and more noise?

      2.5K resolution is great and 5D3′s full frame likely to top out at 1080p, but the 1:1 crop mode could yet reach it’s full potential and give us nearly 4K from a Super 35mm sized area with a very fast card. The 5D3′s DMA memory in the camera does 700MB/s! That is faster than any SSD and not a bottleneck.

      • I’m not sure what you’re referring to with “false detail”. I’ve downloaded the DNGs and the extra image resolution in the BMCC is very real and very noticeable. Yes, it does suffer from aliasing here and there -is that what you are referring to? But for the areas where there isn’t aliasing, it is indeed a sharper camera. And it should be, I would expect so from the specifications.

        Regarding noise, I rarely expect to shoot raw in a situation that I can’t light the scene properly.

        • Look at the false color artifacting on the pebbles in the path up the middle:

          Those artifacts come out clearly on test chart shots, including the one Andrew from EOSHD posted back at the beginning of the BMCC hysteria, comparing it to the Red Epic.

          Look at the test and see how the BMCC renders the boxes with false colour streaks and how false it resolves the individual lines. Here is the link:

          Conclusion from EOSHD: “This problem occurs because the 2.5K sensor does not quite have the resolution to provide each red, green and blue photosite with enough information about detail, such as a fine line which only covers a red photosite, and not the blue for example. ”

          Please inform yourself and google other test that indicate those problems.

          • A comparison of the $3000.00 BMCC 2.5k to the $32,000.00 Red Epic 4k is fair?

          • Also, the BMCC 2.5k is in its first generation.

          • Gene, the Epic is a 17,500 $ camera (at least the Epic-X brain).

            Please read the entire conversation with Sanye. I just said that the 5D Raw does not suffer from these false details and it is just the Alpha version of the hack.

          • So in the talk of comparison to the Red you were trying to make a point about the BMCC v MIII. You did and I miss read. Sorry about that.

            BTW, $17,500.00 for a Red Epic brain—WISH I COULD AFFORD ONE!!! Really! :-)

    • Eric Jolley on 05.14.13 @ 8:11PM

      I don’t get the negative comments. Magic Lantern is an extremely powerful and free upgrade for Canon owners that continues to add value to our cameras. This hack will definitely get better. How much better who knows. But if you already have a 5D Mark III, and were contemplating an additional purchase, say on the BMCC, this will definitely make you rethink that. Especially if you’re on a limited budget. But if you own neither, yeah, I’d say go for the BMCC.

    • If you like a tiny(relative to full frame) sensor, moire and aliasing, terrible form-factor, limited internal battery, poor touch control and a hard-to-see lcd screen, the choice is clear, go with the BMC. The slight increase in resolution is moot when there’s moire all over the place. Also, in lowlight the BMC stands zero chance up against the 5D3.

    • “That’s why I say this is a sociology study for me and not a technicals of video quality study.”

      Indeed, this is one of the more sensible comments I’ve seen on this.

      While this is undoubtedly great for Canon users from a technical perspective, I’ve found it far more fascinating watching people’s reactions :-P

      I’d rather have a Blackmagic camera + a mid range DSLR from the previous generation (as remember, stills technology has been improving at a relatively slower rate in the last few years than it was in the early years of digital photography. While digital video is still accelerating ahead every year at a breathtaking rate. So an older mid range DSLR is quite acceptable for 98% of still photography, while I would like to have the most latest cutting edge dedicated video tools that I can get such as a Blackmagic Design camera).

  • I think the cameras are highly comparable. It seems like a lot of the naysayers are people who jumped on the BMCC bandwagon and are now afraid that Magic Lanter created something better. This is the reality of technology. I wouldn’t say the BMCC is better or worse than the 5D3 but its comparable now. The overall image may go with the BMC because it was built for the image. However, the 5D3 takes stills, is more ergonomic, has a better battery solution, and a full frame sensor for better use of lenses. Not to mention Magic Lantern makes Canon DSLRs a completely different and better product.

    As an owner of a 5D3 I can’t wait to try this.

  • Wow. The thought of comparing the footage of these two cameras two weeks ago was laughable.

    Congratulations, Magic Lantern team! I hope their innovation is well-rewarded over time. Even with RAW aside, they are the only reason my 5D Mark II is still such a useful camera—it’s kept me from upgrading, at least, and for that, I literally owe them thousands of dollars. I hope they can continue to bring such fantastic development to the Mark II and Mark III and really get every drop out of these cameras they can.

    Full frame, 35mm, 14-bit raw recording for <$4k. Absolutely amazing! I couldn't buy a Canon GL2 for that just a few years ago!

  • Good thing they didn’t try and move that 5D3. :-) The jello is off the charts.
    If you already own one, GREAT. Always loved ML.

    Otherwise, I’d keep your powder dry. Much more to come.

    • Apparently, for some, we’re not supposed to pay attention to those issue behind the MarkIII curtain.

  • vinceGortho on 05.14.13 @ 8:25PM

    Any news from kineraw on their mini S35?
    Its supposed to launch next month.

    • nigel Thompson on 05.14.13 @ 9:59PM

      I really want to buy a kineraw but my issue with it is you dont get a proper SDI or HDMI signal out …. seems its some degraded 720 crap just for monitoring ….. not even an 8bit signal

      Was hoping to attach it to my pix240 for prores recording. SHRUG oh well.

  • It seems that very soon we will almost have too many choices. And some people seem to be acting like that’s a bad thing…

    • More accurately, it seems people will take sides no matter what happens. OPTIONS are always a good thing. Ownership bias and/or company loyalty tend to cloud that.

  • Great for everyone. Maybe Nikon, Sony and Panasonic wil be show something new about RAW video. Because now the video fight is only betwen Mark III and Black Magic, who is gonna buy a nikon d800 for video?

  • And the reason why you compare 160 ISO 5D3 picture to 800 ISO BMCC picture is what?!?

    • Presumably because they correspond to roughly the same sensitivity and also because ISO actually doesn’t mean anything in raw?

      A better question would probably be what they were hoping to convey by presenting those numbers.

      • Well, to be more accurate, they don’t.

        When someone that understands the camera does this test alongside a 5D I’m sure that they’ll find 800 ISO on BV1 (Blackmagic Version 1) looks more like 5D @ 1600 ISO.

        Next, it’s not just metadata. 1600 ISO on Blackmagic’s different than 800 ISO, in raw and ProRes. You should likely stick to 800 in all scenarios, it’s what the camera’s balanced around.

        Lastly, the person shooting this test had little idea how to use BV1/Blackmagic, admits it in the Cinema5D post, and yet still goes on to sort the results as definitive. Oh well, though. Misinformation flourishes as usual.

  • While its great seeing all these developments in raw/4k/10/14 bit stuff… I think it’d be swell to see some industry efforts focused on bringing down prices for higher frame rate capabilities:). Yeah I know the fs-700 is out there and a good but, but maybe we could start seeing some more cams with that ability coming in at lower prices, maybe less concern including raw capability and more focus on pretty clean 1080p high frame rate capability around the 3-4k mark.

    Maybe that’s a bit much to hope for soon :P

  • For me it’s more a question of giving money to a company who isn’t holding back technology and will look after customers in the future rather than capturing the market and sitting on new tech. Props to the ML team for all the hard work and being awesome. I can’t wait to give it a try however, my next purchase won’t be a Canon.

  • I would like to see the whole sharpness issue fixed by ML independent of raw. I would love to see even 8bit at that sharpness. The whole workflow is a little out of the way to begin with. At any rate I know they have worked on the encoder to get a higher bit rate. It would be cool if they solved the muddiness and hopefully this would lower the transfer rate and allow for better recording time and maybe develop a sort of MagicLantern log that could retain the dynamic range.

  • Any word on this hack coming to lower end cameras?

  • I am buying a new camera in July. I have shot with and love Canon products. I am still leaning toward BMCC (maybe the BMPC – just because of the global shutter). For me, I would love to shoot raw all the time, for a lot of what I do, it isn’t practical (storage, workflow etc.). 5DIII RAW vs. BMCC RAW – this Canon hack looks pretty legit. But when my only choices are RAW vs. h264, I’d rather go BlackMagic, and shoot raw when I need it. Unless ML does some other amazing stuff (which is totally possible), I think the BMCC offers more practical flexiblity (today).

    Kudos to ML, and they might tweek their code which will cause me to re-think that, but right now, I am still leaning toward BMCC. Competition is great.

    • Well said, this pretty much sums up my thoughts on the matter as well. And beyond that, the 135 format is too large for the bulk of what I shoot. I need to be able to get reasonably deep focus without stopping too far down. Even Super-35 is pushing it. If it’s reliable, the 5D RAW hack will make a fantastic B-camera for DOF shots, and for shooting stills on set. But I really can’t imagine using it as my primary camera.

    • BMPC doesn’t have a global shutter , only the 4K has that. The BMPC has a rolling shutter the same as it’s mid-range bigger brother.

  • does anybody know if it´s possible to build a connector from cf card slot to sata. i think about running a cable out of the cf slot from the mk3 to a ssd drive. then we could use ssd´s direct with the 5d which would be awesome. there are adaptors from cf card to data but i could not find the other way around.

    • That would really be something. A hardware hack for faster data rates really would be awesome. I know high speed cameras have a large physical buffer they fill, and something like that could be very interesting especially if it could give higher frame rates or higher resolution video.

      I’d love to see the guys who do cpu overclocking jump on with the magic lantern guys to find a solution that gave 14 bit raw 22 megapixel images at 120 fps before the camera burst into flames :)

      • It would be best if they could route the RAW footage via HDMI to record to a ProRes HQ recorder like the Ninja 2. Given what I have read, the 5D’s Live view is actually a RAW feed, it would seem quite reasonable to assume that this feed could be diverted via the HDMI out. This is Canon’s chance to repay its faithful consumer base and add value to their product now that the cat is out of the bag.

        • From what I’ve read over on the development forum, that’s not gonna work, as LiveView has already been squashed down to 8bit before sending it down the HDMI lead. Perhaps someone who likes code more than me can correct that.

  • does anybody know if it´s possible to build a connector from cf card slot to sata. i think about running a cable out of the cf slot from the mk3 to a ssd drive. then we could use ssd´s direct with the 5d which would be awesome. there are adaptors from cf card to data but i could not find the other way around.
    maybe the powering of the ssd would be a problem. could also be that the power from the cf slot is enough. the power of an usb is enough…

    • I think they’re looking into that, but at the moment recording isn’t possible whilst the card door is open. Either they have to code around that, or butchering the latch on the card door’s going to be necessary for it to work.

  • Cool. 1DC vs 5DMKIII ML could be interesting too…

  • I would say the BMCC performs better in terms of latitude. And that’s not apples to apples. The BMCC in RAW is 2,5K. They should compare both in that resolution.

    • Agree, side by side comparison of 2.5K upscale on 5DM3 anyone?

    • The 2.5K still bothers me. Couple hundred extra pixels for cropping and sharpness and maybe overall quality? I’m not saying it’s a bad thing I’m just saying it bothers me. But the name is Black Magic ‘Cinema Camera’ Shouldn’t rule cinema distribution out but 2.5K won’t cut it. At least no one has been comparing the BMCC to RED. Comparing it to a 5D Mark III ML is more realistic.

      • That’s 640 more pixels horizontally, it’s significative.

        • Yeah, in terms of image quality it might help but what about distribution?
          If people are going to use this as a ‘cinema camera’ how is 2.5K going to work? I don’t mean to be brash or anything I seriously would like to know. I understand the RED workflow shooting 5k creating 1080p proxies for viewing and editing, and pushing it back out to a higher resolution no problem.

          • You do realize the Alexa isn’t a 4K cam and it seems to have no issues with distribution of the content. Many movies are projected 2K in theatres anyway. Stop drinking the RED Koolaid. 4K isn’t the holy grail Jim would have everyone believe… at least not yet.

          • Most films are still distributed and projected at 2k so lack of resoluction on the BMCC would not be an issue.

          • 2K or even HD acquisition is not a problem for movies being distributed commercially anywhere!

      • James Anderson on 05.15.13 @ 10:41AM


        Ever watched Skyfall in the theater?

  • they didn’t understand how to exposure BMC correctly.

  • BMCC is still sharper.

  • Mark III should smoke BMC in terms of resolution. 22MP is no joke. We need 1:1 comparison.

  • The only problem I have with all of this, is storage. We need some type of way to record either to an external recorder via HDMI or maybe some type of CF to HD converter. I am very excited to shoot raw but I am not willing to drop money on cf cards if its going to cost just as much to purchase a Black Magic Production Camera. I have already invested in San Disk 90mb/s cards and supposedly those wont be fast enough for full 1920×1080. Im thinking if the resolution is so much better than the H264 codec possibly shoot 720p and upscaling will still prove to give a better image with less data.

  • it seems like the 5d raw is better in the highlights. hmm.

  • Hello,

    Questions for Joe please

    My understanding is that currently this would be the largest sensor shooting raw including the red dragon sensor and alexas. What would be the significance of this in terms of overall image quality if any in comparison to other super 35mm sized sensors, assuming that it becomes fully operable?

    In terms of size, would the sensor of mk3 be comparable to the Imax film?

    How does “color science” compare between bmcc and mk3? so far color science has been oneof bmcc’s greatest strengths – a la “baby alexa”

    Also would the BMCC’s global shutter be its greatest feature over Mk3 in terms of shooting motion?

    Any updates on the Panavision 70mm sized sensor digital cinema camera?

    Thank you.

    • Well, I’m not Joe, and he probably knows way more about this than I, but I like writing, so I’ll give you a response as best I can. :-P

      The 5D Mk. III raw hack would allow one to shoot raw video using a larger sensor area than Red’s Dragon sensor at 6K, and definitely larger than the Alexa’s Super 35 mm-sized sensor. However, it would still not be the largest raw-capable camera sensor. Vision Research’s Phantom 65 Gold camera, for example, has a significantly larger sensor at 51.2 x 30.5 mm, approximately the same size as 65mm medium format film. Of course, that camera is a $3,000/day rental.
      Here’s a reference:

      In terms of image quality, there’s no inherent connection between “good” image quality and larger sensors. That said, some of the advantages of using larger sensors tend to come in the form of cleaner and lower-noise images, better high-ISO performance, and higher pixel resolution. One of the greatest strengths of the 5D Mk. III has been its low-light performance; I can’t think of any other raw-shooting camera that is as clean at high ISOs (though the Canon C500 might be a contender since it’s very similar to the C300, a camera already slightly better than the Mk. III in low-light.)

      The sensor of the Mk. III is nowhere near the size of IMAX film. IMAX film is even larger than the size of that aforementioned Phantom 65 Gold camera. Canon’s 5D Mk. III has a 3:2 imaging area of 36 x 24 mm which is usually cropped during video recording to a 36 x 20.3 mm area at an aspect ratio of 16:9. IMAX film, on the other hand, has an imaging area of 69.6 × 48.5 mm at an aspect ratio of about 1:4, which is nearly four times as large.

      Not having tested both the Blackmagic Cinema Camera and raw from the Canon 5D Mk. III myself, I only have what others have produced to go off of. Color science, despite its name, is extremely subjective in nature; different people will find different color representation pleasing. Canon’s DSLRs seem to be known for their pleasing skin tones, for instance, and the added bit depth, at 14 bits per channel, of the raw output from the Mk. III should only enhance that color fidelity. With plenty of BCC footage out in the wild, I’ll leave you to assess it yourself. The video right above in this article directly compares the two. Besides, when shooting raw, while color science does come into play a little bit, it’s not nearly as important as it would be when shooting compressed formats.

      When shooting any fast-moving objects or when moving the camera quickly in a shot (as during a swish pan), rolling shutter will always play a role in your final image. So, yes, the global shutter of the Blackmagic Production Camera 4K (not the Cinema Camera, as you indicated) would completely solve that issue altogether. Not only does the Mk. III have rolling shutter normally, but, from what I’ve seen, it’s even worse when shooting raw video.

      Not being press, I’m not privy to quite the same level of information as Joe may be. I don’t know much, nor have I heard much, about the Panavision 70 mm camera.

      I hope that helped!

  • One additional Q

    If this hack becomes operable in the 1dX- would there be any significant advantages or capabilities over the ML mk3? I remember reading that Canon warned the might of their legal team to anyone hacking the 1dX – in terms of 1dc- makes one wonder what they are worried about………….

    • Since the hack is not available for the 1D X, and people associated with Magic Lantern has already stated that they will not be working on creating a version of the software for that camera, I can’t know for sure what, if any, differences would arise between it and the 5D Mk. III in terms of how they would handle raw video. My best guess is that it would function virtually the same. The 1D X has been said to have a sharper image, slightly more exposure latitude, and noticeably better high ISO/low-light performance than the 5D Mk. III, so those qualities may translate over to raw video as well. One other advantage could come in the form of longer recording times since the 1D X has a faster processor setup and can record still images at a higher frame rate. With improved internal data speeds, one would most likely be able to record for longer periods of time, perhaps at higher data rates. Perhaps, due to the nature of the flexible recording resolutions available in the hack, it could be possible to record at higher resolutions at a 1:1 sensor crop than would be possible on the 5D Mk. III, though that’s all speculation.

      • Recording times on the 5D MkIII are a limitation set intentionally by Canon, and not a limitation of the technology in the 5D3 itself. Its to keep taxation reduced in the purchase price of the cameras – if it were to record 30mins or more, then it falls into the category of a video camera, which is taxed higher – hence the limitation.

  • Still less detail than the BM2.5K, far less shooting time, and the Mark III costs $500 more. Right now the only camera worth my attention is the BM4K.

  • What is all that nasty-a$$ moire on the BMCC? I’m surprised and disappointed, I didn’t think it performed so poorly in that area.

  • So glad I didn’t pick up the BMCC month ago as it turned out it is worth the wait, as a Vancouver wedding photographer & Videographer this will help those of us that has spent our investment on the 5D Mark III for photos and video. Well done Magic Lantern and I can’t wait to work on my personal project when the new firmware release.

  • question…since the mark ll raw video and bmc have moire will this be corrected on the bmc 4k camera coming out?? are there tests out now??

  • We want these hack on the Nikon D800 !!!! Seriously, Please!!
    (not a native speaker)

  • The BMCC still has a stop more DR. Download some DNG files. The BMCC has more. In the shadows the Canon wins. Gotta admit, though, this is pretty amazing.

  • Wow that’s great! Just from watching though, I like the BMCC a little better…. I dunno why. Maybe the color science. Not a whole lot of a difference though!

  • Nicholas Edwards on 05.28.13 @ 11:33AM

    Been using Lightroom 4 and LR Timelapse templets to skip a step in processing of the Raw DNG files, think it is also a superior grading tool compared to adobe camera raw. Saves a step, give it a try!

  • As a beginner in 4k world I am going to buy a BMC 4K to start with. There is a huge set of lenses. I afford just to buy 2-3 lenses. What is the choice?

  • there is NO contest the BMCC blows it away see a real comparison ALL BS, BMCC BLOWS CANON AWAY!!!!!!!!! go see a real comparison review online that actually shows the REAL deal not this laughable joke of a video. hers a REAL review no contest! oh and now a 4K full frame model!! with global shutter. sorry canon its over for you and your overpriced DSLR cameras.

    • That comparison is between the BMC and the 5D MKIII’s built in H264 video. It was made long before Magic Lantern hacked the MKIII and got Raw out of it… The comparison is obsolete and irrelevant…