As most of you are probably well aware, the past few months have been an exciting time for owners of Canon DSLR’s – specifically those with a 5D MKIII. The ingenious team at Magic Lantern have developed new firmware for Canon DSLR’s that now allows you to shoot raw video on your Canon 5D MK III in the format of DNG sequences (not to be confused with Cinema DNG). Not only can you shoot raw, but you can also select a variety of new frame sizes, allowing you to even shoot with anamorphic lenses on the 4:3 2560 x 1280 mode.
Although the firmware will also work on other Canon DSLR’s, to get the most out of it you really need a 5D MK III. For example, if you were to use the firmware on a 550D or 600D, at this time you can only shoot at a maximum resolution of 960 x 540 (if you’re not in a crop mode). It’s certainly still great to have the option of shooting raw on these cameras, but it’s much less practical than having full HD raw capabilities.
Since these developments from the Magic Lantern team have been ongoing for some time, I wanted to wait until the dust settled a bit before writing about it as I was hoping to get a better grasp of exactly where it was going. Before I get into my thoughts on the ML firmware, I should preface it with this: I truly believe that what the ML team is doing is miraculous. They have taken a prosumer level camera that was only designed to shoot heavily compressed h.264 video, and have allowed it to shoot stunning raw video. That is simply amazing. Even more impressive is the fact that their firmware has allowed for the Canon 50D to also shoot raw, which previously had NO video function at all! With that said, I don’t feel that this development is necessarily a game changer – not directly at least. More on that below.
I’ve shot on Canon’s (specifically the 5D/7D) literally countless times over the years. They have paved the way for other video oriented DSLR’s and have proved time and time again to be amazing tools. Even though I’ve always had a bias towards the GH2/GH3 over any canon DSLR (based mainly on resolution/IQ), once the raw hack was announced I started to wonder whether it would be worth it to invest in another DSLR kit, specifically designed for shooting raw on the MK III. Ultimately I decided against it and here is why:
The practicality of any given camera is extremely important to me, and the 5D with raw hack is far from practical right now. I regularly handle raw workflows from EPIC/Alexa/Blackmagic, etc. and have no issues whatsoever dealing with raw material. But typically when I’m shooting on a DSLR as an A camera, it is on lower budget productions that will benefit much more from having an easier set up during production and quicker turnaround time in post. Ae we know, all cameras all have their strengths and weaknesses, and different cameras are of best suited for specific types of jobs. With the 5D (or any DSLR for that matter), the sweet spot has been productions that need high quality video in a small form factor that is low cost and efficient both on set and in post. Unfortunately, what you gain with the 5D hack in terms of raw quality, you lose in terms of the practicality of shooting cheaply and efficiently on a DSLR. For this reason alone, the camera isn’t suitable for the majority of my DSLR based productions.
The cost of media is another big issue for me. If you want to shoot on the 5D MKIII in raw mode, you will need the best CF cards money can buy. And at about $300 – $600 a pop (for a 128gb cf card), you’re only getting about 15 minutes of footage per card. This will be an issue for smaller productions as the cost of the media and the storage (and duplication later) can be very very high. Often when I make this point to people they will counteract it by pointing out the high cost of shooting on Redmags on the RED. But the fact of the matter is, on productions that can afford to shoot RED, the cost of media and storage is not an issue. And while the 5D MKIII in raw looks fantastic, it’s not a RED camera and never will be. Large productions will seldom if ever choose a raw 5D raw over an Alexa or RED. So ultimately this is still a camera for indie film makers on a low budget, not large scale productions. And those indie films need to be very careful about where the money is spent.
Next I want to talk about the image quality, which is a huge improvement over the compressed h.264 you’ll normally get out of a 5D – But for me, it’s not mind blowing enough to justify the workarounds and extra cost. Don’t get me wrong, I feel that the quality from the MK III in raw mode is the best footage from any DSLR today and is truly beautiful. There is no question about it. But a well composed shot on a MKIII raw vs a well composed shot on any other half decent DSLR will not make a huge difference (unless you’re in an extremely poorly lit situation). And it especially won’t matter to clients that are paying for DSLR level production as they will never know the difference. Take a look at this recently released comparison video of the 5D in raw and with h.264 compression. The differences are absolutely there, but you can judge for yourself whether they’ll make or break your project:
The last point I’ll make is the reliability of the camera. So far I have heard very positive things about the reliability of the firmware hack in every day usage, but on any paid shoot I wouldn’t be comfortable using the camera as there is at least some risk that the camera will fail. This risk might be minimal, but I’ve learned my lesson the hard way. On one particular shoot a couple years back, I was shooting on a hacked GH2, which I had previously shot on for hours on end with no issues. But for one reason or another the hack gave me issues on set and I was unable to record more than a couple minutes at a time. Luckily I had a backup camera with me, otherwise we wouldn’t have finished the shoot. Since then, I’ve only used cameras with hacked firmware for my own projects and never on a paid job. There will surely be lots of people that feel it is worth the risk and will use their hacked cameras on paid shoots, but to put it in perspective, no professional shooter that I know has currently hacked their 5D.
So this brings me to my original point of why I don’t think this will be a huge game changer in the indie film world. Even though many will disagree with my points above and workaround it all to use the 5D to it’s full potential, right now those numbers aren’t large enough. As I mentioned above, professionals that make their living using a 5D are unlikely at this point to hack it and risk losing their camera to overheating issues or other unexpected problems that could arise. If the hack proves to be stable over a long period of time, I am sure that many of these DP’s and camera operators would eventually give in to hacking their cameras, but while the developments have been very fast – they are still evolving and still very premature. Just a few days ago ML unlocked a method to get nearly 14 stops of dynamic range out of the camera (as seen in the video above). Again, I think this is mind boggling. But it could very well cause permanent damage to your sensor or cause other problems with the camera. We just don’t know yet.
A year from now when the firmware has been stabilized and everyone knows it’s safe to use it, I think it will simply be too late for this ever to have the impact that many are expecting. By then many other raw cameras will be available that will not only be better in quality, but lower in cost and much more practical. And I’m not just talking about cameras like the Blackmagic Pocket Camera that exist today, I’m talking about the cameras that we’re going to see released in the next year. Whether it’s the Digital Bolex, a new GoPro camera, one of the big manufacturers changing things up, or a completely new company coming out of left field – there will be many many more raw cameras in the coming year. Budget raw cameras are clearly a reality now and there are plenty of manufacturers willing to capitalize on the fact that the independent film community is demanding raw cameras at affordable prices.
Only time will tell, but as brilliant and talented as the ML team is, they are still working with hardware that was never designed to shoot raw and will ultimately be limited by the stability of the camera and that will inevitably affect how quickly (if at all) it is adopted by the general film making community. I only wish Canon would take notice of what Magic Lantern is doing and design cameras with these capabilities and functions in mind from the get go. If they don’t do it soon, they’re going to be out of luck as other smaller manufacturers are now starting to make Canon products look overpriced and in some cases obsolete.
If all this hacking/tech talk is making you feel a bit geared out, take a read from one of my recent non-gear related posts – The Importance Of Story