July 30, 2013

Want to Know How Magic Lantern Really Works & How It's Progressed So Far? Check Out This Video

Magic_Lantern_logo-whiteYou might think there is quite a bit of "magic" involved in the Magic Lantern firmware hack for Canon cameras, but have you ever wondered what's really going on? Back in May, Georg Hofstetter and Michael Zöller spoke at the Open Source 2013 LinuxTag event, explaining exactly how Magic Lantern works and what it's doing on your camera, some of the origins of the project, and how the team continues to innovate -- like bringing RAW video to cameras that had never had it before. Check out the full presentation below.

Here's the description on Vimeo:

Magic Lantern is a Free Software add-on for Canon EOS DSLR cameras, that offers increased functionality aimed mainly at DSLR enthusiasts and power users. It is very similar to CHDK for Canon compact cameras and it runs alongside Canon's own firmware, by hooking into the startup process.

This talk presents a live demonstration of Magic Lantern, showing some of its unique unctionalities and covers the hardware and software internals of Canon's camera operating system DryOS. It also shows that programming your DSLR is quite similar to programming any embedded ARM device in plain C. Finally, since Magic Lantern is not endorsed by Canon in any way, we also present some legal aspects that every firmware modder should be aware of.

What I find truly fascinating with Magic Lantern is that all of these amazing advancements and collaborations have happened in a team structure that is completely different from what you'd probably find at Canon (or most large companies for that matter). While most corporations are very top-down, Magic Lantern actually works horizontally, and there are no team leaders dictating how things should be done and what the developers should be focusing on. This is one of the reasons things have moved as fast as they have. While this can lead to disorganization, it also means that people are only working on what they are passionate about, which in any profession usually leads to better results.

I don't know if a structure like this can work on a much larger scale, but I have a feeling that with less middle managers and bosses and more autonomy with individual workers, you could do some amazing things at all of these camera companies. For example, at Google, employees are encouraged to work about 20% of the time on whatever they'd like, encouraging people to be creative and take risks to come up with new ideas. Could you imagine if a large camera company had a policy like this? I'm sure they would be blowing our minds right now with what's really possible.

Links:

Your Comment

17 Comments

It's easier to do the horizontal structure in the creative fields, such as the software, writing, music or journalism. A 1-hr TV drama episodes are written independently by 4-6 different contributors, with the showrunners then tying the process together. With the large industrial concerns, this is an impossibility. The best someone like Canon can do is to separate medical imaging from the consumer video/photo, though I assume there's still considerable overlap in electronics and assembly.

BTW, great advances in productivity was made possible by the use of the assembly line, where every nut and bolt has to be installed on its own schedule. The Magic Lantern is not involved in the physical manufacturing per se, so they can have various groups working on the different parts of the program. Plus, they are not a profit oriented group either.

July 30, 2013

0
Reply
DLD

I agree with DLD here.

Additionally, when one factors in how much of the production process for Canon and other electronic hardware is outsourced to places like China, the ability to have an equal or equivalent contribution made by all involved doesn't seem likely to occur any time soon.

Knowledge-based economies tend to be horizontal, but the industrial reality of mining and manufacturing which underpins the hardware making that exchange of knowledge possible is as vertical and hierarchical as it's ever been.

July 30, 2013

0
Reply
Dolly

While it is indeed an interesting question, weather or not the "horizontal" structure actually leads to faster an better results than the usual hierarchical order (probably it does, motivation plays a big role here) I do have to disagree with Joe.

It sounds a little like you are trying to imply that the ML team did something that Canon didn't manage to do. I on the other hand am convinced that they definitely have the capacities and could have offered RAW video five years ago, if the actually wanted to. It's just that they fear to cannibalize their higher-level products. To me, this really quite obvious. If Canon had tried to bring RAW video to the 5D it would have probably been a piece of cake for them.

What's so amazing about ML is that they achieved all this as a hack, without having access to Canon dev kits and knowing the system from the ground up (probably some of them actually do by now)

July 30, 2013

0
Reply
Heiko

I'm not sure they could have to be fair as the write speeds of memory cards when I alot of these cameras were brought out would not have been able to handle it. Cost vs speed vs storage makes shooting on CF even right now on anything less than a narrative drama look insane unless someone figures out a way to easily and safely go from a CF slot into a SSD drive it's going to be yr's before all the people who are raving about shooting RAW are going to be able to cover that live event, interview, full documentary day shoot, wedding etc and then go home and process it all. as you could easily spend the price of the camera again on memory cards right now just to shoot a day!

July 30, 2013

-1
Reply
Chris Lambert

But I think ML idea wasn't to create a raw camera for weddings... Raw means a lot of information captured, even on compressed raw formats...

They're trying to push the limits of these cameras and a lot of us should be happy to go out and shoot a video clip, short or whatever and have a full taste of the power the raw gives... It has a price...

Even prores 422 seems a bit overkill for weddings or event shooting depeding on the budget. H264 is just fine and that's where the C100 shines IMO.

July 30, 2013

1
Reply
Alex Mand

Totally agree I just used that as an example as I think a lot of people are screaming RAW is amazing and loosing their heads over something which they are not even slightly ready for unless they are shooting a drama and can afford to have 3-5 minimium cf cards, a dit to handle the encoding etc etc

July 30, 2013

0
Reply
Chris Lambert

don't forget: the camera companies aren't trying to do the same thing as ML. ML are trying to innovate and maximize what they have at hand, and liberate the consumer/user as much as they can. canon takes what they have at hand, has vertical meetings and decides how to hog-tie the product's capabilities, manipulate and hose their customer base. i'll never look at that f'ing company the same after that BS clean hdmi out crap they fed us for six months, just to jerk around BMC.

July 30, 2013

0
Reply
sjk

Beside the lack of sufficiently fast cards, any sort of computation extensive codec generates a lot of heat (anyone seen the fan on the side of SR5?) and quickly drains the battery. Besides, I am unconvinced that ML tweaks are anything more than a small niche from what already is a small niche market (high end DSLR). I have also argued in the past that, by the time a full mod program is developed and released, the market is likely to produce something similar (GH3 or BMD CC) anyway.
.
Still, these type of improvements are, of course, remarkable in increasing the value of the product for those who already own the piece.

July 30, 2013

0
Reply
DLD

The raw hack does not produce more heat. You dont know what you are talking about.

July 31, 2013

-1
Reply
Heather

I have a friend up in Canada that had not shoot video till late last winter. He pick up a Canon t2i. He loves what ML is doing. Even know he is shooting on SD card and getting about 20 seconds per shoot he is using raw video all the time know. He learn the work flow for doing it. Here a link to a video he shoot with the t2i..

http://www.youtube.com/watch?v=ZXpvxNUhgus

July 30, 2013

0
Reply
Michael Bishop

I love ML. I think it makes the Canon DSLR useable as a video camera. In fact, I wish that I had all of the capabilities of ML on my C100.

But, there seems to be a massive misconception by many DSLR filmmakers about how Canon approaches their DSLR product line. Many people seem to think that if the ML team has managed to create all of these extra features in Canon DSLR's, then Canon must be intentionally crippling their cameras in order to protect their profits on their higher end cameras. This may be the case with the 1DX and the 1DC, but it is the not case with their lower end DSLR's. The reason that Canon "intentionally cripples" their cameras is because the vast majority of the buying public does not want these features on their camera. The vast majority of the buying public would actually choose a different brand or model if a salesperson showed them all of the extra features of an ML camera. Most people want simple and easy to use. Even if Canon added all of ML's features and then locked it behind a specialized menu system so that the average consumer couldn't access it, it would still make the average consumer uncomfortable and likely to buy a different camera.

Regarding the ability to output RAW, the number of people who want this is miniscule. Seriously. Imagine the sales pitch. Salesperson "Plus, this camera has the ability to shoot RAW video. All you have to do is buy a number of these really expensive memory cards, be prepared to swap them out every few minutes, and have a massive amount of hard drive space. Once you have the footage you can take it into this complicated program and try to figure out how to make the footage look different than the way you shot it." Customer "But, this Nikon allows me to shoot 20x as much footage on these much cheaper cards. Plus, I can immediately post it to youtube." Salesperson, "But with RAW, it is easier to make your footage look different than what you saw in the monitor". Customer, "Why, would I want to do that?"

July 30, 2013

-3
Reply
Casey

Well being a retied auto worker and shooting video and photography has a hobby I don't think I will ever make a movie. My friends are older and don't want to act in a movie but one day I might want to make a documentary. I keep think of stuff to shoot. I do like to learn about it. That's how I found this site. I join the Indianapolis Camera Club to learn more about photography . So I think that the consumer like me will find what ML is doing is great and hope the camera manufacturers gives use more for our money one day.

Like on the photography side everyone say shoot in raw.

July 30, 2013

1
Reply
Michael Bishop

Sorry, I didn't mean to reply to your video, I just meant this as a general post. I think it is good that people are enjoying using the RAW capability. I only shoot RAW for photographs, but it does mean that I spend far more time finessing my photos than a person with just a point and shoot.

You auto worker comment reminded me of my analogy for ML. I used to have a VW GTI with a turbo and I would visit tuner sites on the web. With just a tweak to the car's chip you could pull 50 ftlb more torque out of the turbo, so everybody on the tuner sites would complain about how VW was crippling their cars. Chipping, however, also caused the car to have a noticeable turbo lag and a tendency to jerk into power. I continued modifying my car and in the end it could do far more than any other car VW produced. I loved it but nobody else even wanted to get in the car with me, much less drive it.

July 30, 2013

0
Reply
Casey

You guys should watch the whole video if you haven't. Magic Lantern has done much, much more than RAW video, but that's obviously a headline spec because this is a more video-oriented site. They've done tons of stuff on the stills side that Canon hasn't bothered to do on high-end or low-end cameras. Stuff like ETTR and taking a photo when there is movement are features people would absolutely want right now, even if it were only on the higher end.

July 30, 2013

-1
Reply
avatar
Joe Marine
Editor-at-Large
Shooter/Writer/Director

Also you have a full set of cheking features, such a peaking, live histogram, vectorscope, zoombox. Timelapse features onboard. Small increments on lower iso (ISO 50).

Also it is rock solid. you can use it in production with no problems or risk of loosing footage or something.

Even RAW is now in good shape for more serious tasks. And you can convert it to prores or cineform and keep a good source material to start with.

ML team are just awesome!

July 30, 2013

0
Reply
Alex Mand

Notice how Canon has left ML and its developers alone? With so many downloads it's bound to have stirred considerable interest in a now sagging camera manufacturer. (off 6% this quarter)

July 31, 2013

1
Reply
RJR

arent they fron the us?? they sound like a loquendo voice ..

August 2, 2013

0
Reply
pablo