Atomos Ninja 2 VS The New RAW Magic Lantern firmware

Hey all,

Just wanted to get some info and opinions on the two new updates that have recently come about for the 5D markIII as I was going to buy an Atomos to record at a higher quality, but am thinking that with this new Magic Lantern upgrade, I might not need to!!

What are the latest details on specs for the RAW Magic Lantern firmware? ....And HOW does it compare to the external recording onto an Atomos Samurai???

10bit vs 14bit I guess is one....any others??

Bring on the Pro's and Con's!!

Cheers!
 
The 5D3 raw workflow is painful. But the image quality is a giant step up from internal recording and a huge step up from uncompressed HDMI to the Atomos.

Right now things are changing so fast, development is literally happening each day now. Things are happening by the hour, so no one can say for sure what are the capabilities and what are the specs.

That said, (as far as I can tell right now, at this moment), this is what is working on the 5D Mark III with a UDMA 7 1000X CF CARD:

Full frame 1920x1280 at 24fps continuosly until the 4GB file size is reached
16:9 1920x1080 at 24fps continuosly until the 4GB file size is reached

Might work but has some tearing and stuttering for now:
1:1 crop mode 2560x720 at 24fps continuosly until the 4GB file size is reached

also heard there may be a 1280x720 48fps mode.
 
Last edited:
Hey all,

Just wanted to get some info and opinions on the two new updates that have recently come about for the 5D markIII as I was going to buy an Atomos to record at a higher quality, but am thinking that with this new Magic Lantern upgrade, I might not need to!!

What are the latest details on specs for the RAW Magic Lantern firmware? ....And HOW does it compare to the external recording onto an Atomos Samurai???

10bit vs 14bit I guess is one....any others??

Bring on the Pro's and Con's!!

Cheers!

There is also a MJPEG mode for the new hack, so you don't necessarily need to shoot RAW.
 
the samples from the new ML i have seen are much shaper than anything direct from the camera, or from a ninja. I would say its a far better choice. Worst case you try it and then later buy or rent a ninja if you are unsatisfied. The new ML is a win win
 
that's the crazy thing with ML raw and 1.2.1. once ML ports to 1.2.1, you potentially have a killer full frame stills camera that also records 4:2:0 24/60p h.264 to CF, has a clean 4:2:2 hdmi out for recording as long as your hard disk will allow, and records 2k RAW at 24p to CF. intriguing, to say the least.
 
^ And that's for 1920x800 and 24 fps

At 1920x1080 and 24 fps, a 64 GB card will hold only 12 minutes of footage (and a 4 GB file will be 45 seconds long, but I think they'll lift this restriction at some point, through file spawn).

And processing those files with optimal image quality (i.e. importing them directly into AE through Adobe Camera RAW: ) is painful, with palyback at less than 2 fps in a fast workstation.

But the image is gorgeous and we'll all jump through all the hoops if we want the best possible IQ. (and you don't lose the ability to record H.264 if that's enough for a particular project!)
 
Hold on...it does Mpeg recording as well?? What quality is the Mpeg?

A minute seems a little crap to me...Anyone know if they are making improvements on it? Like for slow-motion or 25p?

Might hold off until then..
 
Hold on...it does Mpeg recording as well?? What quality is the Mpeg?

A minute seems a little crap to me...Anyone know if they are making improvements on it? Like for slow-motion or 25p?

Might hold off until then..


Sorry to bear bad news Simon, but Magic Lantern does not do any new video encoding. So no on mpeg or mjpeg encoding. It is linear raw in a special single file format. Once on the computer there is a program they wrote called raw2dng that splits the file into the individual raw DNG files for post processing. One DNG for each frame of the video. There is also talk of another program they are working on that may make mjpeg video instead of DNG.

There is a lot of confusion and bad info going around about what can be done with Magic Lantern, mainly because people cannot read and/or comprehend.

What Magic Lantern is doing is grabbing raw video data from a buffer memory area of the camera. This buffer was already there, not made by Magic Lantern. From this buffer area Canon processing makes the live view output to the LCD, h.264 video, and HDMI out.

What is special and exceptional is that the Magic Lantern guys figured out where this buffer memory was, that it contained raw video, how to crop it, and how to write it to the memory card.

I hope this makes sense to everyone and will give you all a better understanding of how the raw video hack is working and what is possible to achieve.
 
A minute seems a little crap to me...Anyone know if they are making improvements on it?

I'm not connected to ML, but what I recall hearing when discussing this is that file spanning (which gives longer record times) is possible, but not a priority at the moment. They'll get there eventually when this is ready for primetime, but I guess there's other more important things to work on first. Patience...
 
Back
Top