Messages

It's ready when it's ready.
There are no timeframes within ML project.

I was hoping for a more reliable answer, but I guess that would be too easy... so I guess I'll have to do my own estimation.

Canon 5d Mark III started alpha around September 2012, never reached beta until early 2015.

Canon 6d is started alpha december 2012, currently in limbo http://www.magiclantern.fm/forum/index.php?topic=3904.0

Canon 7d (not mark II) started alpha december 2012, still in alpha...

So drawing from that info since 70d alpha started January, 2015 this will most likely take much MUCH longer to reach beta and highly likely it won't reach stable release within a reasonable time frame.

I totally agree with the difficulty of developing and reverse engineering. Its very time consuming and can also be scary at times. Having some web development background, I know it can be a little intimidating to think that testing a particular code may just conpletely mess up all your work. In web developing you can back things up, but since you are actually risking an expensive camera you can potentially lose something permanently when testing ML code.

That being said, I would certainly be interested in knowing a time frame by which the project is expected to reach beta. 1-2 months 6-months to a year?

Nice. Yosemite is beautiful. I'm planning on doing some shooting there too but its 6 hours away from my place.

Yes I did notice the grain at 0:52 as well when in full screen HD. Perhaps it's the file you chose to upload to vimeo. They encode it down to h264 5mbps so it's best to upload in h264 for best results. Youtube and vimeo both recommend 10-20 mbps h264 uploads for best results.

If it was in the clip before upload, then I'm not too sure what would cause that. Perhaps you were in auto iso and you shot the scene with too little natural light.

Great work

Apparently with that setting I was only able to get to about 26 minutes of MLV footage (yes I set it in exFAT) before it ran out of space -- so the real question is should I set the resolution to smaller sizes in order to try and pass the 30 minute mark or would you rather have me get a bigger card? lol

I'm thinking of upgrading my SD card for this. As stated before 70d can only take advantage of up to 41MB/s but tests show it's actually in the 38MB/s. If 26 minutes at that resolution would take 64 gb, its probably a good idea to buy the biggest possible capacity of sd card which is the 256 gb.

Extreme pro 95MB/s seems overkill so would this suffice for maximum performance with 70d RAW? -> SanDisk Extreme 256GB U3/UHS-I SDXC with 4K Ultra HD, Up to 80MB/s Read;60MB/s Write- SDSDXN-256G-G46

Great writing on top of the video editting. Where di you get your background music?

How did you shoot this? Did you film it through glass or did you submerge your camera?

This is amazing

No, not really....  As already stated, it's a hardware limitation of the 70D.   If you need 20+ minutes of continuous footage at 1920x1080 24fps 14-bit RAW, then get a 5D Mark III and put ML on it.  The CF card interface of the 5D3 is plenty fast enough to accomplish your needs and ML runs beautifully on it for your RAW shooting pleasure.

5D Mark III is certainly up there when it comes to future purchases. I'm comparing it to GH4 which has 4k, but only has 8 bit color depth.

30 minutes limit in RAW mode for 70D has yet to be confirmed. And there is no 4 GB limit in 70D. Canon implemented file splitt without frame loss in H.264 some time ago for 650D and others.
All other cams are doing seamless recording for +30 minutes with latest ML builds.
Wondering why nobody tested it yet ...

There seems to be a bug in H.264 operation, though. ML (for other cams) will automatically restart recording after 29:59:59 break (loosing about 0.5 seconds during transistion). 70D's switch seems not to work, according to testers.

My mistake 4gb limit was not 70d but due to the fat32 disks. It's not that big of an issue for me since it doesn't stop the recording. 30 minutes is when it stops and that is usually my main issue since I sometimes find myself recording along steep cliffs and slippery terrain that me not needing to go back multiple times to press the record button is not only more convenient but a lot safer.

I do wonder, again forgive my lack of knowledge of ML, but could I possibly adjust the bit depth to perhaps 10 bit and record close to 1080p? That would be the ideal trade off?Something like 1536x864? 1536x864x24x10/8=40mbps.

So it seems to be a hardware limitation.

As far as im aware of the current limitations of 70d are the 4gb, 30 minute limit, and also the 41 mbps write speed.

So if i plan on shooting 30 minutes i would need to reduce the resolutions to around 1500x600p. (1500x600x24x14/8=38mbps) Drawing from this, ML Is preferable for film projects that usually have 1 minute scenes and not for documentary type of productions that have 20+  minutes of continuous footage.

At this point im wondering if a 1500x600 RAW footage would beat a 1920x1080 h264. That will give me less detail when it comes to resolution but more detail when it comes to color depth. Dont take this as a slgith against ML and all the work that went into it,  but since 1080p is now becoming industry standard, im guessing 1080p with h264 would probably be a better bet than a 14 bit raw video at half the resolution since i doubt upscaling 600p to 1080p would cut it even if it is 14bit.

Sorry if this had already been adressed. I read about 10 pages and i havent seenit adressed directly. I havent installed the alpha build but im planning on testing it once it gets to beta. 

The current duration time for maximum resolition of 1800something x 764 is at 15 seconds. I read somewhere that there is a "movie restart" function to circumvent the 29:59 recording limit but not used for the higher resolution raw files. 

My question would be is the short duration time being worked on? Meaning do the developers intend to eventually increase the recording time to continuous for maximum resolution.

Im new to magic lantern so im not sure if this is a common ML quirk. Im planning on using ML for recording hour long footages so im interested in knowing developments regarding this issue.

wrong assumption. digital sensor, digital camera, digital file format.

Makes sense. I'm still wondering if say a 720p RAW video would still be better than a 1080p h264 since the 70d Magic lantern is still in development and it only has 720p as with continuous shooting and I'm planning on shooting hour long videos.

2. You can convert MLV to DNGs (image sequence that AE can import) using any of several converters found here.
If you choose to use MLVFS then you get "virtual" cDNGs more quickly than making real ones, so you save some time.
Here's the workflow I use (may or may not be right for you depending on your setup).

Thanks! I bookmarked your workflow. What I'm worried about is MLVFS since it seems to be a bit tricky when using windows.

Thanks! Thats was a useful link

I have a few questions regarding the workflow associated with using Magic Lantern. I'm very new to this and I'm planning on taking the plunge. I'm the guinea pig and I will probably recommend ML to my friends who are using 70d in case I figure this out.

1. What is the average files size of a RAW video file? Assuming its 1920 x 1080 RAW video files for 30 minutes.

2. How do you import RAW video files into Adobe After Effects? Does it work natively with AE or do you need to install a plugin for it to work with RAW video?

3. I'm assuming RAW is not a digital format, so it upscales pretty good. Would movies saved in RAW be future-proof when 4k starts becoming more ubiquitous?

Im planning on installing ML as well and i need for continuous shooting similarly. Ive read that the 30 minute limit is for tax reasons, the 4gb limit sd card limitation but it doesnt interupt the shooting it just makes a new clip.

Im wondering the same thing. It would literally add tons of value to this version of Ml if it removes that limitation.

Also wondering of 1080 continuous RAW is in the works.

Its keeps pausing for me every few seconds. Is that just my connection or is that the actual footage?

Im planning on using 70d ML for a long cotniuius shot but im not sure if 720 on ML looks better than 1080p on h264.

This certainly helps the comparison. Thanks for the demo.