Canon 70D

[josepvm]

All other ML supported cameras do not have a Dual-Pixel sensor, so yes, the 70D is different.

If you think you can do it better and faster, simply do it yourself. All the code is available to everybody. But complaining will not speed up the development.

[jeddo_drone]

All other ML supported cameras do not have a Dual-Pixel sensor, so yes, the 70D is different.

If you think you can do it better and faster, simply do it yourself. All the code is available to everybody. But complaining will not speed up the development.

How do I do This? Is it easy enough to do? I would like to have a go and and hopefully get it released to a few other 70D users. If anyone else could help to make a network of open source 70D ML developers it would me much appreciated.

[josepvm]

There is already a network of open source ML developers, built around this forum.

The source code is here (as announced in the opening post of this thread):  https://bitbucket.org/nikfreak/magic-lantern/commits/branch/70D-merge

And the information needed to build the code and to help in development is already available in this ML forum. Start reading here: Getting started with development

But no, it's not easy to develop Magic Lantern, because it's a reverse engineering project. Canon will not help you with any info about how their cams work. Everything needs to be discovered by disassembling and analysing the firmware. This needs a good knowledge of ARM assembly, and countless hours of learning and experimenting .. with no warranty at all that you will achieve ever a viable way to get the camera do something you want ... that's why you cannot anticipate the rate of development. It's ready when it's ready.

There are simple tasks in which you can help to improve and debug Magic Lantern, with minimum programming skills. And any help is very welcomed here.

But keep in mind that the whole project has been possible thanks to a small group of highly skilled ant talented developers. And they do it simply for the fun of doing it, and spending the hours they can and want to spend of their free time.

 

[jeddo_drone]

There is already a network of open source ML developers, built around this forum.

The source code is here (as announced in the opening post of this thread):  https://bitbucket.org/nikfreak/magic-lantern/commits/branch/70D-merge

And the information needed to build the code and to help in development is already available in this ML forum. Start reading here: Getting started with development

But no, it's not easy to develop Magic Lantern, because it's a reverse engineering project. Canon will not help you with any info about how their cams work. Everything needs to be discovered by disassembling and analysing the firmware. This needs a good knowledge of ARM assembly, and countless hours of learning and experimenting .. with no warranty at all that you will achieve ever a viable way to get the camera do something you want ... that's why you cannot anticipate the rate of development. It's ready when it's ready.

There are simple task in which you can help to improve and debug Magic Lantern, with minimum programming skills. And any help is very welcomed here.

But keep in mind that the whole project has been possible thanks to a small group of highly skilled ant talented developers. And they do it simply for the fun of doing it, and spending the hours they can and want to spend of their free time.

Thanks for the info. It's really confusing to see where you can do little jobs. Do you know how I can help to do a few small jobs while they do other stuff. I don't know much about development.

[josepvm]

Follow the forum dicussions, specially in the "Developing Magic Lantern" section, including the subforums located there (General Development, Modules repository, Reverse engineering) and you will see the features that are being developed or improved. The most common help needed is to test and debug these features in all cams, so you can help staying tuned and testing on your 70D eveything needed.

But first you need to learn how to build the ML source code, setting up a development machine. Doing it in Linux is usually easier (It's possible also to build ML with OSX and even Windows, but the configuration is harder). If you don't have already a  computer running Linux, you can use a VirtualBox to run a Linux virtual machine with the ARM compiler and other software needed to build ML.   

[starbase64]

better slowly than nothing

[Walter Schulz]

I don't know much about building a development machine on OS X. Just followed dfort's instructions and everything is running smooth on Hackintosh (PC-based OS X machine).

[dhagen]

Need a little help, and searching general ML didn't seem to help any. I'm using a 70D with this ML firmware (which is pretty awesome by the way) to turn off all overlays and use the HDMI out for a video stream. The problem I'm having is that after a while (camera is on AC power) it "sleeps", taking the shutter down and asking me to hit "play/pause" to resume video mode. Is there a way to keep it from timing out like that other than randomly pressing the shutter halfway to keep the camera "awake"? Tried to look through all menu options and google around, but not having any luck with options. Thanks!

[Walter Schulz]

Try Prefs tab -> Misc key settings -> Sticky HalfShutter ON

Documentation in https://builds.magiclantern.fm/#/features -> Prefs -> STICKY_HALFSHUTTER (link)

[nikfreak]

Sticky Halfshutter was reported as "buggy" long time ago and is disabled since then. Compile from source and enable it in features.h to test if it works and then report back your findings.

[janjan]

Hi,
thanks for all the hard work.
Any idea if the disabled focus menus, which enable focus staking, will every be fixed?

[Wille_W]

Hello!

New user of ML. Very nice product!

I have a problem though that i saw was listed as fixed earlier. I can't get the zoom to af point to work in picture preview. Anyone that know the status on this?

chrisfrancia:
Upgraded from Alpha5b to Alpha7b(ICEMAN).

New build is fantastic!!! On the look out for bugs, will post as I test things in new posts. So far,

1. Change log items
-Audio Meters: OK
-Movie Crop: OK
-Zoom to AF Point: OK

[nikfreak]

Anyone that know the status on this?

Working for me. You better read 1st line of 1st post and do it.

[hiddenmagic]

Dear Magic Lantern team,

First of all many thanks for the work you all did on developing magic lantern for canon 70D!

I have been following this forum topic since the start, and I have downloaded ML software several times, but I didn't install it yet.

In order to intercut and use footage from the 70D APS-C sensor with compressed but sharp/clear (100mbit/s, 8-bit)UHD footage from a small sensor mirrorless camera, I have to use  ML RAW video in the largest format possible.  As I sadly have read, continuous 1080p 16:9 RAW is not possible on 70D hardware.

There are many alpha/beta versions of magic lantern.
If I want fps override enabled, which one is the most stable, and can I still download it?
Then, if that would be possible, can I set footage at 20fps 1080p 16:9 continuous or even 12 fps and speed up  2,5k footage in post to 24p, or make 24interleaved video in post from 12fps progressive?

Do you have other suggestions ? As I understand, bitrate modifications aren't possible neither clean 422 1080p hdmi, in any version of alpha or beta software?

Many thanks, in any case!!.

[OlRivrRat]

      @ HiddenMagic

   To get Limited FPS Override You will need the > 8th August: Alpha-7a2 < Build ~ It is No Longer supported by the

M'L' Wizards & so not available here for DownLoad but I keep it in My archives & can send You a Copy if You EM Me Direct @

[email protected]

   Let Me know which Version You need.

                                 ORR ~ DeanB

[jeddo_drone]

Hi magic lantern team. I finally got my 80mb/s 64gb Sandisk SDXC cards delivered yesterday. I mounted magic lantern in the cards and it is working fine until I stop recording (not in raw) and the camera shuts itself off. I then, after waiting for it to turn back on itself for 3mins, turned it off by the switch. Then the Red LED light just stays on for about 2 mins then the light turns off. I checked the card and found multiple crash logs after this happens over 3 times over a span of 3 recordings. Would you like me to post the crash logs on this forum? I really need help ASAP. If I can't get this fixed I will need to dismount magic lantern until the firmware gets updated. Thanks in advanced.


Sent from my iPod touch using Tapatalk

[nikfreak]

Post exact instructions (step by step) to reproduce it. Also list your settings / modules activated

[jeddo_drone]

Thanks. Lets star with the activated modules:
- adv_int
- autoexpo
- bench
- deflick
- dot_tune
- ettr
- file_man
- lua
- mlv_play
- mlv_rec
- mlv_snd
- pic_view
- raw_rec
- silent

Movie tab settings:
- REC indicator: remaining time
- REC key: half shutter
- Gradual Exposure: on
- HDR video: off
- image fine tuning: default
- movie tweaks: default
- RAW video (MLV): off
- RAW video: off
- delayed start: default

Steps to crash:
- press record
- move around a little to make a basic shot
- stop recording (crash may happen)
- press play (crash may happen)

Update: the crash is happening far less often now but I have another problem with RAW video.

Settings:
- RAW video (mlv): on, 1824x1026, 1.62x, 16:9
- RAW video: 1824x1026, 16:9, card warm up 64MB

Steps:
- I press record and it says preparing then starts recording.
- then it says Frame skipped, record stopping.
It then stops record and when I go to review the video it is really choppy and pixelated.

I have tried both raw and raw mlv but neither works.




Sent from my iPod touch using Tapatalk

[dmilligan]

Disable lua.mo. It uses lots of memory and is the likely source of the crash.

Your raw video "problem" is not a problem, but expected behavior. In camera preview has to use the slow CPU to do rendering, therefore it is slow and low resolution. Copy the files to a computer, convert them and play them back, they will be fine.

[nikfreak]

Too much modules activated at once.
I think it could be the LUA or mlv_snd module causing your issue but that's just a guess.

I would welcome it if you find some time and disable all modules. Afterwards enable them one by one but try to have always only one activated and reproduce the error please.
Should be the best way we can find the module which causes your error.

[OlRivrRat]

   Just for Ref' ~ I Usually run these Modules with No Issue ~

- adv_int
- autoexpo
- bench
- deflick
- dot_tune
- dual_iso
- ettr
- file_man
- mlv_play
- mlv_rec
- mlv_snd
- pic_view
- silent

                                 ORR ~ DeanB

[ShootMeAlready]

Hidden, there's a lot more to consider then just raw & 16:9
For instance 70D can shoot raw, but overheat is a big issue, so ISO 800 is about the max to prevent Hot Pixels.
Most will have better low light with H264 which shoots nice at ISO 1600/3200.   Ofcourse flat pic style will extend your DR.
You can shoot raw continuous at 1600 @ 2:39 @ 24 fps.  I would suggest ISO 400 max, because it not only controls Hot Pixels
but minimizes low light shadow banding. 
All-I compression does a very good job of limiting shadow banding, so higher bitrate wont improve on that.
Really if you need 16:9 then H264 flat pic, with ALL-I is very versatile and gets you best overall picture.

422 is possible only through HDMI out recording.  Canon limits res. to 1600x800, so you have to crop in post.
However why crop 422 H264, when you can get 14bits depth and shot raw at about the same resolution (no crop needed). Either needs a rescale.  I can appreciate the 422 HDMI out, if you want to live stream at 720p.  Good for blogging.
720p @ 24fps shoots @ 16:9 is continuous raw, thats your best 16:9 mode for 14bit raw.

Now shooting raw 1080p / 16:9, @ some low fps, means you have to frame interpolate.  Ive seen a few folks do this for
proof of concept brief clips. But no one recommends this, and it means your rendering raw & using a third party app to interpolate (which varies based on the app).  A rather lengthy workflow, if time is money and you want 1080p raw just buy a 5DmkIII and have a life.  If you really need 16:9 14 bit raw, its still the cheapest way to go.    Still you lose the DPAF & touchscreen AF, which is why most people want to shoot video with a 70D in the first place. 

[Starmaker5280]

Since I am new to the forum on here... Am I just not seeing the download for the 70D?

[jhon_wane]

Hello,
I'm trying hard to get DUAL ISO to work, without much success. 
I'm using AETTR+DUAL ISO (usually 100-1600) to get the maximum data from the exposure.

When i take a picture of blue sky,clouds and trees/grass the exposure for the sky is perfect but everything else is pitch black in the dual iso CR2 file. When i drag the file into CR2HDR (latest version, from last build,dec 2015) the DNG file is exactly the same - no data in the blacks, and I need to drag the exposure a lot to the right in order to see some data. This adds a lot of noise.
I am using LR6 or ACR 9.1 - the results are the same. not good.
Maybe i dont know how to use this function, but i thought that after the DNG conversion the blacks/shadows will pop out automatically.
They do not pop out at all.

Any help will be welcomed, I'm very frustrated...

[jeddo_drone]

Thanks guys. This is much appreciated. I'll do s little more testing later on and see if it is completely gone. Also, what is this FPS override? I feel like I'm missing out on something. 


Sent from my iPod touch using Tapatalk