Magic Lantern 2024 summary, 2025 preview

[names_are_hard]

Did 2024 ML seem quiet to you?  In reality, it was busy!  Read on for a summary of 2024, and a preview for important plans for 2025.

What happened in 2024?
Raw video works in experimental form for some modern cams (7D2, R).
Raw video works fairly well for other modern cams (200D, 6D2).  This is not finished, but it's good enough to be released for testing.
Time limits for MP4 / MOV recording increased beyond 29m59s.
Significant documentation for developers added.
Build system greatly simplified.
The above two items mean it is easier than ever for developers to join ML project.  Got programming friends who like cameras?  Ask them to consider joining!
EDMAC breakthroughs on new cams; we know how to find equivalent functionality on new cams that depends on EDMAC.

What will happen in 2025?

There will be a Magic Lantern release.  This will cover old and new cams.

There will be a new official repo; actively maintained and regularly releasing official builds.
There will be an official build process, fully documented and accessible to anyone: it will be easy for any dev to make a standard build.
Raw video will be improved on some modern cams, so it is fully usable.  Planned: 200D and 6D2.
More non-video features will be implemented for new cams.
High demand features from 3rd party builds will be merged into official ML.  Planned: improved raw video LV preview (AKA "Crop Mood") and SD speed tuning.

2025 should be an exciting year for ML; the first year with an official release for any cam beyond Digic 5.

[iaburn]

It looks like Magic Lantern is more alive than ever, second youth 

Better documentation and easier build process is more important than it looks, I'm sure many capable people gave it a try and gave up too soon...

Thanks for the report and all that work!

[kitor]

Let me add to the news:
We re-gained access to (what we thought was) long gone official Magic Lantern Youtube channel - yes, the one with ML v2.3 promo video from 12 years ago! (and also to ML Vimeo account)

It took almost 2 years of digging and trying to contact people who in many cases left this community many years ago. But we got lucky and after I was able to get in touch with @Bart we found out that by sheer luck we was still able to get access there and get me rights needed for the channel management.

[onlykhaz]

   Great News! The 2025 Will be a great year for ML no doubt. Thanks to all involved in keeping the fire alive!

[pmiramon]

Great, this is very, very good news! I thought with Dane's amazing 1620p 4:3 mode we had hit the ceiling. 

[whitelight]

Amazing! Looks like a lot of videographers dreams are going to be turned into reality. Many Thanks to all people involved.

[falcon]

Wow, very nice! Do you have an approximate release date?
Will they have working AF?

[names_are_hard]

The approximate release date is: 2025.  As more work is completed, this may get refined.

I'm not sure what you mean by AF "working".  ML shouldn't change how AF functions for a given cam.  The newer bodies tend to have better, faster AF, and ML won't change that.

[andy kh]

good news

[ol-mark]

Oh, god! Looking forward to having published at least building instructions for 6DM2!

[Ernani]

With a well developed build and its DPAF, the 200D could become the ML holy grail! Will keep a close eye to these developments. So good to see Magic Lantern alive!

[Desander]

<Mod intervention
Full quote removed. Please quote sensible and regard forum/internet etiquette!>

hope to see a build for my trustworthy 200D soon. Thanks for all the hard work.

[names_are_hard]

Thanks for all the words of encouragement!  Progress continues in the background.

Replying to a few different posts:

Build instructions are the same for all cameras and are simple.  Clone repo, install a few system dependencies (e.g. a compiler), be in platform/CAM_OF_YOUR_CHOICE, run: make -j8.  I don't recommend you do this unless you're active on discord and have some sense of the maturity of the build you want to run.  Always test your build env is sane by trying builds in qemu-eos first.

There is an existing published build for 200D, here:
https://www.magiclantern.fm/forum/index.php?topic=26681.0

That is missing features that will be in the 2025 release, but it's easy to install.

DPAF on 200D seems to work fine in video modes.  Cam is, so far, limited to 60MB/s write (it's 40 stock, so ML is giving a 50% boost ), and will not have lossless MLV at release.

[morgan20]

Great news! Hope to see 5d4 getting some love from ML. I might join the dev team if I get to acquire the camera for a not too high price.

[names_are_hard]

5D4 is very unlikely to be in the 2025 release.  While I own this cam and have an ML build for it, it has serious stability issues for which the cause is not known.  Debugging that kind of stuff often takes a long time, so it is not being prioritised for this release.

It's in a good place for someone that enjoys debugging.  It can run ML up to menus, and log to disk or UART.  Just needs time spent determining and fixing problems.

[vastunghia]

Thank you so much names_are_hard for your efforts and for rekindling ML's spirit!

My personal hope, FWIW, is that your tremendous work will make it simpler to achieve anything close to a 'Crop Mood' for 5D3. 5D3 is the most powerful and capable cam that is Raw-enabled at this time IMO, but lacking a reliable live preview (particularly on HDMI out) is a deal breaker.

Looking forward to reading more updates!

[names_are_hard]

Improvements for 5D3 can happen in this repo, and are marginally easier than before.  This repo is faster to build, and works with modern tools.  No difficult installations of obscure compiler versions, etc.  Many bugs affecting all cams have been fixed.

That said, it's not a large factor.  Presumably the work is not being done because nobody is trying.  I don't own a 5D3 so I'm not doing any work for that cam.

You don't need my work to get improved features for 5D3; you need a dev who has a 5D3 and wants to spend time on it.

[vastunghia]

Thanks for clarifying, though it was already pretty clear to me. Just hoping your work could help some dev to be more productive in tweaking the 5D3. Oh, and of course rekindle attention towards the ML program, which could also mean attracting new devs.