Magic Lantern by Tramm Hudson getting close

[Chris Barcellos]

"Wizard" Tramm Hudson posted update on status tonight. He's working on "err70 problems that I've read about were mirror problems or CF issues. This is definitely software..."

He goes on to say:

"It's my major project for the weekend. This is the last major hurdle before a public release of the Magic Lantern firmware since the other big-ticket items have been knocked out:
* AGC disabled
* Selectable gains and filters
* Onscreen meters
* Zebra stripes
* Button remapping
* Audio monitoring on A/V jack
* Framework for new extensions"

What an amazing job--.

A lot of stuff covered already... wow do we owe you a lot man...

[Sean Seah]

Sounds really AWESOME!!!

[Tramm Hudson]

Originally Posted by Chris Barcellos

* AGC disabled
* Selectable gains and filters
* Onscreen meters
* Zebra stripes
* Button remapping
* Audio monitoring on A/V jack
* Framework for new extensions"

And last night I stumbled on the left/right audio levels in the DIGIC IV rather than the mono one, so it now has dual audio meters. I'm still not happy with the ballistics and the sampling time for the meters, but we'll get to that eventually.

I might have also found a work-around for the err70 problem. Writing 32-bit word at a time and writing monotonic addresses in the lvram rather than randomly jumping around seems to prevent it from showing up. I have re-written the meter drawing code to do a scanline at a time, and the printf() code to write the font by entire words rather than pixels. Since making that change I have not had any err70 aborts.

The other major fix that needs to happen is to detect when the camera is displaying menus. Right now it overwrites the entire screen with the zebras, which prevents Canon's menus from being displayed.

[Tom Daigon]

Your commitment and motivation to making this fine camera even better is
a treat to behold. As a token of the masses of us that are indebted to you for
your abilities....is there any way we can return the favor? Maybe voluntary
monetary contributions to you or your favorite charity or...?
This expenditure of creative positive energy on your part deserves to be acknowledged
and rewarded in some manner.

[Allan Tabilas]

I agree. Tramm, your commitment to the firmware is incredible. Let us know if you are taking donations, etc.

[Chris Hurd]

I'll ask Tramm if he can set up a PayPal donation page that references DV Info Net. Thanks for the suggestion,

[Glen Elliott]

Nice! I wonder if Tramm single handedly will put a fire under Canon once again for yet another firmware update. :) Man, give us an inch.....

[Jon Fairhurst]

Originally Posted by Tramm Hudson

I'm still not happy with the ballistics and the sampling time for the meters, but we'll get to that eventually...

Yeah, the sampling time needs to be much longer. Ideally, you'd be able to fade the edges of the sampling window, but that's probably not possible. With my 500 Hz tone calibration tests, the meters are highly inconsistent, I assume because it catches the tone at various phases.

[Tramm Hudson]

Originally Posted by Chris Hurd

I'll ask Tramm if he can set up a PayPal donation page that references DV Info Net. Thanks for the suggestion,

Thanks to everyone who has offered donations. The software will be free when it is released next week and if you want to donate to help with the development, I've setup a dvinfo PayPal donation link: https://www.paypal.com/cgi-bin/websc...ton_id=6273230

I'm also looking for used film-making equipment. Send me an email if you have an older Steadicam, rails systems, matte box or lenses that you aren't using anymore.

The full source will also be released under the GPL so another way that some folks will be able to help is by writing new software to run on everyone's cameras. While it isn't directly based on CHDK, I've benefited from their GPLed sources and the hard work that they did in reverse engineering Canon's P&S cameras. Without their initial effort, I would not have been able to make such rapid progress on the 5D Mark II.

[Chris Hurd]

Thanks Tramm! I've copied your post and made it a "sticky" at the top of the forum:

http://www.dvinfo.net/conf/canon-eos...tion-link.html

[Tom Daigon]

And as one of the iniital contributors to the cause I just want to say thanks!

[Marc Fairorth]

Tramm,

Thank you for your efforts. Where will you be posting the code? Can you say a few words about the development environment?

Regards,

Marc Fairorth
Austin, TX

[Mark Nicholson]

I know there are at least a few of us out there checking various forums everyday waiting to hear about the firmware getting released. WE do really appreciate it, and I hope when that happens at least a few people out there will give you an IMDB credit for the work you've done!

Now the moment 24p is announced... the internet will go crazy.

[Tramm Hudson]

Originally Posted by Marc Fairorth

Where will you be posting the code?

There will be a public wiki for documentation, development and source distribution. It's not ready to go live yet, but I'm hoping after Jon's tests this weekend the firmware will be ready for wider distribution. I'll announce it here as soon as it looks stable enough for beta users to start playing with it.

Can you say a few words about the development environment?

It is very similar to CHDK's requirements: an arm-elf-gcc cross compile setup and some perl scripts. I do the hacking on my Linux desktop box and my OS X laptop ("sudo port install arm-elf-gcc" was all it required).

There are only a few lines of assembly in the entire code tree. I've made an effort to jump into C code as quickly as possible to lower the barrier to entry for other coders. I have also avoided copying any Canon code into my firmware so that it can be distributed without any encumberances. I have, however, written headers based on my reverse engineered firmware so that it is possible to create DryOS tasks, interact with some of the system structures and devices, and call into functions in the ROM image.

The Magic Lantern firmware is structured as a set of loosely coupled .c files. They add entry points to a ELF segment to indicate which tasks to startup at boot time once the grungy bits are done, simply by virtue of being linked into the image. This allows new functions to be added just by adding the .o file to the list of object files in the Makefile.

[Tramm Hudson]

Originally Posted by Tramm Hudson

It's not ready to go live yet, but I'm hoping after Jon's tests this weekend the firmware will be ready for wider distribution. I'll announce it here as soon as it looks stable enough for beta users to start playing with it.

Jon's tests are going great, but one of the alpha users reported a fairly major bug: the digital zoom in live view gets "stuck" and it doesn't return to full screen mode. I'm working on that one now, which is the last hurdle before public availability.

I promise!