Messages

There's no point in going above ISO 1600 on DIGIC IV.

http://www.guillermoluijk.com/article/isos5dmkii/index.htm

TP log is quite dangerous. It hooks directly in the property task and pretty much all other Canon tasks that send messages.

DM log will flood the CPU with processing all those debug messages (of course, if enabled). In LiveView it can't run without disabling a bunch of them, at least here.

They were written for developers to understand how Canon firmware works; they are not designed for the end-users (not even as diagnostic tool).

I've added a small buffer optimization to the raw rec simulator:

https://bitbucket.org/hudson/magic-lantern/commits/0db037b3b960

With the numbers from the example above:

1152x460 23.976fps 20.0MB/s => with current ML method, the simulation says 198 frames.
After optimization (only reordering and splitting the buffers differently) => 880 frames;

At 21.0 MB/s => 729 frames with current method, 2880 optimized.

Notice there's a slight difference from the previous numbers, because the latest script also models the slight speed loss from using small buffers.

Can you implement this buffering strategy in raw_rec and check if my theory is valid?

DM/TP? these are hacker modules, not for end-users...

That's old: http://magiclantern.wikia.com/wiki/Unified/UserGuide

new one: http://wiki.magiclantern.fm/userguide (but I usually have to hit F5 a few times for it to load)

Set them as 0 in the raw data, this means "bad pixel".

Did you even read the docs?

http://wiki.magiclantern.fm/userguide#movie_record_50d

Well.. small formatting quirks for the new wiki (which seems down anyway). Copy/paste the output and you'll see it needs some manual adjustments.

On the old wiki I didn't have to do any such adjustment, just ran the script and pam! the user guide wiki page was updated.

I've used a plain text editor. You need latex to compile it. Of course, you can add files, sections, whatever.

All the scripts are in the doc directory (mkdoc.py is for pdf, mkdoc-cam for help bmp's, wikidoc for wiki output - but needs some tweaks).

Not sure if it's enabled in 650D, but on 5D3 it's working.

Well, it is actually in the ISO submenu.

Highlight rolloff improvements are already there, in the image finetuning menu (only for movie). They work nicely with HTP.

Here's how the document is split:

Code Select Expand


include(o, "FEATURES.txt");
include(o, "MANUAL.txt", 1);
include(o, "MENUSTART.txt");
include(o, "MN-AUDIO.txt");
include(o, "MN-EXPO.txt");
include(o, "MN-OVERLAY.txt");
include(o, "MN-MOVIE.txt");
include(o, "MN-SHOOT.txt");
include(o, "MN-FOCUS.txt");
include(o, "MN-DISPLAY.txt");
include(o, "MN-PREFS.txt");
include(o, "MN-DEBUG.txt");
include(o, "MENUEND.txt");
include_indent(o, "FAQ.txt");


So, each menu has a text file. There are also text files for other large sections (e.g. FAQ). The master document is created by the scripts.

The syntax is RST ( http://docutils.sourceforge.net/docs/user/rst/quickstart.html ). You can just write it as plain text, since the formatting is easy to fix afterwards.

To add images, commit them like normal files. To include them in doc, take the syntax from examples:

Code Select Expand


.. figure:: 02.expo.png
    :align: center
    :width: 5cm


Of course, you can use a visual editor for editing, then commit them to the repo and the changes will appear in both the PDF and the BMP versions. If anyone fixes the wiki scripts, we can update the online version as well, all from a single source.

It probably doesn't detect the HTP setting; ask the guys to fix it.

Yes, if you look at the latest changes, these PDFs and BMPs will be included in the nightly zip.

if the road man wants to print a manual, must go to repository, find the right txt among a myriad of other files, and print it ? or, worse, compile the txts together ? How ?

make docs

This will make PDF and BMP files, like the ones in ML 2.3 zip.

If you don't like that feature, let me know and I'll just delete it. Two years of work gone out of the window.

So you have noticed a consistent slowdown of 0.1 MB/s? I don't think this is enough to jump to the conclusion that the latest build is any slower.

If you reuse the existing manual (from the repositiory), it can be compiled in many output formats (PDF, wiki, HTML, BMP for in-camera help, or whatever you may want to add.

Also I think translations should be done a bit later.

Why not edit it directly here? https://bitbucket.org/hudson/magic-lantern/src/tip/doc/

See my earlier posts. In x5 mode, you can't get good image without pink preview (well, grayscale now).

If you have correct color preview while recording in x5, expect colored dots that you can't remove, because their position is not fixed (they follow high-contrast areas). This happens only on 5D2 and 50D afaik.

If you run the speed simulation script, you would notice that at 1152x460 you get 603 frames if the card write speed is 20.9 MB/s and 891 if the speed is 21.0. For reproducing the results: buffers 32x2+8, 23.976 fps, timing variation 0.

So these reports about slowdown are quite pointless.

2048x1080 is technically possible with a fast card but the devs just haven't enabled it in the test builds

What?!

@squig: this is the 10th reminder to update your tests. In raw, the intermediate ISOs are identical to multiples of 100.

Follow focus. It's in the raw nightly builds.

On raw2dng output it works.

With the hex editor, figure out the offset of black_level, or just append a video file that has the correct level, and a little more frames.

gamentine: this was fixed a few days ago: the display turns on, takes the picture, then turns off right away. Image review should be off.

Can you upload a video?