Messages

Discussion about the announced changes regarding Tragic Lantern has been moved here:  http://www.magiclantern.fm/forum/index.php?topic=11080.0




Thanks.

Yes, the thread I've linked is pretty much about that, but this process is very difficult without cooperation from 1%. In the past, all of us ML devs have attempted to port many of the TL changes blindly (without a camera to test on) and we have also tried to ask for the help of other developers (with only partial success; the progress was not as fast as we would like, and these ports are not yet as robust as on 5D3 or 5D2, for example).

Fortunately, he did share some of his changes lately (mostly from 7D/6D). But this should be the normal development flow, just like all other code contributions to ML, not "backflow" as 1% calls it.

Nanomad also has an EOS-M and he did some good progress with TL backporting. But, since ML is a spare-time project for all of us, progress is not always as fast as we would like, so we prefer not to waste further development resources on private forks.

If you are not familiar with the how open source contributions should work, you can get an idea here:
http://programmers.stackexchange.com/questions/112458/forking-an-open-source-project-nicely
http://tirania.org/blog/archive/2010/Dec-31.html
http://jamesdixon.wordpress.com/forking-protocol-why-when-and-how-to-fork-an-open-source-project/
http://www.kaizou.org/2013/04/three-golden-rules-open-source/

Alex,

Thank you for explaining these things. Hopefully, 1% will respond and tell us things from his point of view. Maybe it can all be worked out for the best.

Best regards,
Gary

That's why we are trying to undo the community split, so everybody is invited to switch to Magic Lantern (where most of the innovation happens anyway). If there are still differences between ML and TL, these should be backported (but for EOS-M I can help mostly with advice, because I don't use one myself).

If any of you still wants to continue with the community split, you are free to do that on some other website, not here.

Is 1% the only dev who has an M? I remember way back that Max was willing to donate his M camera to ML to use but no one accepted his offer. I think he said he had two M cameras.

Who took ML code, tweaked it a little and refused to give the changes back?

I agree. Why would anyone not share and give back to the parent? Can you take what is current in TL and see what is there and apply it to ML? I am sorry I am not a dev and understand the workings of it all. How will it all end up? I mean, will ML lack what TL has done or will it just take longer for ML to develop what was there? Or, will ML just not try to progress in those areas for certain cameras? What does TL have that ML does not have right now for say, my camera, the M? From what I have seen, on ML I do not see video hacks and RAW recording, just MLV. Probably more, I don't know.

Who started the community split in the first place? Who took ML code, tweaked it a little and refused to give the changes back? Who took over the ML development threads and left the main codebase outdated?

FYI, more than 90% of what you are running is Magic Lantern code, and more than 70% of it was written with my own hands.

Please study the commit history and give credit where credit is due.

I was not aware of a commit history. I am not taking sides and I have respect for everyone. I will edit my post. My concern is to not split the community. We all need each other. it would probably be good to also hear from 1% and to try and work things out. I know you all have tried in the past and it's obvious it will now take some more effort. We need everyone to work together.

I hope you, Alex, and some others realize that the rest of us out here do not know what is going on behind the scenes or the the real history, such as the development for the M and some other cameras.

I hope the major forum changes that were announced do not impede any development for the M, 6D and 7D. Splitting the community is not good. I don't know why such claims are made to scare people away from using TL when I have used the daily builds on my M since last July without having any problems. Nor have I seen anyone reporting any terrible problems from using TL. I am not taking sides and I have no disrespect for anyone here at ML.

[not]

At the moment it is fixed. The ffmpeg argument currently is "-vcodec prores_ks -profile:v 3". The colour format that uses is YUV444 10bit. The quality is quite high, and of course it's using (slow/expensive) AMaZE for demosaicing.

Fast/lower quality proxies is not a feature that anyone has requested to the enhancements list: https://bitbucket.org/baldand/mlrawviewer/issues

thanx. I wanted to use the highest quality prores.

Keep up the good work. this is a great app.

I am using MlRawViewer 1.1.3 . Is there anyway to select which version of ProRes we can use use for a proxy export? Which version is it currently using? 444, HQ, etc.

i can't get ProRes to work with this latest version 1.9. It creates a mov file, but nothing is there when I try playback using VLC or QT.

Thanks g3gg0, 1% and everyone here for taking the time on this topic. I hope it eventually has a solution for all the cameras, including the M.

mlv_rec -> ML code
h264 -> canon binary

-> not possible the way you explained it

please explain why it can't happen. If I explained it wrong, then how should it be explained?

It can be patched the same way as on 600D, again the meta data would probably have to be rewritten. I thought about doing it but figured why bother since its supposed to auto split the files. Its more work to re-write the videos just for 4 frames.

I am talking about not missing any frames when shooting long, non stop events that are splitting the files. He said he thinks it always loses four frames between each 4gb split it records. Can ML/TL record a small 4 frame clip so we then have three clips to rejoin with nothing missing? Or, can ML possibly keep recording into a buffer while it sets up to record the next clip. That buffer then feeds the writing of the second clip, always having a latency of four frames, but not a problem since in post it all gets rejoined. 

Or, maybe we can have an extra small buffer area that only accumulates every small four frames sections and when the whole event ends, the buffers records files to the card. in post, we get an app that reassemble all the clips in proper order into one continuous file. Maybe it has some sort of checking that assures it was all compiled into one with no errors and then it deletes all the splits from the hard disk to save storage space.

I do not have the knowledge of how how TL/ML actually does any of the recording. It all reminds me of the pre read ability we had in Sony VTR's years ago when digital tape recording was introduced. Now, we have other video cameras on the market that record split files and never lose any frames. I think Sony and Panasonic can do it.

Are you suggesting a temporary split file for the 4 frames to be saved to a separate file when the camera creates the 2nd file?  So we would end up with 3 files in the end for a 30 minute clip, or would you suggest having ML join the files at the end of the record session to get one continuous 29:59 file at the end of it?


I am not sure the best way to implement this, but finding out how the > 4 GiB file setting for mlv_rec works may be helpful in determining how easy this is to implement for H.264.

I would say any way that can preserve non stop playback in post after rejoining the files. No dropped frames. It would be great to have ML do it in camera so we end up with one complete joined file that has no dropped frames. But if ML cannot do it all in camera, then an app we can use in post that can join the 3 files (or more, depending on the size of the memory card) into one complete, continuous file.

I was wondering if it was possible for ML to keep recording those four frames or so into a memory buffer and then it records those four frames into another split file when the next clip starts recording. Or it accumulates the four frames each time it starts a new clip. Meaning, the next clip is always four frames later than real time and so on for the whole shoot to one memory card. Maybe if we did this for a long shoot, it might have accumulated at most a 28 frame delay (just guessing, but making a point) which wouldn't even be noticed after it all gets rejoined in post without any dropped frames.

I prefer the correct values being printed in ML.  If I want to know what the Canon rounded values are, I look at the Canon feedback.

That is what I have been doing, using the Canon menu for shutter setting.

Canon actually does this rounding for you, and ML is supposed to print it properly (but somebody needs to sit down and fix it somehow on the M).

I understand, thank you for explaining it for me. I wish I could help more on the dev end as you say someone still needs to fix it for the M. 

Well... one has to double-check FRAME_SHUTTER_TIMER and get_current_shutter_reciprocal_x1000.

There is a big quirk here from Canon (all other cameras are well-behaved): EOS-M does not configure LiveView with the real shutter speed while not recording (so the readings will be only relevant while recording). Same for FPS.

So, for standby, one might want to rely on APEX units. Also, if one can't figure out how to fix the bug, just undefine FRAME_SHUTTER_TIMER (don't leave the wrong value) and the code will fallback to APEX (which is portable, but not very exact - Canon does some unusual rounding and ML doesn't copy it - and the APEX method doesn't account for FPS override, for example, but it's easy to fix).

In photo mode, you can set exposure time of 100 seconds?! are you sure about that?

sorry, I just do not understand it. I never had to do math to set the shutter speed for video. I tried making the shutter adjustments and nothing seemed to be changing except the picture display was getting darker. When I checked in the Canon display, it was up over 200 yet the ML display kept showing 0". Is this method something more that still photographers need? Can this be something that could be a user pref setting so we all can choose what method we need? Can anyone tell me why TL is showing the 1/48 and 1/64 and Canon shows settings like 1/50, 1/60 and so on? Is there something wrong with the TL and Canon way of displaying shutter speed that will ruin our video footage? We need to set the shutter in video to be double the frame rate most of the time.

Normally, the first key press / scrollwheel / whatever should disable the screen right away (at least this is how it works on all other cameras).

For anyone who can reproduce the issue and has basic command-line skills: can you run hg bisect to find the buggy changeset?

The new magiclantern-Nightly.2014Mar17.EOSM202 works now on the M. The splash also goes away with a single push on the set button. I don't understand the shutter display in the ML display. It says 0.-1, -1818  ?? What happened to the usual 1/48 and 1/64 and so on like the Canon menus show and the TL versions? How do we get the shutter to go to the usual settings for video? I had to go into the Canon menus to set it to 1/50 and 1/60 as needed.

I just tried going back to ML Mar 11th and everything worked a few times in a row of off and rebooting. I was able to shoot an MLV raw file, but it has no MLV playback.
The splash screen is annoying. It appears we have to wait ten seconds or more for it to shut off. I tried swiping and nothing got it to shut off. isn't there some sort of swipe or button push to shut the splash screen off?

This was without any module enabled?

(it's important for narrowing down)

yes, no modules were loaded. The camera has been sitting for a few minutes and now it powers up with no modules loaded. I tried three times in a row and all three times it powers up with no modules.

EDIT, after I typed the last message here, I tried again and it's back to not working with the steady red led again. Waited a bit, tried again, it worked. Powered off, waited a bit and the red led hang came back.

What modules do you have loaded?

Can you try booting without them?

I have all the modules loaded, six of them, except arkanoid. I just tried shutting off all modules and rebooted and it worked. I powered off and tried to boot again and it went back to the steady red led and hang... battery pull.

nanomad, i just thought that maybe cus I do not have eoscard used on my card, that might be doing all of this? I don't know, just guessing. But I never had any problems like this on ML builds before a couple of days ago.

I just tried magiclantern-Nightly.2014Mar16.EOSM202 on my EOSM and it works just fine. What firmware version do you have?

Can you try renaming the ML folder and then re-extracting the one from the zip?

I have the latest firmware as far as I know. I am using the fir that always comes with the TL builds.

I just tried a new download of ML Mar 16, deleted the ML and autoexec on the card like I always do, then extracted the zip and copied the two files onto the card. This time, the red led does shutoff like normal when I inserted the card, but when I try to power on, it looks as if it is going to do it with the led flashing then after a couple of seconds it goes to steady red led and nothing happens except it hangs the camera and then a battery pull.

EDIT- I tired booting again and this time it worked. I have never had the steady red led hang up before except for these last two days using the ML nightly. Strange.
BUT, now the intro splash screen is stuck. How do I get past this? In the past, I just swiped the screen again to get to the ML menus.

EDIT 2, now the ML menus showed up after a minute of just sitting there with the splash screen showing. Something is acting strange doing this. Never had this behavior before.

EDIT 3, did my usual settings in ML menus. Turn off audio headphones, turn on all the modules, Debug listing 2 set to Reload modules after a crash. I shut off the camera and then rebooted and now it's back to the steady red led and it hangs there.  :-(

LAST EDIT- after trying to boot again for a few more times, it finally does start to work. Something is very wrong with the builds I have tried (many times)  the last few days.

ML ver Mar 16th and Mar 15th will not work at all on the M. I only get a steady red led showing. Nothing else happens.

Ok one more question. When you do a reinstall, do you also disable any of the bootflags before? Camera or card or both

Lastly, can you do the tests suggested by Alex? They would help tremendously

I can try to do the tests. where do I get the minimal fir and minimal autoexc files? what do you or Alex suggest I do for starting out with my current sandisk 32gb card that has the latest TL and I did not use eoscard on it? I would guess a normal windows 7 format to clear it all? Anything else I need to do to start??