Tragic Lantern for EOS M

Started by coutts, April 17, 2013, 01:43:28 AM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

Rewind

Quote from: maxotics on September 07, 2013, 09:12:24 PM
Was it difficult.  Was it just off by a pixel?  What do you think happened?  THANKS AGAIN!
Well, I don't know what exactly happened, because i don't have a EOS-M and i don't know, how it was organized earlier. I just picked up your raw file, and reconstruct the pixel map from it, here it is. Then I just corrected the dot data file to correspond the locations from this chart.
It was not so difficult due to i've done it already for my 650D, but to read someone else's code is a real pain you know ))

maxotics

Quote from: Rewind on September 07, 2013, 09:25:06 PM
Well, I don't know what exactly happened, because i don't have a EOS-M and i don't know, how it was organized earlier. I just picked up your raw file, and reconstruct the pixel map from it, here it is. Then I just corrected the dot data file to correspond the locations from this chart.
It was not so difficult due to i've done it already for my 650D, but to read someone else's code is a real pain you know ))

Thanks Rewind!  Awesome work.  How are all those dots represented in a few lines of numbers in the dotData file?  I find discussions about it in the thread, but no definite reference of what he decided to do.

Rewind

Quote from: maxotics on September 07, 2013, 09:28:54 PM
Thanks Rewind!  Awesome work.  How are all those dots represented in a few lines of numbers in the dotData file?  I find discussions about it in the thread, but no definite reference of what he decided to do.

I doubt my poor english skills will allow me to explain all this stuff correctly, but there are repeatitive sequences in dot locations, which may be characterized by a small amount of data. The comments at this part of the source code were guided me in the right direction:
  * Add a "grid block" of dots to the dot set
     *
     * @param x0 the offset of the first dot column, relative to the image center in RAW coordinates (ignore ActiveArea)
     * @param dy0 the offset of the first dot row, relative to image center in RAW coordinates (ignore ActiveArea)
     * @param dy1 the offset of the last dot row, relative to image center in RAW coordinates (ignore ActiveArea)
     * @param stepX the x-offset between to grid dots
     * @param stepY the y-offset between to grid dots
     */
    public void addCoordinates(int x0, int dy0, int dy1, int stepX, int stepY)
    {
        gridInfo.add(new int[]{x0, dy0, dy1, stepX, stepY});
    }


These variables x0, dy0, dy1, stepX, stepY are exactly what we see in dot data file.
Hope this will be helpful )

maxotics

VERY HELPFUL!  You're the man REWIND!  Now I can shoot tonight and know I'll be pink-free in post :)

maxotics

HI Rewind, sorry to be a pain.  Can we just use one line (dot patterns) as an example? 

7, -359, 331, 24, 30

I understand the 24 and 30.  I can see how grids of dots have those x and y offsets.

But I can't find the first reference dot using the 7 and -359.  How would you calculate that position on the grid, if we talk about  in terms of 0,0 to 1280,720? 

I think I'm close!  THANKS!!!

Rewind

Quote from: maxotics on September 07, 2013, 10:27:52 PM
How would you calculate that position on the grid, if we talk about  in terms of 0,0 to 1280,720? 
These positions are related to the center of the frame. I used Photoshop to take them:


As you can see by the rulers, the center is on the cross of guides, then you just counting positions.

maxotics

Thanks Rewind, I think I'm almost there.

7 gets the horizontal "column" position from center, so (1280/2) + 7 = 642 or x = 642
-359 get the vertical "row" position from center, so (720/2) - 359 =  1 or y = 1 (but starting from top down?)
720/2 + 331 = 691 which corresponds to the last "row" point, or in x/y 642/691 ?

What seems implicit, and please let me know if this sounds right, is that they assume the dots to the farthest right or left are unknown.  So these patterns would work for both 1280x720 and 1600x720 say.

PDR essentially works from the center out, to the top-most dot and bottom most, and then to the end of each side, whatever the width

Is that right?  Thanks!

WiGgLr

Is anyone who owns the 18-55mm lense NOT experiencing the shutter bug?

zuzukasuma

@rewind, thanks for the update. I'd like to have a destination folder option on the software you've created, is it possible? I know its not killing the drive speed like cr2hdr.exe but I'm keeping 2 drives for RAW footage and pre-edit footage.
in a complicated relationship with eos m.

Malakai

Quote from: WiGgLr on September 07, 2013, 11:52:25 PM
Is anyone who owns the 18-55mm lense NOT experiencing the shutter bug?

I think its all 18-55mm lenses that are affected. Its still an unknown bug too. Between a couple of us on here have posted a few DM logs to see if 1% or another dev can nail the issue but its drawing a blank so far.
Hunting for that elusive EOS M shutterbug!!


AnotherDave

Just got/hack ML on the EOS-M (also have it on 5D3 for a while) and I have a quick question - Does Focus Stacking work on here?  I don't see it in the menu, but maybe I missed it? :-/


haomuch

Seems every body is still talking about video staff. Makes me doubt if I really bought a photo camera. Anyway, I just wonder if 1% are still interested in working on the magic zoom function or improving the live view auto zoom function. Thanks again.

jerrykil

Quote from: Malakai on September 08, 2013, 03:12:49 AM
I think its all 18-55mm lenses that are affected. Its still an unknown bug too. Between a couple of us on here have posted a few DM logs to see if 1% or another dev can nail the issue but its drawing a blank so far.
how do i reproduce this bug? has this been posted, i searched quickly and i couldn't find any exact directions. does a power cycle fix it for you?

zuzukasuma

Quote from: haomuch on September 08, 2013, 06:08:59 AM
Seems every body is still talking about video staff. Makes me doubt if I really bought a photo camera. Anyway, I just wonder if 1% are still interested in working on the magic zoom function or improving the live view auto zoom function. Thanks again.

well, it definitely not ready for pro-use. there are still lots of trouble in RAW footage, especially dual-iso footage requires too much cpu cycles to make something out of it. also there are other problems, dots, green/pink screen issues, fps override crashes, overheating, small battery etc.

get a 5D.III and be happy, but costs 10 times more than M.
in a complicated relationship with eos m.

Rewind

Quote from: zuzukasuma on September 08, 2013, 02:26:29 AM
@rewind, thanks for the update. I'd like to have a destination folder option on the software you've created, is it possible? I know its not killing the drive speed like cr2hdr.exe but I'm keeping 2 drives for RAW footage and pre-edit footage.
All the credits goes to Foorgol and Mixer2, i am just playing around with interpolation algorithm (which by the way is pretty much acceptable now) and trying to optimize the tool for my 650D, so that i can work with it right now.
But this is just a partial solution, I hope the devs sooner or later will polish the tool. I'm not a java programmer, and it is formidable for me to explore their code.
But I'm glad to help )

WiGgLr

Quote from: jerrykil on September 08, 2013, 06:59:11 AM
how do i reproduce this bug? has this been posted, i searched quickly and i couldn't find any exact directions. does a power cycle fix it for you?

From what I've read, you just switch the camera on with the 18-55mm attached and you won't be able to take photos until you either double tap the power button, or reattach the lens with the power on. The bug will reappear every time the camera is turned off.


Malakai

Quote from: gary2013 on September 08, 2013, 03:20:44 AM
The only thing that is so different about this one lens, and not even the pancake brother, is that it uses IS controlled electronically from the cam body. The pancake has no IS and yet it works fine with ML. Just this one zoom lens has a problem.

Gary

Does this happen for the 11-22mm ef-m? That has Electronic IS i believe.

Quote from: jerrykil on September 08, 2013, 06:59:11 AM
how do i reproduce this bug? has this been posted, i searched quickly and i couldn't find any exact directions. does a power cycle fix it for you?

Just turn the camera on with ML bootloading and the shutter wont activate, power cycle or rotate lens and it will work until you next power off. Is not a permanent fix.
Hunting for that elusive EOS M shutterbug!!

zuzukasuma

Quote from: Rewind on September 08, 2013, 09:15:15 AM
All the credits goes to Foorgol and Mixer2, i am just playing around with interpolation algorithm (which by the way is pretty much acceptable now) and trying to optimize the tool for my 650D, so that i can work with it right now.
But this is just a partial solution, I hope the devs sooner or later will polish the tool. I'm not a java programmer, and it is formidable for me to explore their code.
But I'm glad to help )

I see. no prob, they are here too :) I hope they can combine it with other softwares too, I'd like to have a "Magic Lantern RAW-Dual ISO-HDR Video Suite" in the future.

posted on dual iso thread but you might miss it,

having problem with cr2hdr.exe on bright areas like sunset, the original DNG is here (from RAW-Dual ISO file) https://www.transferbigfiles.com/c58348c4-8282-4dcd-b58b-59e1dbb9a684/Uw7m4GTs2sg8-yDtjvYYjA2
in a complicated relationship with eos m.

jerrykil

Quote from: Malakai on September 08, 2013, 11:52:38 AM
Does this happen for the 11-22mm ef-m? That has Electronic IS i believe.

Just turn the camera on with ML bootloading and the shutter wont activate, power cycle or rotate lens and it will work until you next power off. Is not a permanent fix.

mine works fine so far. i;ve had the bug on the old firmware 1+ML but it was spontaneous. right now i have my 18-55 in and no problems as of yet

Malakai

Quote from: jerrykil on September 08, 2013, 12:53:22 PM
mine works fine so far. i;ve had the bug on the old firmware 1+ML but it was spontaneous. right now i have my 18-55 in and no problems as of yet

Are you using 1%'s latest ML version with the EF-M 18-55mm? What firmware are your lens and camera on? This could be the breakthru we need to find the fix to this bug! If yours is working then there may be something in the logs that one of the devs can see the difference between us having the bug and you not having it.

Can you upload the contents of your SD card, just the ML folder, autoexec and the fw file. Zip them all together and ill grab them and test them in my M to see if theres any difference.
Hunting for that elusive EOS M shutterbug!!

jerrykil

I'm running from tragic lantern 6D binaries i compiled a few commits ago. maybe something they did fixed it recently? let me know if these files work:

http://www.bot-fly.com/tragiclatern_eosm202_sept4.zip

Here is another one i built today. still no shutter bug for me:
http://www.bot-fly.com/magiclantern-v2.3.NEXT.2013Sep08.EOSM202.zip

Malakai

Quote from: jerrykil on September 08, 2013, 04:40:16 PM
I'm running from tragic lantern 6D binaries i compiled a few commits ago. maybe something they did fixed it recently? let me know if these files work:

http://www.bot-fly.com/tragiclatern_eosm202_sept4.zip

Here is another one i built today. still no shutter bug for me:
http://www.bot-fly.com/magiclantern-v2.3.NEXT.2013Sep08.EOSM202.zip

Neither versions worked :( can you confirm your M FW and lens FW, mine are 2.0.2 and 2.0.0. @gary2013 can you give these a go too see if they do anything for you.
Hunting for that elusive EOS M shutterbug!!

jerrykil