magiclantern for T2i/550D 1.0.9: very soon

[indy arm]

this is my system log:

Wed Dec  1 23:01:08 2010
     0:    11.311 [STARTUP]
K270 ICU Firmware Version 1.0.9 ( 4.3.6 )
     1:    11.367 [STARTUP]
ICU Release DateTime 2010.09.02 12:51:21
     2:    11.616 [SEQ] CreateSequencer (Startup, Num = 6)
     3:    11.846 [SEQ] NotifyComplete (Cur = 0, 0x10000, Flag = 0x10000)
     4:    12.877 [MAGIC] Magic Lantern 0.1.9 (5dbd73e0f8f1+ (550d) tip)
     5:    12.897 [MAGIC] Built on 2010-12-01 20:47:45 by user@ubuntu1004desktop


Indy

[indy arm]

Hi,

Please find ML for Rebel T2i/550D ported to firmware 1.0.9.

"magiclantern.fir" file size as been optimized, so that CO2 footprint is reduced, as well as decryption/signature check time ;-)

I added a configuration variable to disable bootdisk flag, see mlantern.cfg.

Already tested 10-15 times, so risk is very low, but never zero...as usual.
http://bitbucket.org/hudson/magic-lantern/downloads/magiclantern-0.1.9-rc0_550d_fw109.zip

Indy

2010/12/1 indy arm <arm.i...@gmail.com>

[phil]

Hey, very nice. I didn't know there's a new firmware version.
What did you have to do to port it?

phil

On Dec 3, 9:12 am, indy arm <arm.indi...@gmail.com> wrote:
> Hi,
>
> Please find ML for Rebel T2i/550D ported to firmware 1.0.9.
>
> "magiclantern.fir" file size as been optimized, so that CO2 footprint is
> reduced, as well as decryption/signature check time ;-)
>
> I added a configuration variable to disable bootdisk flag, see mlantern.cfg.
>

> Already tested 10-15 times, so risk is very low, but never zero...as usual.http://bitbucket.org/hudson/magic-lantern/downloads/magiclantern-0.1....
>
> Indy
>
> 2010/12/1 indy arm <arm.indi...@gmail.com>

[Renny Hayes]

Does this need to be installed OVER 1.0.6? I just get a "LOADING"
screen and then it shuts off when trying to install from scratch.

[arm.indy]

Renny,
It is for T2i/550D, firmware 1.0.9
Previous Magic Lantern for T2i/550d was for fw 1.0.8.

Phil,
I had to find the DryOS functions addresses that are used by ML, this
is the file stubs-550d.109.S.
plus some checks, understand the building process and update my
signing tool.

Indy

[Emma m]

What are the features in this release? uprated h264 bitrate? audio
level changing? audio out?

> --
> http://magiclantern.wikia.com/
>
> To post to this group, send email to ml-d...@googlegroups.com
> To unsubscribe from this group, send email to ml-devel+u...@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/ml-devel?hl=en

[arm.indy]

This is only a port to firmware 1.0.9.

As with the previous version for fw 1.0.8 you can tune audio levels
via the configuration file mlantern.cfg.
I just added a variable "magic.disable_bootflagf" which is self
explaining.

please read the wiki.

Is there any interest for bitrate tuning ? In any case, it might only
be available through config file, as menus are not implemented yet.

Indy

On 4 déc, 17:27, Emma m <models...@gmail.com> wrote:
> What are the features in this release? uprated h264 bitrate? audio
> level changing? audio out?
>

[Renny Hayes]

I get it now. I didn't realize 1.0.9 had been released.

[arm.indy]

I was impatient for this update:
- to check any security changes = nothing
- how difficult it is to port ML to another fw version = mission
completed this time. fw are almost identical.

Indy

[Nicolas Marechal]

You need to update to the Canon Firmware 1.09 first then you can apply ML 1.09. Works for me !

[Piers]

I'm gonna bite: what does the magic.disable_bootflagf do? Does it not
run ML that time, not run ML forever? I have to admit I've never been
clear on bootflag versus autoboot (or whatever, I'm not even clear on
what everything's called)

PG

[John Calvin Young]

Definitely interested in higher bitrates if it can be implemented as
an option (even with only being via configuration files). I haven't
used ML on my t2i yet, but am definitely planning to use it on my next
short film, which is shooting this month...

John Calvin Young

[arm.indy]

This version has been downloaded 274 times, and I had (almost)
feedback.
is it stable ?

I'm considering adding h264 tuning features from Trammel/5d2 branch,
please be patient.

Indy

[robert]

Here: http://magiclantern.wikia.com/wiki/550D I see these versions of
ML alpha for 1.0.8, they are in the different threads of this forum:
* 8 Aug 2010 Update 550D beta, now with gain control
* 11 Nov 2010 magiclantern for 550D, compiled with shorter config
file
* 23 Nov 2010 Recording Internal and external audio simultaneously
* 7 Dec 2010 fixed cropmarks
Which of these versions listed has been ported to this 1.0.9 version?

[Alex]

As far as I know, it's this one:

* 11 Nov 2010 magiclantern for 550D, compiled with shorter config file

The newer ones will be ported after you (ML users) will test them on 1.0.8 and confirm they work well and they are stable.

[Alex]

Porting of the new features on 550D/1.0.9 is almost done, we are
preparing for a new official release. Stay tuned!

On Dec 8, 7:04 pm, Alex <broscutama...@gmail.com> wrote:
> As far as I know, it's this one:
> * 11 Nov 2010 magiclantern for 550D, compiled with shorter config file
>
> The newer ones will be ported after you (ML users) will test them on 1.0.8
> and confirm they work well and they are stable.
>

> On Wed, Dec 8, 2010 at 6:52 PM, robert <nagyr...@gmail.com> wrote:
> > Here:http://magiclantern.wikia.com/wiki/550DI see these versions of

> > ml-devel+u...@googlegroups.com<ml-devel%2Bunsu...@googlegroups.com>

> > > > > For more options, visit this group athttp://
> > groups.google.com/group/ml-devel?hl=en
>
> > --
> >http://magiclantern.wikia.com/
>
> > To post to this group, send email to ml-d...@googlegroups.com
> > To unsubscribe from this group, send email to

> > ml-devel+u...@googlegroups.com<ml-devel%2Bunsu...@googlegroups.com>

[Pedro Barbosa]

the feature "Time remaining estimate indicator" is already ported from
5d or it will be in the future? sometimes appears a time indicator in
top right corner or the lcd but it disappears as soon as i go to the
ML menu

[Alex]

What does this feature do on the 5D? Show how much time you have left
for recording? I can't see anything in the top right corner; what
steps did you follow to make it show?

I can't find anything meaningful in the code about this feature...
maybe I don't know where to look. I've found timecode and timed-start,
but I guess you are not talking about those.

> --
> http://magiclantern.wikia.com/
>
> To post to this group, send email to ml-d...@googlegroups.com
> To unsubscribe from this group, send email to ml-devel+u...@googlegroups.com

[Trammell Hudson]

On Mon, Dec 13, 2010 at 01:20:31AM +0200, Alex wrote:
> What does this feature do on the 5D? Show how much time you have left
> for recording?

It is an estimate of the time remaining based on the space on the
card, but since it doesn't take into account any of the bit rate
changes since Canon's 1.1.0 firmware, I doubt that the value is very
realistic any more.

The value should probably be updated to read in MB instead of
min:sec.

--
Trammell

[Alex]

Yes, MB is very useful, too.

In which source file is this feature implemented in 5D2? I can't seem
to find it...

[Pedro Barbosa]

when i turn on the camera...in the top right corner shows a red time-
counter...

about the bitrate...it seams that it is different from the 5d to the
550d...but what matters is that it works and increases the
bitrate...but it was nice if we could understand work 550d works

[Alex]

Okay, found it!

PROP_HANDLER( PROP_REC_TIME )
{
...
}

A first hypothesis is that PROP_REC_TIME=0x02010009 may have a
different address in 550D. Trammell said in the comments:
// time remaining. output at 1 Hz while recording

Enabling Debug->Draw prop and disabling Video->global draw helps
seeing property changes in real-time. I've saw the following two
properties which are counting down: 0x8004000d and 0x80030005. They
have identical values, and at the end of the movie they are 0. Not
sure about the units (seconds?)

[Trammell Hudson]

On Mon, Dec 13, 2010 at 01:30:26AM +0200, Alex wrote:
> Yes, MB is very useful, too.
>
> In which source file is this feature implemented in 5D2? I can't seem
> to find it...

PROP_HANDLER( PROP_REC_TIME ) in zebra.c. That file has turned
into the dumping ground for most of the onscreen stuff (sort of like
debug.c is the storehold of random testing code that doesn't work
yet).

--
Trammell

[Alex]

Seems to work!

Will be in the next version. With QScale control enabled, the value is
not accurate any more.

Also added an option to turn off QScale, which works :)

[Alex]

As a preview for the new 550D/1.0.9 release, here is the first draft
of the user guide, which will be included in the zip file in plain
text format.

http://magiclantern.wikia.com/wiki/550D/UserGuide

I kindly ask everyone who knows how to use Magic Lantern to check and
improve this guide.

The sources for the wiki page are in RST format (and I think they are
human-readable enough to be included in the zip file without
modification):

http://bitbucket.org/hudson/magic-lantern/src/7e914d4e1035/FEATURES.txt
http://bitbucket.org/hudson/magic-lantern/src/7e914d4e1035/CONFIG.txt

For rendering the RSTs to Wiki markup, I've used Pandoc:
http://johnmacfarlane.net/pandoc/try . You can improve this guide
either by pushing to the repo (for the developers) or by editing the
wiki (and I'll commit the changes to the repo).

The features described in the guide are the ones from the latest 1.0.8
experimental build, with minor changes here and there.

[Alex]

P.S. If you want to try the pre-release build, mail me.

[Piers]

As suggested off line, what about a bytes remaining and/or raw canon number displayed so that user has a fair idea of time left without us having to do possibly (very) bogus time estimations?

[Alex]

As suggested off line, I don't know how to read the bytes remaining.
The value displayed is the raw canon number.

[Alex]

It would be nice to have some good screenshots in the user guide, so
if you have a (good) second camera, please make them and send them
here.

[Alex]

I've tried to merge Indy's ideas with the install guide from the main
page (which is much more complete imo). Also formatted the guide as
RST, to be included in plain text in the release ZIP.

Comments & improvement are welcome. It's still not clear for me how to
make cards bootable on Mac. Should we leave the old script, or does
Indy's new script work on Mac, too?

On Tue, Dec 14, 2010 at 11:35 PM, arm.indy <arm.i...@gmail.com> wrote:
> I have started an "installation instructions" page for 550D:
> http://magiclantern.wikia.com/wiki/550d_install
>
> Feel free to improve it
>
> Indy
>
> On 14 déc, 22:32, Matt Dennie <matts...@gmail.com> wrote:
>> Don't just open bootCF.exe. Open a command prompt first, change directory to
>> where bootCF is, and the execute it from there.
>>
>> ex.  Start=>Run=>cmd  type   cd C:\"location of bootCF.exe" (press
>> enter)
>>
>> then type bootCF.exe and follow the regualr instructions.
>>
>> If the change directory thing messes you up, put bootCF where the command
>> prompt opens by default: C:\Users\"User Name" for windows 7, C:\Documents
>> and Settings\"User Name" for XP (I think)
>>
>> On Tue, Dec 14, 2010 at 1:16 PM, cristian paradiso <guit...@yahoo.it> wrote:
>> > Hi there, I have a 32gb sd card that I want to make it bootable in
>> > order to make my ML on it but if I do it with the windows version of
>> > bootcf it wont work to me when I run it it opens a terminal window
>> > that disappear immediately and won't let me do anything, I have a mac
>> > also but I dont know where to get the MAC version of it and the post
>> > below doesnt really explain how to do it. Any help please? Can someone
>> > help me on how to do it? I uploaded the firmware to 1.0.9 from 1.0.8
>> > and now I want to ML it, but I need first to make my 32gb sd bootable
>>
>> > "For Mac users, the terminal commands listed in the original 550d/T2i
>> > hack announcement should work regardless of card size"

>>
>> > --
>> >http://magiclantern.wikia.com/
>>
>> > To post to this group, send email to ml-d...@googlegroups.com
>> > To unsubscribe from this group, send email to

>> > ml-devel+u...@googlegroups.com<ml-devel%2Bunsu...@googlegroups.com>

[Alex]

Seems I've hit a bug in google groups.

Original message was:

> I have started an "installation instructions" page for 550D:
> http://magiclantern.wikia.com/wiki/550d_install
>
> Feel free to improve it
>
> Indy

On Dec 15, 12:32 am, Alex <broscutama...@gmail.com> wrote:
> I've tried to merge Indy's ideas with the install guide from the main
> page (which is much more complete imo). Also formatted the guide as
> RST, to be included in plain text in the release ZIP.
>
> Comments & improvement are welcome. It's still not clear for me how to
> make cards bootable on Mac. Should we leave the old script, or does
> Indy's new script work on Mac, too?
>

[Alex]

Also added Renny's suggestion about bootcf. Thanks!

[Alex]

Here's the source code for the wiki page:
http://bitbucket.org/hudson/magic-lantern/src/tip/INSTALL.txt

Is it human readable enough for including it as-is in the release (as
plain text), or it's better to compile it into a PDF like the attached
one?

Same question for FEATURES.txt and CONFIG.txt.

If you like the PDF, there are some small problems to be solved with
the wiki links and inline images in the user guide. If the text it's
fine... just make zip and we have the new release :)

[Piers]

Indy's script would work on Mac if you set the shell var to something Mac-like like /dev/disk1s1*. Except he then goes and explicitly refers to "/dev/sdc1" instead of his var ;-)

should work, will try tonight on my non-booting SDHC card (which is actually a microSD, but I don't think that would matter).

PG

(* the contradiction that is OSX, theres nothing Mac-like about "/dev/disk1s1"!!)

[Piers]

I noticed last night while changing my still record mode from L-Hi down to S-hi that the time remaining estimate went UP, so I speculate that PROP_REC_TIME is in fact the remaining stills estimate? Or at least related to it.

PG 

[Alex]

There are two possible addresses for PROP_REC_TIME (I wrote them in a
thread somewhere here). In video mode, they seem to have the same
value. Maybe in video mode they are the remaining time for video, and
in photo mode they are the remaining number of photos (or only one of
them)?

The latest sources include a better draw_prop mode (i.e. easier to
read).

[Alex]

Can you fix the script, so we'll be able to say "for Linux and Mac" in
the docs?

On Dec 15, 4:56 am, Piers <pie...@gmail.com> wrote:
> Indy's script would work on Mac if you set the shell var to something
> Mac-like like /dev/disk1s1*. Except he then goes and explicitly refers to "
> /dev/sdc1" instead of his var ;-)
>
> should work, will try tonight on my non-booting SDHC card (which is actually
> a microSD, but I don't think that would matter).
>
> PG
>

> (* the contradiction that is OSX, theres *nothing* Mac-like about
> "/dev/disk1s1"!!)

[Alex]

After playing a bit with the latex template, here are the new PDF
docs. Doc sources & build script are commited.

[Alex]

Here is the latest pre-release build. Please test it, stress it a bit,
and confirm if everything is OK (or report the problems).

[Alex]

We have a new Win32 GUI tool for making cards bootable from Pel! It
seems good enough to replace CardTricks & bootCF.

Check the docs: http://magiclantern.wikia.com/wiki/550d_install

On Dec 17, 10:47 am, Alex <broscutama...@gmail.com> wrote:
> Here is the latest pre-release build. Please test it, stress it a bit,
> and confirm if everything is OK (or report the problems).
>

> On Wed, Dec 15, 2010 at 11:00 AM, Alex <broscutama...@gmail.com> wrote:
> > After playing a bit with the latex template, here are the new PDF
> > docs. Doc sources & build script are commited.
>

> > On Wed, Dec 15, 2010 at 9:08 AM, Alex <broscutama...@gmail.com> wrote:
> >> Can you fix the script, so we'll be able to say "for Linux and Mac" in
> >> the docs?
>
> >> On Dec 15, 4:56 am, Piers <pie...@gmail.com> wrote:
> >>> Indy's script would work on Mac if you set the shell var to something
> >>> Mac-like like /dev/disk1s1*. Except he then goes and explicitly refers to "
> >>> /dev/sdc1" instead of his var ;-)
>
> >>> should work, will try tonight on my non-booting SDHC card (which is actually
> >>> a microSD, but I don't think that would matter).
>
> >>> PG
>
> >>> (* the contradiction that is OSX, theres *nothing* Mac-like about
> >>> "/dev/disk1s1"!!)
>
> >> --
> >>http://magiclantern.wikia.com/
>
> >> To post to this group, send email to ml-d...@googlegroups.com
> >> To unsubscribe from this group, send email to ml-devel+u...@googlegroups.com
> >> For more options, visit this group athttp://groups.google.com/group/ml-devel?hl=en
>
>
>

>  magiclantern-2010dec17.550d.fw109.PRERELEASE.a1ex.zip
> 821KViewDownload

[Alex]

Fixed Live View startup bug & LCD remote shoot instability. Check it out!

[robert]

It works, thanks!

On Dec 17, 2:18 pm, Alex <broscutama...@gmail.com> wrote:
> Fixed Live View startup bug & LCD remote shoot instability. Check it out!
>

>  magiclantern-2010dec17.550d.fw109.PRERELEASE.a1ex.zip
> 849KViewDownload

[robert]

Any chance to implement in the future overriding of the 29:59 and 4GB
file size limits? (exFat formatting is supported in 550D, which
doesn't have this limit anymore)

[Alex]

I don't think so; most internal data structures seem to be 32-bit.

[robert]

Hmm... guess that only matters if we're talking about file sizes. If
you look at the 29:59 time limit, that must only be an internal check
somewhere, the software periodically checks the recording time, and if
time > 29:59, stops.

Since we can alter the bitrate now via ML, I think we could live with
lower bitrates allowing recording longer than 29:59 in less than 4GB -
that would still be a life saver. Maybe you could just remove or alter
the time-duration check function.

How about this approach?

On Dec 17, 11:34 pm, Alex <broscutama...@gmail.com> wrote:
> I don't think so; most internal data structures seem to be 32-bit.
>

[robert]

Just a quick calcuulation shows that using a qscale parameter to set
the bitrate to about 8000-9000 kbps would allow about 1 hour of
recording in 4000MB. That bitrate is still very good for long-shot
everyday purposes (eg. HDTV streams on digital cable/satellite are
usually about 7-12 MBps with h264 encoding...)

[Alex]

The idea is not bad. If there's a simple check, and you know where to
look and how to override it... it might be possible.

[Antony Newman]

Robert,

"29:59 time limit"

I think you are talking about the counter that multiplies 1800 x Framerate

I haven't tried to find the Address that R0 points to (5D2 2.0.4) in the following routine ... but I think that is what you are looking for.

AJ.

============================================

ROM:FF851950 ; R0 = address of the maximum number of Frames that a movie can last for
ROM:FF851950
ROM:FF851950 AJ_MOVW_GetMaxFrameCount_n_FramePerSec  ; CODE XREF: AJ_MovieRecorder_MovRecState.c+458 p
ROM:FF851950                 STMFD   SP!, {R4-R6,LR}
ROM:FF851954                 LDR     R5, =aAJ_0x1EE0_MOVWptr_struct_0x00_to_0x04_taskName
ROM:FF851958                 MOV     R6, R0
ROM:FF85195C                 LDR     R0, [R5,#4]     ; aAJ_0x1EE0_MOVWptr_struct_0x04_ptr_to_struct
ROM:FF851960                 ADR     R4, aMovw_getmaxframecount ; "MOVW_GetMaxFrameCount"
ROM:FF851964                 CMP     R0, #0
ROM:FF851968                 BNE     loc_FF851988
ROM:FF85196C                 MOV     R3, R4
ROM:FF851970                 ADR     R2, aSNotInitialized ; "%s : Not initialized"
ROM:FF851974                 MOV     R1, #6
ROM:FF851978                 MOV     R0, #0x30 ; '0'
ROM:FF85197C                 BL      TH_DebugMsg
ROM:FF851980                 MOV     R0, #0x17
ROM:FF851984                 LDMFD   SP!, {R4-R6,PC}
ROM:FF851988 ; ---------------------------------------------------------------------------
ROM:FF851988
ROM:FF851988 loc_FF851988                            ; CODE XREF: AJ_MOVW_GetMaxFrameCount_n_FramePerSec+18 j
ROM:FF851988                 LDR     R0, [R0,#0xC]   ; Struct_A = * [aAJ_0x1EE0_MOVWptr_struct + 0x04]
ROM:FF851988                                         ; R0 = * [Struct_A + 0x0C]
ROM:FF85198C                 MOV     R3, R4
ROM:FF851990                 ADR     R2, off_FF8519EC
ROM:FF851994                 MOV     R1, #3
ROM:FF851998                 AND     R0, R0, #0xFF
ROM:FF85199C                 BL      TH_DebugMsg
ROM:FF8519A0                 LDR     R0, [R5,#4]     ; Struct_A = * [aAJ_0x1EE0_MOVWptr_struct + 0x04]
ROM:FF8519A4                 LDR     R0, [R0,#0xC4]  ; R0 = * [Struct_A + 0xC4]  = Frames per second
ROM:FF8519A8                 CMP     R0, #0
ROM:FF8519AC                 LDRNE   R1, =1800       ; AJ: Maximum length of movie = 30 minutes
ROM:FF8519B0                 MULNE   R0, R1, R0
ROM:FF8519B4                 STRNE   R0, [R6]        ; AJ: save number of frames allowed  at parm.R0
ROM:FF8519B8                 MOVNE   R0, #0
ROM:FF8519BC                 LDMNEFD SP!, {R4-R6,PC}
ROM:FF8519C0                 MOV     R3, R4
ROM:FF8519C4                 ADR     R2, aSFramepersec0 ; "%s : FramePerSec = 0"
ROM:FF8519C8                 MOV     R1, #6
ROM:FF8519CC                 MOV     R0, #0x30 ; '0'
ROM:FF8519D0                 BL      TH_DebugMsg
ROM:FF8519D4                 MOV     R0, #1
ROM:FF8519D8                 LDMFD   SP!, {R4-R6,PC}
ROM:FF8519D8 ; End of function AJ_MOVW_GetMaxFrameCount_n_FramePerSec

[robert]

Cool, that must be it! Add a zero to 1800, the limit would be 300
minutes, that's 5 hours max. But if you add two zeroes, that would be
50 hours limit - which is infinite time from the camera's
perspective :)

> ROM:FF8519AC                 LDRNE   R1, =*1800  *     ;* AJ: Maximum length
> of movie = 30 minutes*

[Alex]

Fixed. I'm looking at the clock in order to scale the value given by
Canon code, and seems to work well in my test. Code is online.

> It is an estimate of the time remaining based on the space on the
> card, but since it doesn't take into account any of the bit rate
> changes since Canon's 1.1.0 firmware, I doubt that the value is very
> realistic any more.

[Alex]

In 550D/1.0.8:

AJ_MOVW_GetMaxFrameCount_n_FramePerSec at FF04B23C:

...DebugMsg's...
if *(196 + MEM(7780)) != 0:
*(arg0) = 1800*MEM(196 + MEM(7780))
...DebugMsg's...

Called by:
mvrRecStart+1160: AJ_MOVW_GetMaxFrameCount_n_FramePerSec(12 + unk_SP)

So it seems to be stored in a local variable in mvrRecStart.

> ROM:FF8519AC                 LDRNE   R1, =*1800  *     ;* AJ: Maximum length
> of movie = 30 minutes*

[lichtjaar]

Hi Alex,

Do you mean your fix enables to rec 50 hours, within the file size
limit of 4GB.
Or is it now also possible to generate bigger than 4 GB files on a
( ExFAT / FAT64 ) filesystem?

bye bye

Lichtjaar

[lichtjaar]

Hi Alex,

Have you already managed to record something like 60 minutes in a 8GB
file on ExFAT / FAT64 ?

On Dec 18, 11:30 pm, Alex <broscutama...@gmail.com> wrote:

[Alex]

No. Before the fix, the time remaining for recording was only accurate
when qscale was not used (because Canon code assumed a fixed
bitrate... i.e. 4 GB = 12 minutes). If qscale was used, the value
displayed was meaningless.

Now, the time displayed in the upper right corner should be good even
if qscale is used (although it's not possible to give an accurate
estimation, because qscale means variable bitrate).

The QScale setting may enable you to record around 1 hour in 4 gb at
the -1 setting. (see http://vimeo.com/17742479 ). In older builds,
QScale = 16 could mean around 3 hours in 4 GB, but the quality is too
low.