Menu

Show posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Show posts Menu

Messages - Greg

#26
M50 bit depth

Single - 14-bit
Continuous Low - 12-bit?
Continuous High - 12-bit?

Histograms (single, continuous low, continuous high) :


If you want to use 14-bit exposure bracketing, you need to quickly press the shutter button, continuous modes are not 14-bit.
#27
M50 1/200s ISO 100 - White Level

Exif -  11892

Looking at the raw data:
R - 16383
G - 15500 safe limit for my camera
B - 16383

Green channel clipping 15750 :


???
#28
If you remember which prop and you can load ML, you can try to create a minimal autoexec.bin which will try to restore this prop before err is initialized.
It worked on my 500D a few years ago.

If this does not work, you need to contact unbrick.io
:o
#30
Diaphora - IDA PRO plugin for comparing the firmware :
https://github.com/joxeankoret/diaphora

M50 1.0.1 vs 1.0.2
#31
General Chat / Re: The 'when' question
April 26, 2019, 03:25:35 AM
WOW :o
@a1ex, patreon will not be enough. Devs need more money for the ML event in Costa Rica.
@dfort, when ML party? Watch out for falling coconuts!
#32
General Chat / Re: The 'when' question
April 25, 2019, 07:30:10 PM
@dfort, where is your canon, I see an iphone in exif?
;)
#33
Alex, if you need energy try this (lime, lemon, blue is a special mixture only for devs) :

#35
Camera-specific Development / Re: Canon 80D
March 28, 2019, 12:14:10 AM
Quote from: reddeercity on March 27, 2019, 11:38:48 PM
Seeing EOS R @ 179MB/s does the 80D have the same interface but just limited by canon firmware ?

I doubt that 80D has a UHS-II connector.
#36
Quote from: srsa on March 09, 2019, 03:33:45 PM
EOS M50 (1.0.2)

Firmware Version 1.0.2 incorporates the following fix:
1. Improves reliability of communication with external flashes.


It now works with yongnuo flash triggers.
#37
Quote from: Dmytro_ua on March 05, 2019, 07:32:16 PM
These 4 candidates seems to be more interesting (as ML has nothing common with Nikon):
Nikon has more features, focus stacking, n-log 10bit video, etc.
I have not seen ML on R and 5D IV.
#41
You did not need this software to change the language restriction.
PROP 0x01000012 -> http://magiclantern.wikia.com/wiki/Properties/550D
Just compare in qemu with the second 6D.
#42
Reverse Engineering / Re: Sensor Factory Calibration
January 02, 2019, 08:52:59 PM
ISO 100-1600 (analog) changing the first value (A) by 1 affects column_black_level by 1.48
ISO 3200-12800 (digital) changing the first value (A) by 1 affects column_black_level by 2.96
Each camera may react differently to changing these values depending on the SHAD configuration.

update :
- stdev thershold,
- validation,
- save to file,

https://bitbucket.org/Gr3g01/ml-cmos-cal/src/iso-research/modules/cmos_cal/



To run on another camera need :
if (is_camera("500D", "1.1.1"))
{
    PROPAD_GetPropertyData = 0xFF05A57C;
    SENSOR_FULL_WIDTH = 4832;
}
- calibration ratio (1.48) may be different than 500D
#43
Reverse Engineering / Re: Sensor Factory Calibration
December 31, 2018, 12:20:45 AM
Using this module you can change the value of B and click "Read column sensor value" and look at diff.
500D also affects column_black_level - https://www.magiclantern.fm/forum/index.php?topic=23086.msg209326#msg209326

500D
[EDMAC#9] Starting transfer from 0x30A10C to <15>, 4832x4, flags=0x10000


Here, you can also see some 2 parts :
5D3
[EDMAC#24] Starting transfer from 0x174E100 to <15>, 2968x4, flags=0x40020000
[EDMAC#24] Starting transfer from 0x174B2A0 to <15>, 2968x4, flags=0x40020000


Maybe someone forgot about it and treated it again like optical black.
#44
Reverse Engineering / Re: Sensor Factory Calibration
December 30, 2018, 11:04:40 PM
Quote from: g3gg0 on December 30, 2018, 01:40:12 AM
whats their content in your camera?
I do not remember, I just changed some values in RAM related to EDMAC#8 and I saw more defective pixsels (photo mode cr2).

I think that this code is more dangerous now.  :P
https://bitbucket.org/Gr3g01/ml-cmos-cal/commits/642112e37a254c4965280663694b87c6c45ed723?at=iso-research

For the safety of users who do not know what they are doing, the code downloaded from the repository will not be compiled.
//static int (*PROPAD_GetPropertyData) (unsigned int property, void* data, int *length) = 0xFF05A57C; // 500d.111 stubs.h
#45
Reverse Engineering / Re: Sensor Factory Calibration
December 30, 2018, 12:26:02 AM
Canon photo mode function :

if (!raw_iso)
  assert("Iso != 0", "Param\\HivshdParam.c");
PROPAD_GetPropertyData(dword_FF447038[raw_iso - 64], (int)&v2, (int)&v3);
return v2;


Canon live view function :

if(raw_iso < 72)
  raw_ iso = 72;
if(raw_iso > 104)
  raw_iso = 104;
PROPAD_GetPropertyData((signed int)(raw_iso - 72) / 8 + 0x1050000, (int)&v2, (int)&v3);
Then it calculates the offset in the table (depends on the mode and position of the crop mode)


500D VFPN correction (live view)
0x1050000 - ISO 100
0x1050001 - ISO 200
0x1050002 - ISO 400
0x1050003 - ISO 800
0x1050004 - ISO 1600 - 12800




Live view on newer cameras works differently, but the address property seems to be the same.
5D3 ROM string "LiveView Gain Table Ver0.2" :
0x1050000
0x1050001
0x1050002
0x1050003
0x1050004
0x1050005
0x1050006
0x1050007
0x1050081
0x1050082
0x10500D1


BTW, we can find hotpixels table, also possible focus pixels.
For 500D this is EDMAC#8 [DEFM] engine.
#46
Reverse Engineering / Re: Sensor Factory Calibration
December 29, 2018, 01:37:51 AM
500D properties :
VFPN correction (photo mode CR2 and FRSP)
0x0101004D - ISO 100
0x01010050 - ISO 200
0x01010053 - ISO 400
0x01010056 - ISO 800
0x01010059 - ISO 1600
0x0101005C - ISO 3200
0x0101005F - ISO 6400
0x01010062 - ISO 12800

HFPN correction
0x01010096 - ISO 100-12800 FRSP
0x010100A1 - ISO 100-12800 CR2


Attempt to read the RAM address and size using g3gg0's code :
#47
Reverse Engineering / Re: Sensor Factory Calibration
December 18, 2018, 02:20:03 PM
I thought that your clone from Santa again wants to change the value in the ROM of my 500D.  :P
#48
Reverse Engineering / Re: Sensor Factory Calibration
December 17, 2018, 06:11:11 PM
I will not modify these properties. I'm using this with ADTG modifications. So if I started the camera without ML, the pictures will be worse than the factory calibration.
Maybe I did not clearly write what I mean by the publication of the module. I just wanted to create something that will present effects on other cameras. To see if it is useful.

We're in Reverse Engineering, so I'm trying to show how these registers work, RAM editing is more secure when I do not know what it does.

If someone can not find the address in RAM and does not understand what this code does, he should not use it at this stage.
#49
Reverse Engineering / Re: Sensor Factory Calibration
December 16, 2018, 04:31:41 PM
These logs are incomplete.

Quote from: aprofiti on December 16, 2018, 02:29:57 PM
- In EDMAC Tool menu I have set Log very 2000us
Try 100us or 50us.
#50
Reverse Engineering / Re: Sensor Factory Calibration
December 16, 2018, 12:19:00 AM
@Danne,
This log looks like a live view.
We need a regular pic log.

Live View
I looked at the live view in 500D :
EDMAC#9: addr=0x442784 conn=15 cbr=0 name='''''' size='''32x208''' //vertical column
EDMAC#9: addr=0x485fd4 conn=15 cbr=0 name='''''' size='''32x165''' //horizontal line


This looks like a 4x1664 buffer and 4x1320
Looking at hexeditor and RAM, the data in this buffer looks very similar to those in photo mode.

Changing these registers does not affect ML raw. ML takes raw before FPN correction.

Change the value of several columns:


Several lines:


I do not plan further research in live view/video mode. I do not have time for it.