ADTG and CMOS registers

Started by a1ex, June 25, 2013, 11:01:20 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Audionut

ADTG2[8882]
ADTG2[8884]
ADTG2[8886]

These are interesting.  Appear to change the sensitivity of the color channels.

They're @ ADTG6 also.







ADTG6[21] is doing some funky stuff.  0x2aab is like some crop mode in LV.  0x2bab-0x2bde is like some gain control.

Greg

500D:
ADTG3[0] - change the color temperature warm - cold Maybe, sensitivity of the blue channel.

Andy600

ADTG1[13] = everything goes red
ADTG1[1061] = changes shutter, seems quite random
Colorist working with Davinci Resolve, Baselight, Nuke, After Effects & Premier Pro. Occasional Sunday afternoon DOP. Developer of Cinelog-C Colorspace Management and LUTs - www.cinelogdcp.com

g3gg0

Help us with datasheets - Help us with register dumps
magic lantern: 1Magic9991E1eWbGvrsx186GovYCXFbppY, server expenses: [email protected]
ONLY donate for things we have done, not for things you expect!

1%

QuoteADTG1[1061] = changes shutter, seems quite random

This is shutter in zoom mode I think. For some reason still doesn't want to go over 110... this is the only camera I have with this behavior.

Andy600

Quote from: 1% on August 31, 2013, 03:51:59 AM
This is shutter in zoom mode I think. For some reason still doesn't want to go over 110... this is the only camera I have with this behavior.

How is the 5d MkII shutter?

Could it be something to do with the 50D not having native 24/25p and needing FPS override? FPS Override modes (hi jello, low light etc) increase/decrease shutter speed. Maybe there is another timer address?
Colorist working with Davinci Resolve, Baselight, Nuke, After Effects & Premier Pro. Occasional Sunday afternoon DOP. Developer of Cinelog-C Colorspace Management and LUTs - www.cinelogdcp.com

Andy600

Quote from: 1% on August 31, 2013, 03:51:59 AM
This is shutter in zoom mode I think. For some reason still doesn't want to go over 110... this is the only camera I have with this behavior.

I think this might be shutter in zoom: ADTG[105f] N 0x5d7
Colorist working with Davinci Resolve, Baselight, Nuke, After Effects & Premier Pro. Occasional Sunday afternoon DOP. Developer of Cinelog-C Colorspace Management and LUTs - www.cinelogdcp.com

a1ex

Quote from: Andy600 on August 30, 2013, 08:55:19 PM
from 50D 2.35:1 aspect ratio (crop mode)

CMOS[4]   0x481 -> 0x489 looks like vertical interlacing or Dual ISO or something

This looks like some really extreme HDR (columns 1 and 2 are pretty much normal, columns 3 and 4 look like they were taken at 1/50000)

a1ex

Quote from: Andy600 on August 31, 2013, 02:15:55 AM
ADTG1[1061] = changes shutter, seems quite random

Try adding this line:

    {DST_ADTG, 0x1061, 1, "changes shutter"},

=> that 1 means the register is using NRZI values, so it should no longer be random.

Andy600

Quote from: a1ex on August 31, 2013, 08:56:59 AM
Try adding this line:

    {DST_ADTG, 0x1061, 1, "changes shutter"},

=> that 1 means the register is using NRZI values, so it should no longer be random.

Yes, that did the trick. It's 1x shutter control. Just noticed that it's what 1% has already used.
Colorist working with Davinci Resolve, Baselight, Nuke, After Effects & Premier Pro. Occasional Sunday afternoon DOP. Developer of Cinelog-C Colorspace Management and LUTs - www.cinelogdcp.com

Andy600

Quote from: a1ex on August 31, 2013, 08:50:22 AM
This looks like some really extreme HDR (columns 1 and 2 are pretty much normal, columns 3 and 4 look like they were taken at 1/50000)

I also found some less extreme variations between dark and light lines at some settings and some alternated each line, not 2x2 (need to find it again because my battery died). Could this be useful for a psuedo Dual ISO video mode on the 50D? The shot was a frame grab from raw video so it definitely is recordable but vertical rather than horizontal. Not sure if a horizontal reg exists that does the same. I couldn't get cr2hdr to process it because I couldn't re-save as raw after rotation.

BTW, is a similar module for Digic possible? I remember ages ago running a build on the 600D that had some Digic reg viewing thing but I didn't really understand it at the time.
Colorist working with Davinci Resolve, Baselight, Nuke, After Effects & Premier Pro. Occasional Sunday afternoon DOP. Developer of Cinelog-C Colorspace Management and LUTs - www.cinelogdcp.com

1%

I think digic poke is included.

Greg

500D:
ADTG3[0] - 0x149 - color channel
ADTG3[3] - 0x14a - color channel

These values ​​should be the same?

marekk

For 60D it should be:

FRAME_CMOS_ISO_START = 0x407458FC; // CMOS register 0000 - for LiveView, ISO 100 (check in movie mode, not photo!)
FRAME_CMOS_ISO_COUNT =  6; // from ISO 100 to 3200
FRAME_CMOS_ISO_SIZE  =  30; // distance between ISO 100 and ISO 200 addresses, in bytes

PHOTO_CMOS_ISO_START = 0x4074464C; // CMOS register 0000 - for photo mode, ISO 100
PHOTO_CMOS_ISO_COUNT = 6; // from ISO 100 to 3200
PHOTO_CMOS_ISO_SIZE  =  18; // distance between ISO 100 and ISO 200 addresses, in bytes


but I don't know how to set:
CMOS_ISO_BITS = ;
CMOS_FLAG_BITS = ;
CMOS_EXPECTED_FLAG = ;


Quote from: v8rrc on August 30, 2013, 02:18:35 PM
}
    else if (streq(camera_model_short, "60D"))
    { 
        /*
        100 - 0
        200 - 0x024
        400 - 0x048
        800 - 0x06c
        1600 -0x090
        3200 -0x0b4
        */
        is_60d = 1;   

        FRAME_CMOS_ISO_START = 0x407458fc; //
        FRAME_CMOS_ISO_COUNT =          7; //
        FRAME_CMOS_ISO_SIZE  =         32; //

        PHOTO_CMOS_ISO_START = 0x4074464c; //
        PHOTO_CMOS_ISO_COUNT =          6; //
        PHOTO_CMOS_ISO_SIZE  =         18; //

        CMOS_ISO_BITS = 3;
        CMOS_FLAG_BITS = 2;
        CMOS_EXPECTED_FLAG = 0;

Would this work?

1%

It looks like similar values to 600D.

marekk

I've just tested it and I think it works.

Quote from: 1% on September 01, 2013, 04:04:40 PM
It looks like similar values to 600D.

a1ex

Quote from: Andy600 on August 31, 2013, 07:50:24 PM
I also found some less extreme variations between dark and light lines at some settings and some alternated each line, not 2x2 (need to find it again because my battery died). Could this be useful for a psuedo Dual ISO video mode on the 50D?

That would be some sort of dual shutter, which will be a lot more powerful in really extreme situations (e.g. shooting from a dark cave outside), but will have some more motion artifacts.

Postprocessing will be the same, as long as the pattern is two bright lines / two dark lines (or columns, since rotation is trivial).

Greg

CMOS[4] - scroll horizontal in 5x zoom mode

1%

So those are the instant zoom area moves? I was moving it in 1X mode, maybe thats why it acted so odd.

Audionut

Is PHOTO_CMOS_ISO_COUNT a simple count of the amount of ISO steps?

100
200
400
800

would = PHOTO_CMOS_ISO_COUNT = 4?

nanomad

Yes if 1600 has the same cmos the value as 800
EOS 1100D | EOS 650 (No, I didn't forget the D) | Ye Olde Canon EF Lenses ('87): 50 f/1.8 - 28 f/2.8 - 70-210 f/4 | EF-S 18-55 f/3.5-5.6 | Metz 36 AF-5

Audionut

There's some weird settings in dual_iso.c

6D has a count of 7 to ISO 12800 and 25600  ???  Copy/pasting comments?  And why only 7, it should be doing atleast ISO 12800?

nanomad

Well, that depends on the hardware. You should check what ISOs make the CMOS register change and what no. Sometimes the comments can be a bit off. I don't have a 6D but maybe 1% can comment on it
EOS 1100D | EOS 650 (No, I didn't forget the D) | Ye Olde Canon EF Lenses ('87): 50 f/1.8 - 28 f/2.8 - 70-210 f/4 | EF-S 18-55 f/3.5-5.6 | Metz 36 AF-5

1%

6D is like this:


6D 03
- 100 - 0
- 200 - 11
- 400 - 0x22
800 - 0x33
1600 - 0x44
3200 - 0x55
6400 - 0x77
128K+ 0x77


0x77 is the highest the ISO goes which is 6400.

Audionut