S100 porting thread - page 27 - DryOS Development - CHDK Forum  

S100 porting thread

  • 286 Replies
  • 71942 Views
*

Offline nafraf

  • *****
  • 1306
Re: S100 porting thread
« Reply #260 on: 17 / July / 2014, 07:55:56 »
Advertisements
Test version for 101b.
« Last Edit: 08 / September / 2014, 11:54:34 by nafraf »

Re: S100 porting thread
« Reply #261 on: 17 / July / 2014, 09:35:09 »
I think that the firmware only sets 0x7a94 once, not continuously.

The crash occurred on switch release.
If you immediately zero 0x7a94 on switch release it still crashes.


Re: S100 porting thread
« Reply #262 on: 17 / July / 2014, 10:14:15 »
I think that the firmware only sets 0x7a94 once, not continuously.
The crash occurred on switch release.
If you immediately zero 0x7a94 on switch release it still crashes.
What's curious to me is that "one press" mode in CHDK (and presumably the equivalent in SDM) works. 

In that mode,  when the USB power is applied,  the code does the equivalent of an immediate "full_press" of the shutter button.  I believe that causes the camera to focus and set exposure right away and then proceed to shoot - where it gets trapped by the "wait_untill_remote_button_released" hook.

In two press mode,  the initial press looks like a "half_press" the camera.  It will focus and set exposure and apparently look for "faces".  Quickly releasing and repressing the remote button triggers the additional "full-press" shot and the "wait_untill_remote_button_released" hook.  Going there with face recognition events queued up seems to be a bad thing if the shooting task "waits".   It's hard to understand why that would be the case as with sync disabled it goes through exactly the same code except for not waiting.


Ported :   A1200    SD940   G10    Powershot N    G16

Re: S100 porting thread
« Reply #263 on: 17 / July / 2014, 10:27:17 »
(and presumably the equivalent in SDM) works. 

Yes.

I will now be offline for many hours.



Re: S100 porting thread
« Reply #264 on: 17 / July / 2014, 10:41:07 »
Quote
Test version for 101b.
yes this version works - no crashes

Werner
« Last Edit: 17 / July / 2014, 10:49:50 by wbloos »

Re: S100 porting thread
« Reply #265 on: 17 / July / 2014, 10:51:58 »
Quote
Test version for 101b.
yes this version works - no crashes
Very interesting.

Suggests that nafraf's fix of moving the "wait_for_remote_release" farther down the shooting sequence gives the camera time to "cleanup" all the face events it detected prior to the shooting wait.

Obviously,  a much cleaner solution to the "two press with S100" problem.
Ported :   A1200    SD940   G10    Powershot N    G16

*

Offline nafraf

  • *****
  • 1306
Re: S100 porting thread
« Reply #266 on: 17 / July / 2014, 12:28:39 »
yes this version works - no crashes
Thanks. The patch is now checked in for CHDK 1.2 and 1.3. Changeset 3516 and 3517

Re: S100 porting thread
« Reply #267 on: 18 / July / 2014, 15:53:50 »
The patch is now checked in for CHDK 1.2 and 1.3.

Yes, good work, that is the 'proper' fix.

I doubt that any other cameras with face-detect have the same problem.



Re: S100 porting thread
« Reply #268 on: 15 / November / 2014, 13:47:47 »
A couple of things I noted today while working on MF testing.  I'll drop them here in case I don't get around to fixing them.

  • using the set_zoom() command in scripting crashes the camera.  I tried a changing a couple of the #defines in platform_camera.h but no luck.  Needs a deeper investigation.
  • when shooting DNG images, if the DNG Crop Size is set to [ Full ] then you get purple images. Setting to [ Active ] or [ JPEG ] works properly.
Ported :   A1200    SD940   G10    Powershot N    G16

*

Offline philmoz

  • *****
  • 3123
    • Photos
Re: S100 porting thread
« Reply #269 on: 15 / November / 2014, 15:23:59 »
when shooting DNG images, if the DNG Crop Size is set to [ Full ] then you get purple images. Setting to [ Active ] or [ JPEG ] works properly.

CAM_ACTIVE_AREA_Y1 is odd (11) - despite the warning in camera.h that it must be even ???
When you choose full for the DNG crop, CAM_ACTIVE_AREA_Y1 is set to 0 in the image file.

So the Bayer decoding is starting on different rows in the Bayer matrix - odd vs. even.

This also means the cam_CFAPattern value is wrong.

If you take a Canon RAW image, you can use the Adobe dng_validate program (from the DNG SDK) to get the correct values for cam_CFAPattern, and active area. Use the -v option on dng_validate.
Might also be worth checking the COLOR_MATRIX values - the comments in platform_camera.h say the values are from the G15.

Phil.
CHDK ports:
  sx30is (1.00c, 1.00h, 1.00l, 1.00n & 1.00p)
  g12 (1.00c, 1.00e, 1.00f & 1.00g)
  sx130is (1.01d & 1.01f)
  ixus310hs (1.00a & 1.01a)
  sx40hs (1.00d, 1.00g & 1.00i)
  g1x (1.00e, 1.00f & 1.00g)
  g5x (1.00c, 1.01a, 1.01b)

 

Related Topics