supplierdeeply

IXUS 170 Porting

  • 67 Replies
  • 26301 Views
*

Offline blackhole

  • *****
  • 642
  • A590IS 101b
    • Planetary astrophotography
Re: IXUS 170 Porting
« Reply #50 on: 25 / March / 2017, 08:54:57 »
Advertisements
I 'll post it  on the Monday,power supply in my PC is dead.Did you try the test build 3?

Poslano sa mog GT-I9301I koristeći Tapatalk


*

Offline momo

  • *
  • 16
Re: IXUS 170 Porting
« Reply #51 on: 26 / March / 2017, 03:38:04 »
I 'll post it  on the Monday,power supply in my PC is dead.Did you try the test build 3?
I didn't, the camera hasn't come yet, I don't even know the f/w version. Will see...

*

Offline momo

  • *
  • 16
Re: IXUS 170 Porting
« Reply #52 on: 31 / March / 2017, 06:32:19 »
Did you try the test build 3?

@blackhole

My cam happens to be f/w 1.00a, so I did some testing with your build 3, although I really see little reason for testing. As I understand, you own such a camera, and so can test as well as any. Unless we find some obscure bug.

Anyway, my basic testing included the following, all done in P-mode (note that I'm not familiar with the recent CHDK layout and functions, so must've missed something):
- overrides work
- DNG/RAW works, although I don't see much sense using it with that [admin: avoid swearing please] of a lens. The DNG colors are horrible, as might be expected. Still, I'll see what I can do about it with a color target (it is old, but cannot be worse than the DNG colors)
- flicker in shooting mode is more than tolerable, judging by what I read about recent cameras. Do you have a special solution for that?
- no crashes so far

I believe you and srsa_4c should submit the code to SVN, and for alpha-build.

*

Offline blackhole

  • *****
  • 642
  • A590IS 101b
    • Planetary astrophotography
Re: IXUS 170 Porting
« Reply #53 on: 31 / March / 2017, 11:52:16 »
Quote
My cam happens to be f/w 1.00a, so I did some testing with your build 3, although I really see little reason for testing. As I understand, you own such a camera, and so can test as well as any. Unless we find some obscure bug.
I do not have such a camera, otherwise I would have tested myself.
Quote
The DNG colors are horrible, as might be expected. Still, I'll see what I can do about it with a color target (it is old, but cannot be worse than the DNG colors)
Are the colors just a different tonality compared with jpeg or looks like it's all covered with green or pink color?
Quote
- flicker in shooting mode is more than tolerable, judging by what I read about recent cameras. Do you have a special solution for that?
No.
Quote
I believe you and srsa_4c should submit the code to SVN, and for alpha-build.
Attached is the source code used to create the test3



*

Offline momo

  • *
  • 16
Re: IXUS 170 Porting
« Reply #54 on: 31 / March / 2017, 16:31:35 »
Quote
As I understand, you own such a camera, and so can test as well as any.
I do not have such a camera, otherwise I would have tested myself.
Understood. Your effort is much appreciated.
Quote
Are the colors just a different tonality compared with jpeg or looks like it's all covered with green or pink color?
I attach the histogram of a shot. You can call it a blue tint, but it means the colors (the DNG color matrix) are off. As said, I'll take care of it, eventually.

*

Offline srsa_4c

  • ******
  • 3728
Re: IXUS 170 Porting
« Reply #55 on: 02 / April / 2017, 14:58:37 »
Attached is the source code used to create the test3
Seems to be OK (corrected and cleaned up, when compared to the original).

Some notes:

- I don't think the startup is completely correct - a fw routine (sub_FF82D94C) was left in its original form, which means the startup flag written by CHDK (in sub_FF8203C4_my) is overwritten during boot.

As a result, I would expect the camera to start in a certain mode (playback?) regardless of the button used to start it. Pressing and holding the ON/OFF button would likely result in the same start as pressing the playback button.

@momo: can you confirm this?

- The sigfinder's hit for time() is the right one - the ixus160 port was (or still is) using a different function at the time I created this.

Did not verify everything, there might be other issues.


edit:
The camera does start with the current code because the loader routine sets the reboot flag (0x12345678). But the startup flag is indeed overwritten in sub_ff86b1bc.

edit2:
Thinking about it some more, I'm no longer sure that starting in rec mode is impossible, using the current code.
« Last Edit: 02 / April / 2017, 20:31:10 by srsa_4c »

*

Offline blackhole

  • *****
  • 642
  • A590IS 101b
    • Planetary astrophotography
Re: IXUS 170 Porting
« Reply #56 on: 03 / April / 2017, 09:29:55 »
Quote
- The sigfinder's hit for time() is the right one - the ixus160 port was (or still is) using a different function at the time I created this.
I noticed it, but I was not sure why you did it so I left as it was.
Quote
edit:
The camera does start with the current code because the loader routine sets the reboot flag (0x12345678). But the startup flag is indeed overwritten in sub_ff86b1bc.

edit2:
Thinking about it some more, I'm no longer sure that starting in rec mode is impossible, using the current code.
We can only wait for a response from someone the user.

It would be good if someone with a camera can upload one DNG file.
No one has yet answered whether "firmware update method" to load CHDK works or not.

*

Offline momo

  • *
  • 16
Re: IXUS 170 Porting
« Reply #57 on: 06 / April / 2017, 10:07:50 »
As a result, I would expect the camera to start in a certain mode (playback?) regardless of the button used to start it. Pressing and holding the ON/OFF button would likely result in the same start as pressing the playback button.

@momo: can you confirm this?
The camera starts in rec mode with the on/off button, and in playback mode with the playback button. Indeed always starts in playback, sorry for misreporting.

It would be good if someone with a camera can upload one DNG file.
No one has yet answered whether "firmware update method" to load CHDK works or not.
Here is a pair dng/jpeg, and I'm confused. Now the dng colors appear to be about right, whereas the first time I tried they were definitely wrong, see the histogram some posts above. I'll sort it out on occasion.
https://drive.google.com/open?id=0B6QXKSCcVxHSMzl0dWZ5eGRiT00

F/w update starts the cam in review mode. Can be switched to rec and works, but I didn't test extensively.
« Last Edit: 06 / April / 2017, 22:19:55 by momo »


*

Offline reyalp

  • ******
  • 11586
Re: IXUS 170 Porting
« Reply #58 on: 06 / April / 2017, 13:21:42 »
Here is a pair dng/jpeg, and I'm confused. Now the dng colors appear to be about right, whereas the first time I tried they were definitely wrong, see the histogram some posts above. I'll sort it out on occasion.
How are you viewing / converting the DNGs? Some programs like raw therapee and irfanview for example do an auto adjustment which can look really bad depending on the scene. In raw therapee, the "reset" button in the exposure tab will give you a neutral adjustment.
Don't forget what the H stands for.

*

Offline momo

  • *
  • 16
Re: IXUS 170 Porting
« Reply #59 on: 06 / April / 2017, 15:39:51 »
How are you viewing / converting the DNGs? Some programs like raw therapee and irfanview for example do an auto adjustment which can look really bad depending on the scene. In raw therapee, the "reset" button in the exposure tab will give you a neutral adjustment.
Usually Adobe Camera Raw or, for the purposes of color analysis, dcraw.

But you are right, with the first shots there obviously was some weirdness with camera settings and scene (all auto, strange scene and lighting), and then ACR picked white balance ‘as shot’ which was profoundly wrong (and it probably was wrong indeed in DNG, for some reason that I’m not ready yet to investigate).

Anyway, here’s my take at the color matrix

Code: [Select]
983906    -205236   -115756
-352094    1162018    190075
 -26846    157633    459011
It results in WB multipliers somewhat better than before. Not substantial though. But I may’ve made mistakes in calculations, I’ll re-check some day.

Incidentally, how do folks determine the color matrix these days?
« Last Edit: 06 / April / 2017, 15:51:52 by momo »

 

Related Topics