Bug Reports against Recent Builds -- Report bugs here

  • 281 Replies
  • 52517 Views
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #190 on: 15 / October / 2012, 18:00:27 »
Advertisements
@waterwingz
So, this issue can not be repeated reliably. Some race condition perhaps? PAUSE_FOR_FILE_COUNTER?

Anyway, if anybody can come up with a reliable way to reproduce the issue, feel free to post here. Until then, keep the JPEGs beside the DNGs ;)

I reproduced and fixed this bug:
shooting.c:
// AUTOISO:EXIF
short shooting_get_iso_market()
{
    short iso_mode = shooting_get_canon_iso_mode();
    if ((iso_mode < 50) || (conf.iso_override_koef && conf.iso_override_value>0) || (conf.iso_bracket_koef && conf.iso_bracket_value))
    {
        short dsv;
        get_property_case(PROPCASE_DELTA_SV, &dsv, sizeof(dsv));
        return (short)((double)shooting_get_iso_market_base() * pow(2, (double)dsv/96.0));
    }
    return iso_mode;
}


The highlighted expression evaluates to true even when ISO bracketing is disabled in the settings, causing the wrong code path to be executed when manual ISO is activated. get_property_case(PROPCASE_DELTA_SV) always returns dsv = 0 when manual ISO is engaged, so the ISO is always set to 100 * pow(2,0) = 100.

While not a real solution, I commented out the offending statement as a workaround, as I don't use ISO bracketing. It is not very useful, as read noise does not change with ISO on these cameras - just underexpose and push the RAW in post.
« Last Edit: 15 / October / 2012, 18:02:18 by evilr00t »

*

Offline philmoz

  • *****
  • 2936
    • Photos
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #191 on: 26 / October / 2012, 03:46:29 »
@waterwingz
So, this issue can not be repeated reliably. Some race condition perhaps? PAUSE_FOR_FILE_COUNTER?

Anyway, if anybody can come up with a reliable way to reproduce the issue, feel free to post here. Until then, keep the JPEGs beside the DNGs ;)

I reproduced and fixed this bug:
shooting.c:
// AUTOISO:EXIF
short shooting_get_iso_market()
{
    short iso_mode = shooting_get_canon_iso_mode();
    if ((iso_mode < 50) || (conf.iso_override_koef && conf.iso_override_value>0) || (conf.iso_bracket_koef && conf.iso_bracket_value))
    {
        short dsv;
        get_property_case(PROPCASE_DELTA_SV, &dsv, sizeof(dsv));
        return (short)((double)shooting_get_iso_market_base() * pow(2, (double)dsv/96.0));
    }
    return iso_mode;
}


The highlighted expression evaluates to true even when ISO bracketing is disabled in the settings, causing the wrong code path to be executed when manual ISO is activated. get_property_case(PROPCASE_DELTA_SV) always returns dsv = 0 when manual ISO is engaged, so the ISO is always set to 100 * pow(2,0) = 100.

While not a real solution, I commented out the offending statement as a workaround, as I don't use ISO bracketing. It is not very useful, as read noise does not change with ISO on these cameras - just underexpose and push the RAW in post.

Thanks for that.

This was introduced with the new menu system where the 'off' value for conf.iso_bracket_value is now -1 not 0.

Should be fixed in revision 2229 (trunk).

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)

Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #192 on: 27 / October / 2012, 04:40:56 »
This was introduced with the new menu system where the 'off' value for conf.iso_bracket_value is now -1 not 0.
Does this mean that some other values in a CCFG2.CFG files created under 1.1.0 will appear to be inverted in the current dev trunk ?

*

Offline philmoz

  • *****
  • 2936
    • Photos
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #193 on: 27 / October / 2012, 05:06:19 »
This was introduced with the new menu system where the 'off' value for conf.iso_bracket_value is now -1 not 0.
Does this mean that some other values in a CCFG2.CFG files created under 1.1.0 will appear to be inverted in the current dev trunk ?


Possibly; but I think the code is ok for the most part - I had a quick look and didn't see any other obvious problems.

You might have issues if you try and use a 1.2 saved CCFG2.CFG file with release 1.1 - having thought about it the ISO bug would probably be triggered in 1.1 in this case.

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)


Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #194 on: 27 / October / 2012, 05:16:16 »
You might have issues if you try and use a 1.2 saved CCFG2.CFG file with release 1.1 - having thought about it the ISO bug would probably be triggered in 1.1 in this case.
Which could explain why I could not reproduce the bug even though I had saved images that proved it had happened.  Good - thanks !

Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #195 on: 05 / December / 2012, 16:34:50 »
sx40hs Firm Ver. GM1.00F, CHDK Ver: 1.1.0-2247

With 50 ASA x 1 selected in CHDK, camera/shutter works OK without flash, but when Flash selected no flash happens and camera freezes with a clunk with lens out.  Camera wakes up by a press of the On/Off button but without CHDK installed.

The problem repeated about 6 times whilst I was identifying the source, but not liking the sound of the 'clunk' when camera froze I stopped at that point!

BillBal

Fairly sure this is a bug, but after reporting the Torch mode as such - who knows  :D  ?!

*

Offline philmoz

  • *****
  • 2936
    • Photos
Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #196 on: 06 / December / 2012, 05:55:00 »
sx40hs Firm Ver. GM1.00F, CHDK Ver: 1.1.0-2247

With 50 ASA x 1 selected in CHDK, camera/shutter works OK without flash, but when Flash selected no flash happens and camera freezes with a clunk with lens out.  Camera wakes up by a press of the On/Off button but without CHDK installed.

The problem repeated about 6 times whilst I was identifying the source, but not liking the sound of the 'clunk' when camera froze I stopped at that point!

BillBal

Fairly sure this is a bug, but after reporting the Torch mode as such - who knows  :D  ?!

Thanks for the report.
70 seems to be the lowest value the SX40 will accept without crashing when using the flash - it mostly seems to happen in AUTO or Scene modes.
I found similar issues on my other cameras at different ISO's (e.g. G12 crashes if ISO set below 50).

The next version on the autobuild server will limit the minimum ISO value that CHDK will try and force for these cameras.
You can still select a lower value in the menu; but CHDK won't use anything below 70 for the SX40.

This is not really a problem as anything below 70 did not alter the image in any way I can see.

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)

Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #197 on: 06 / December / 2012, 16:05:09 »
OK thanks Phil.  I was in C! mode, with a copy of a Manual mode, so neither Auto nor Scene.  Lens fully zoomed with a close-up lens on (to photo a mosquito's wing).  I will stick with ASA 70 then, but CHDK still gives better than the 100 ASA Canon went down to, so not complaining   :)  !


Re: Bug Reports against Recent Builds -- Report bugs here
« Reply #198 on: 06 / December / 2012, 19:14:15 »
I will stick with ASA 70 then, but CHDK still gives better than the 100 ASA Canon went down to, so not complaining
Isn't  CHDK ASA 70 =  Canon ASA 100 anyway ?  (per the ongoing conversation about "market" ISO vs "real" ISO)

int_incr multiplier needs to be initialized
« Reply #199 on: 14 / December / 2012, 15:08:41 »
The int_incr in core/gui_menu.c should be initialized to 1 by default.
Currently, if shortcuts are used to adjust manual focus, the int_incr is zero unless you opened the CHDK menu first.  Zero is not a very useful value for a multiplier -- the keys to increase or decrease have no affect.
Making sure it's initial value is 1 fixes the problem.
Code: [Select]
Index: core/gui_menu.c
=============================================================================
--- core/gui_menu.c     (revision 2375)
+++ core/gui_menu.c     (working copy)
@@ -37,7 +37,7 @@
 static int          x, y;
 static int          w, wplus, num_lines;
 static int          len_bool, len_int, len_enum, len_space, len_br1, len_br2, cl_rect;
-static int          int_incr;
+static int          int_incr=1;
 static unsigned char *item_color;
 
 //-------------------------------------------------------------------
This is for 1.2 unstable, I haven't checked if this affects 1.1.

« Last Edit: 14 / December / 2012, 15:10:12 by SnowLeopard »
A810 1.00e

 

Related Topics