icewing-user Mailing List for iceWing
Brought to you by:
floemker
Menu
▾
▴
icewing-user — Discussion about iceWing and its plugins
You can subscribe to this list here.
| 2005 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(6) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2006 |
Jan
|
Feb
(1) |
Mar
(3) |
Apr
(6) |
May
(5) |
Jun
|
Jul
|
Aug
(5) |
Sep
|
Oct
(14) |
Nov
|
Dec
|
| 2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(1) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2008 |
Jan
|
Feb
|
Mar
|
Apr
(2) |
May
(1) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2009 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
(5) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2012 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(1) |
Nov
|
Dec
|
|
From: Frank L. <flo...@gm...> - 2012-10-03 00:07:13
|
Hello,
After more than three years of development a new iceWing version,
version 0.11, with lots of new features, enhancements, and
polishing is now available.
More information about iceWing is available at:
http://icewing.sourceforge.net
Some highlights of the new release:
- Support for more cameras, e.g. GenICam cameras via the aravis
library, more V4L2 cameras, cameras from IDS, and many more.
- A GUI for changing different grabber properties on the fly.
- Enhanced OpenCV 2.x support.
- Plugins can register callbacks for all widgets.
- Support for the newest versions of all the dependencies.
- And lot's more.
In detail the changes are:
- Added callbacks for all widgets. Any number of them can be attached to
a widget and can be called when the widgets value changes or when the
widget is removed.
- The text of separator widgets is now bold, so it stands out more.
- Set the size of the entry widget for int's based on the max allowed value.
- New function opts_value_get() to get the value of a widget based on it's name.
- Support _NET_WM_ICON: More window managers will now show a window icon.
- New function iw_img_convert_layout() to convert between planar and
interleaved images.
- New function prev_copy_to_image() to get the content of a prevBuffer.
- iw_img_yuvToRgbVis() supports interleaved images.
- Added iw_opencv_render() for displaying OpenCv 2.x cv::Mat images directly.
- Added iw_opencvMat_from_img(), iw_opencv_to_img(), and iw_opencv_to_img_interleaved()
for conversion to/from OpenCv 2.x cv::Mat images.
- Enhanced record plugin:
- Support interleaved images.
- New mode to automatically deactivate recording after one plugin round.
- New CMake options to disable the autodetection of libraries.
- Support xdg-open and okular for viewing the iceWing documentation.
- Added RGB image display in the grab plugin.
- New options "-nyuv" and "-nrgb" for the grab plugin to make output
names of images configurable.
- Enhanced handling of duplicated frames when compiled without FFmpeg.
- Show only a warning and do not exit when grabbing an image did not work.
- Remove any double entries in the grabbing options which are shown
on the grabImage page.
- Added new aravis grabbing driver, supporting GenICam compatible cameras.
- Added new NET iCube grabbing driver.
- Added new IDS Imaging uEye grabbing driver.
Currently not all image widths are supported.
- Enhanced firewire driver:
- Added option 'propX' to set and get different properties.
- The format7 ROI can be changed on the fly.
- Option to set the iso speed.
- Option to reset the bus before starting capturing.
- Autodetection of the needed bayer pattern.
- Always return the newest image in the queue for the dc1394 V2 library,
as it was already done for the V1 library.
- Optionally use the image embedded timestamp from different point grey cameras.
- Enhanced V4L2 driver:
- Support for additional pixel formats by using libv4lconvert for
unknown formats.
- It seems that for some cameras/drivers setting a property before
grabbing an image does not take effect. So set the properties during
opening the device and after grabbing the first image.
- Do not fail but show only a warning if setting a property failed.
- Grabbed color images can be interpreted as gray images with an
increased width or height.
- Correct/complete support for bool and menu properties.
- Enhanced MATRIX VISION impact acquire driver:
- Added option to load a configuration file.
- On init try longer to get a valid image, the camera might take
some time to get up.
- Additional support for the pixel formats
ibpfMono12Packed_V2, ibpfYUV422_10Packed, ibpfYUV422_UYVY_10Packed,
ibpfBGR101010Packed_V2, ibpfBGR888Packed, ibpfYUV422_UYVYPacked,
ibpfYUV444_10Packed, ibpfYUV444Packed, ibpfYUV444_UYV_10Packed,
ibpfYUV444_UYVPacked
- Framework for changing grabber properties and other grabber
parameters on the fly.
- Optional GUI for changing all grabber properties on the fly.
Bug fixes:
- Wrong position of the tips for the log histogram.
- opts_float() corrections:
- Do not set longer strings as specified.
- Enhanced width of the entry widget.
- Ensure that no value gets out of the specified value range.
- Some cameras need a longer timeout for grabbing an image via V4L2.
- Fixed bayer decoding methods bilinear interpolation and smooth
hue transition interpolation on 64bit systems.
- Setting stringenter widgets changed only the widgets content,
not the associated variable.
- Fixed TSC reading on 64bit systems.
- Fixed compile error in icewing-control with gcc 4.4 (and newer ones).
- Fixed setting menu hot keys for GTK2 with icewing-control.
- Small resource leak fixes in error cases.
- -iconic option works now with GTK2.
- Explicitly link against all libraries which are used directly,
needed by newer linkers.
- Fixed non uchar OpenCV <> iceWing multi-planed image conversion
for images with a widthStep != 0.
- A press on the "read next/prev..." image navigation button was sometimes ignored.
- Fixed hanging shmdata cleanup on the receiver side.
Bye,
--
Frank Loemker
|
|
From: Frank L. <flo...@Te...> - 2009-06-21 20:24:39
|
Hello, On Wed, Jun 17, 2009 at 10:29:55AM +0200, Christian Lang wrote: >> On Tue, Jun 16, 2009 at 12:28:10PM +0200, Christian Lang wrote: >>> In case one uses "opts_value_set" with an value of higher precision in >>> terms of post decimal positions, the value is truncated to that >>> implicitly given precision by the call to "opts_value_set". In addition, >>> gtk produces a beep to indicate that issue (this is at least my >>> interpretation of the situation). >>> >> >> Hmmm, the beep should not happen. During my tests with gtk 2.8 and gtk 2.12 >> no beep was produced. And I would call it a bug if one occurs. SVN head should now no longer beep when opts_value_set() is called. Bye, -- Frank Loemker |
|
From: Christian L. <chr...@un...> - 2009-06-17 08:30:08
|
Hello Frank,
thanks a lot! With the help of the "float_get_digits" function, I
modified my code so that it does no longer set values with a too high
precision, which solved the problem for me.
Kind regards,
Christian
Frank Loemker wrote:
> Helo,
>
> On Tue, Jun 16, 2009 at 12:28:10PM +0200, Christian Lang wrote:
>
>> I have a question about the "opts_float_create" and "opts_value_set"
>> functions of iceWing. When "opts_float_create" is used to create a slider,
>> the given range for the value seems to determine the precision of that
>> value implicitly.
>>
>
> Yes, that's right. The corresponding code is:
>
> static int float_get_digits (float left, float right)
> {
> int digits = 2;
> int i = 1/(1.01*(right-left));
>
> if (right-left >= 100) return 1;
> while (i > 0) {
> digits++;
> i = i/10;
> }
> return digits;
> }
>
>
>
>> In case one uses "opts_value_set" with an value of higher
>> precision in terms of post decimal positions, the value is truncated to
>> that implicitly given precision by the call to "opts_value_set". In
>> addition, gtk produces a beep to indicate that issue (this is at least my
>> interpretation of the situation).
>>
>
> Hmmm, the beep should not happen. During my tests with gtk 2.8 and gtk 2.12
> no beep was produced. And I would call it a bug if one occurs.
>
> There is a gtk setting which sounds as if it might be involved:
> "gtk-error-bell"
> When TRUE, keyboard navigation and other input-related errors will
> cause a beep. Since the error bell is implemented using
> gdk_window_beep(), the windowing system may offer ways to configure
> the error bell in many ways, such as flashing the window or similar
> visual effects.
>
>
>> I attached a small example which shows the described behavior. It can be
>> compiled with "g++ -shared `icewing-config --cflags` -o libTestPlug.so
>> TestPlug.cpp" and started by "icewing -l ./libTestPlug.so". When the slider
>> is set to a value with two post decimal positions, it is not changed by the
>> call to "opts_float_create". But if a value with three post decimal
>> postions is used, it is truncated to two post decimal positions and the
>> "beep" occurs.
>>
>> My question is now: How exactly is the precision of the value determined? I
>> would like to now this to avoid unintended changes in the value and also to
>> avoid the "beep" every time a value with too high precision is set.
>>
>
> Bye,
>
>
|
|
From: Frank L. <flo...@Te...> - 2009-06-16 22:50:24
|
Helo,
On Tue, Jun 16, 2009 at 12:28:10PM +0200, Christian Lang wrote:
> I have a question about the "opts_float_create" and "opts_value_set"
> functions of iceWing. When "opts_float_create" is used to create a slider,
> the given range for the value seems to determine the precision of that
> value implicitly.
Yes, that's right. The corresponding code is:
static int float_get_digits (float left, float right)
{
int digits = 2;
int i = 1/(1.01*(right-left));
if (right-left >= 100) return 1;
while (i > 0) {
digits++;
i = i/10;
}
return digits;
}
> In case one uses "opts_value_set" with an value of higher
> precision in terms of post decimal positions, the value is truncated to
> that implicitly given precision by the call to "opts_value_set". In
> addition, gtk produces a beep to indicate that issue (this is at least my
> interpretation of the situation).
Hmmm, the beep should not happen. During my tests with gtk 2.8 and gtk 2.12
no beep was produced. And I would call it a bug if one occurs.
There is a gtk setting which sounds as if it might be involved:
"gtk-error-bell"
When TRUE, keyboard navigation and other input-related errors will
cause a beep. Since the error bell is implemented using
gdk_window_beep(), the windowing system may offer ways to configure
the error bell in many ways, such as flashing the window or similar
visual effects.
> I attached a small example which shows the described behavior. It can be
> compiled with "g++ -shared `icewing-config --cflags` -o libTestPlug.so
> TestPlug.cpp" and started by "icewing -l ./libTestPlug.so". When the slider
> is set to a value with two post decimal positions, it is not changed by the
> call to "opts_float_create". But if a value with three post decimal
> postions is used, it is truncated to two post decimal positions and the
> "beep" occurs.
>
> My question is now: How exactly is the precision of the value determined? I
> would like to now this to avoid unintended changes in the value and also to
> avoid the "beep" every time a value with too high precision is set.
Bye,
--
Frank Loemker
|
|
From: Christian L. <chr...@un...> - 2009-06-16 10:29:13
|
Dear all, I have a question about the "opts_float_create" and "opts_value_set" functions of iceWing. When "opts_float_create" is used to create a slider, the given range for the value seems to determine the precision of that value implicitly. In case one uses "opts_value_set" with an value of higher precision in terms of post decimal positions, the value is truncated to that implicitly given precision by the call to "opts_value_set". In addition, gtk produces a beep to indicate that issue (this is at least my interpretation of the situation). I attached a small example which shows the described behavior. It can be compiled with "g++ -shared `icewing-config --cflags` -o libTestPlug.so TestPlug.cpp" and started by "icewing -l ./libTestPlug.so". When the slider is set to a value with two post decimal positions, it is not changed by the call to "opts_float_create". But if a value with three post decimal postions is used, it is truncated to two post decimal positions and the "beep" occurs. My question is now: How exactly is the precision of the value determined? I would like to now this to avoid unintended changes in the value and also to avoid the "beep" every time a value with too high precision is set. Thank you very much. Kind regards, Christian |
|
From: Frank L. <flo...@Te...> - 2009-06-07 07:56:12
|
Hello,
After three years of development a new iceWing version,
version 0.10, with lots of new features, enhancements, and
polishing is now available.
For the list of changes since the last release candidate of iceWing,
version 0.10rc1, see the end of this message.
More information about iceWing is available at:
http://icewing.sourceforge.net/
Some highlights of the new release:
- Gtk+ 2.x support, Gtk+ 1.2 is as well still supported.
- Greatly extended grabber support.
- Extended movie support.
- New CMake based build system.
- And lot's more.
In detail the changes are:
- New CMake based build system (CMakeLists.txt, cmake/, ...).
- pkg-config support.
- Optionally support GTK V2.x, GTK V1.2 is as well still supported.
Assume UTF-8 if GTK V2.x is used.
- Renamed the page 'Other' to 'GrabImage1'.
- Added main menu to the main iceWing window; moved 'Quit' and buttons
from the 'Other' page to the menu.
- All main menu entries and most of the image context menu items
can be remote controlled.
- The planes, which are rendered in image windows (normally 0 or 0,
1, 2) can be freely changed.
- Changed default color space for the "-sp" option to RGB.
- Added 'Close Window' to the image context menus.
- Draw histogram tips in a different color.
- Info Window:
- New layout.
- Display of the 'grab' point in the image windows.
- Possibility to average the color across a square area.
- Support pattern matching for setting widgets.
- Extended movie support:
- Added support for writing RAW444, RAW420, FFV1, and MPEG4 encoded video
files to the main program and the record plugin.
- Allow reading and writing of RAW444 and RAW420 movies without any
external dependencies.
- Expand instead of crop MJPEG's to a size divisible by 16.
- Support saving of PNG (JPEG) images, if the source image has more
than 4 (3) planes.
- Added display of the icewing.pdf manual to the help menu.
- Support all Latin-1 glyphs for text rendering.
- Enhanced plugins page in the "Plugin info" window:
- List all installed plugins.
- Allow to filter the displayed plugins.
- Show plugin name, arguments, and help message.
- Arguments for new plugin instances can be specified.
- Plugins are searched, and can be installed, additionally at
${ICEWING_PLUGIN_PATH}/plugins/ (colon separated variable) and
~/.icewing/plugins/.
- Possibility to specify plugin dependencies ("ident", "ident()", and
"ident(plugs)") for plug_observ_data() and plug_observ_data_remove().
- Added native INET sockets based communication protocol to the remotectrl
plugin and the icewing-control program.
- Allow to specify mnemonics ('_'), accelerators ('<accel>'), and stock
icons ('<"stock" icon>') for menus.
- iceWing option "-d" now only disables the process() function.
- New plugin shmdata to distribute data via shared memory and local
sockets to other local processes.
- Extended record plugin:
- Support saving of >8bit PPM images.
- Support for new movie encodings RAW444, RAW420, FFV1, and MPEG4.
- Use the second thread for movie saving.
- Allow to save images of different plugins in different movie files.
- Do not parse ".gtkrc" (the GTK lib does it already), but parse "~/.icewing/gtkrc".
- The grabber device options can be changed during runtime.
- New bayer decoding method: Adaptive homogeneity-directed interpolation.
- Stereo decoding for 16 bit mono source images.
- Added new MATRIX VISION impact acquire grabbing driver.
- Enhanced firewire driver:
- Support format 7 (partial image grabbing modes).
- Support version 2 of the libdc1394 library.
- Enhanced V4L2 driver:
- Support resolutions greater than 768 x 576.
- Support printing / setting the video standard.
- Support more camera properties (control class and ones with none continuous IDs).
- noselect option to work around a zoran kernel driver bug.
- Stereo and bayer decoding.
- Support for the MJPEG pixel format.
- Stereo and bayer decoding for the unicap driver.
- Escape closes most of the dialogs.
- New plugins chapter and new program chapter in the documentation.
Bug fixes:
- Ignore all modifier keys besides shift and control for the page select hotkey.
- Fixed possible deadlock if no hotkey for an option page could be found.
- Fixed option parsing for the alpha camera driver.
- Fixed '-x' for the remotectrl plugin.
- Use the quality setting in the record plugin.
- During saving of movies: Crop/Extend frames to the size of the first frame.
- Fixed possible endless loop during reading of a movie if the movie does
not start with image number 1.
- Bayer decoding of images with an odd height.
- Incompatible changes:
- Renamed the define DEBUG to IW_DEBUG
- Renamed and slightly changed the defines DEFAULT_WIDTH/HEIGHT to
IW_PAL_WIDTH/HEIGHT.
The changes since iceWing 0.10rc1:
- Small documentation enhancements, especially much more complete index.
- Bug fixes:
- Fixed reinit of the grabber after a failed grabber init.
- Enhanced installation of empty directories for the CMake build system,
patch by Patrick Holthaus.
Bye,
--
Frank Loemker
|
|
From: Frank L. <flo...@Te...> - 2009-04-23 00:11:08
|
Hello,
A first release candidate for the upcoming 0.10 version of iceWing is now
available. Lots of new features, enhancements, and polishing were added
during the last three years since the last release.
Please download it, compile it on as many systems as possible,
and stress test it.
Some highlights of the new release:
- Gtk+ 2.x support, Gtk+ 1.2 is as well still supported.
- Greatly extended grabber support.
- Extended movie support.
- New CMake based build system.
- And lot's more.
In detail the changes are:
- New CMake based build system (CMakeLists.txt, cmake/, ...).
- pkg-config support.
- Optionally support GTK V2.x, GTK V1.2 is as well still supported.
Assume UTF-8 if GTK V2.x is used.
- Renamed the page 'Other' to 'GrabImage1'.
- Added main menu to the main iceWing window; moved 'Quit' and buttons
from the 'Other' page to the menu.
- All main menu entries and most of the image context menu items
can be remote controlled.
- The planes, which are rendered in image windows (normally 0 or 0,
1, 2) can be freely changed.
- Changed default color space for the "-sp" option to RGB.
- Added 'Close Window' to the image context menus.
- Draw histogram tips in a different color.
- Info Window:
- New layout.
- Display of the 'grab' point in the image windows.
- Possibility to average the color across a square area.
- Support pattern matching for setting widgets.
- Extended movie support:
- Added support for writing RAW444, RAW420, FFV1, and MPEG4 encoded video
files to the main program and the record plugin.
- Allow reading and writing of RAW444 and RAW420 movies without any
external dependencies.
- Expand instead of crop MJPEG's to a size divisible by 16.
- Support saving of PNG (JPEG) images, if the source image has more
than 4 (3) planes.
- Added display of the icewing.pdf manual to the help menu.
- Support all Latin-1 glyphs for text rendering.
- Enhanced plugins page in the "Plugin info" window:
- List all installed plugins.
- Allow to filter the displayed plugins.
- Show plugin name, arguments, and help message.
- Arguments for new plugin instances can be specified.
- Plugins are searched, and can be installed, additionally at
${ICEWING_PLUGIN_PATH}/plugins/ (colon separated variable) and
~/.icewing/plugins/.
- Possibility to specify plugin dependencies ("ident", "ident()", and
"ident(plugs)") for plug_observ_data() and plug_observ_data_remove().
- Added native INET sockets based communication protocol to the remotectrl
plugin and the icewing-control program.
- Allow to specify mnemonics ('_'), accelerators ('<accel>'), and stock
icons ('<"stock" icon>') for menus.
- iceWing option "-d" now only disables the process() function.
- New plugin shmdata to distribute data via shared memory and local
sockets to other local processes.
- Extended record plugin:
- Support saving of >8bit PPM images.
- Support for new movie encodings RAW444, RAW420, FFV1, and MPEG4.
- Use the second thread for movie saving.
- Allow to save images of different plugins in different movie files.
- Do not parse ".gtkrc" (the GTK lib does it already), but parse "~/.icewing/gtkrc".
- The grabber device options can be changed during runtime.
- New bayer decoding method: Adaptive homogeneity-directed interpolation.
- Stereo decoding for 16 bit mono source images.
- Added new MATRIX VISION impact acquire grabbing driver.
- Enhanced firewire driver:
- Support format 7 (partial image grabbing modes).
- Support version 2 of the libdc1394 library.
- Enhanced V4L2 driver:
- Support resolutions greater than 768 x 576.
- Support printing / setting the video standard.
- Support more camera properties (control class and ones with none continuous IDs).
- noselect option to work around a zoran kernel driver bug.
- Stereo and bayer decoding.
- Support for the MJPEG pixel format.
- Stereo and bayer decoding for the unicap driver.
- Escape closes most of the dialogs.
- New plugins chapter and new program chapter in the documentation.
Bug fixes:
- Ignore all modifier keys besides shift and control for the page select hotkey.
- Fixed possible deadlock if no hotkey for an option page could be found.
- Fixed option parsing for the alpha camera driver.
- Fixed '-x' for the remotectrl plugin.
- Use the quality setting in the record plugin.
- During saving of movies: Crop/Extend frames to the size of the first frame.
- Fixed possible endless loop during reading of a movie if the movie does
not start with image number 1.
- Bayer decoding of images with an odd height.
- Incompatible changes:
- Renamed the define DEBUG to IW_DEBUG
- Renamed and slightly changed the defines DEFAULT_WIDTH/HEIGHT to
IW_PAL_WIDTH/HEIGHT.
If everything goes well, the final 0.10 release should follow shortly.
Bye,
--
Frank Loemker
|
|
From: Christian L. <chr...@un...> - 2008-05-05 08:15:13
|
Hello, >> I encountered a problem with the "opts_value_set" function of iceWing. >> Please consider the attached "minimal" test plugin which shows the issue. >> >> [...] >> > > Thanks a lot for the bug report. Hopefully the bug is now fixed. Please > svn up and test it. It was a missing return value from the callback of > the window delete event. So, depending on the compiler, the used libs, > the exact code path -- or something like that -- the main window was > deleted to early, resulting in the SegV later on. > Thank you for fixing the bug. It works fine now. > A > return TRUE; > is missing here (or a return FALSE), to indicate wether other plugins > should be called. > Thank you for that hint. Bye, Christian |
|
From: <flo...@Te...> - 2008-04-29 21:35:56
|
Hello,
On Mon, Apr 28, 2008 at 10:13:06AM +0200, Christian Lang wrote:
> I encountered a problem with the "opts_value_set" function of iceWing.
> Please consider the attached "minimal" test plugin which shows the issue.
>
> Compiled with "g++ -shared `icewing-config --cflags` -o libTestPlug.so
> TestPlug.cpp", the plugin bevaves as expected until iceWing is closed.
> When an "older" iceWing version with gtk1 is used and exited via the
> "Quit" button, everything is fine. But if iceWing is closed via the
> close-Item ("x") of the window manager, iceWing aborts with a
> segmentation fault. This also happens with a "newer" gtk2 version of
> iceWing (which has no "Quit" button).
Thanks a lot for the bug report. Hopefully the bug is now fixed. Please
svn up and test it. It was a missing return value from the callback of
the window delete event. So, depending on the compiler, the used libs,
the exact code path -- or something like that -- the main window was
deleted to early, resulting in the SegV later on.
> #include <gui/Ggui.h>
> #include <gui/Goptions.h>
> #include <main/plugin_cxx.h>
>
> namespace ICEWING {
>
> class TestPlug : public Plugin {
>
> public:
>
> TestPlug (char *name) : Plugin(name), value(0.5) { }
>
> void Init (grabParameter *para, int argc, char **argv) {
> plug_observ_data(this,const_cast<char*>("start"));
> }
>
> int InitOptions () {
> int p = plug_add_default_page(this,NULL,(plugPageFlags)0);
> opts_float_create(p,"value","set value",&value,0.1,1.0);
> return p;
> }
>
> bool Process (char *ident, plugData *data) {
> opts_value_set("value",&value);
A
return TRUE;
is missing here (or a return FALSE), to indicate wether other plugins
should be called.
Bye,
--
Frank Loemker
|
|
From: Christian L. <chr...@un...> - 2008-04-28 08:13:06
|
Dear all,
I encountered a problem with the "opts_value_set" function of iceWing.
Please consider the attached "minimal" test plugin which shows the issue.
Compiled with "g++ -shared `icewing-config --cflags` -o libTestPlug.so
TestPlug.cpp", the plugin bevaves as expected until iceWing is closed.
When an "older" iceWing version with gtk1 is used and exited via the
"Quit" button, everything is fine. But if iceWing is closed via the
close-Item ("x") of the window manager, iceWing aborts with a
segmentation fault. This also happens with a "newer" gtk2 version of
iceWing (which has no "Quit" button). Here is the gdb backtrace, tested
with an iceWing version got from the svn repository last week:
(gdb) file icewing
[...]
(gdb) run -l ./libTestPlug.so
[...]
> plugin.c ( 674): TestPlug1: Processing data from 'MainLoop'...
> plugin.c ( 674): TestPlug1: Processing data from 'MainLoop'...
> plugin.c ( 674): TestPlug1: Processing data from 'MainLoop'...
(icewing:29231): GLib-GObject-WARNING **: invalid uninstantiatable type
`(null)' in cast to `GtkObject'
(icewing:29231): Gtk-CRITICAL **: gtk_object_get_user_data: assertion
`GTK_IS_OBJECT (object)' failed
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6d3fb90 (LWP 29232)]
0x0808fb91 in opts_float_set (opts=0x8b16398, value=0x89621b8) at
gui/Goptions.c:1363
1363 if (f_val>=adjustment->lower && f_val<=adjustment->upper) {
(gdb) bt
#0 0x0808fb91 in opts_float_set (opts=0x8b16398, value=0x89621b8) at
gui/Goptions.c:1363
#1 0x0808e923 in opts_value_set (title=0xb7082ecb "value",
value=0x89621b8) at gui/Goptions.c:615
#2 0xb7082dde in ICEWING::TestPlug::Process () from ./libTestPlug.so
#3 0x080dcfeb in PluginProcess (plugDef=0x8962194, ident=0x8961d80
"start", data=0x8b1dc28)
at main/plugin_cxx.cpp:85
#4 0x08073add in plug_process (plug=0x8962208, ident=0x8961d80 "start",
data=0x8b1dc28) at main/plugin.c:677
#5 0x08077ee9 in plug_main_iteration () at main/plugin_comm.c:508
#6 0x0806a894 in iw_mainloop_start (para=0x858120c) at main/mainloop.c:532
#7 0xb76af192 in start_thread () from /lib/libpthread.so.0
#8 0xb752002e in clone () from /lib/libc.so.6
The problem does not occur always, but in the majority of runs. It seems
to dependent somehow on the runtime of iceWing, whether it aborts with a
sigmentation violation are exits normally, probably some thread issue.
Regards,
Christian
|
|
From: Marc H. <mha...@te...> - 2007-07-23 13:59:34
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hey, Some of my student just did a new binarization plugin (old Otsu method: N. Otsu, "A Threshold Selection Method from GrayLevel Histogram", IEEE Trans. Systems, Man and Cybernetics, Vo l. 9, 1979, pp. 62-69.). Use it if you want a global adaptive binarization of your images. Available on request and based on OpenCV (mainly histogram routines). Cheers, Marc - -- - ------------------------------------------------------------------------ Dr.-Ing. Marc Hanheide Phone: +49(521)106-5159 Angewandte Informatik Fax: +49(521)106-2992 Technische Fakultaet EMail: mha...@te... Universitaet Bielefeld Room: N5-100 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGpLRE7yyxwySZfrgRArcXAJ9coVHMHoz5wwN+cFNPkvdNd9DLPwCfZKii +Frrua7xd7hlN6NePhCasqE= =rNLl -----END PGP SIGNATURE----- |
|
From: Axel H. <ah...@Te...> - 2006-10-10 21:24:18
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Dear Frank! You're the best! The patch worked. :-) Thanks a lot, now it works perfectly alright. Have a nice evening, cheers, Axel Frank Loemker wrote, On 10.10.2006 23:13: > Hello, > > On Tue, Oct 10, 2006 at 01:22:09PM +0200, Axel Haasch wrote: >> you're correct, I'm using the svideo input, not composite. Therefore I just >> tried out your proposals but neither doesn't work. >> >> In the following you see the output of the correct input=1: S-Video. I kicked out >> most of the redundant lines but nevertheless it looks as bad as with the wrong >> input used. >> >> ... >> >> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QBUF - type=1, index=0 >> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QBUF - type=1, index=1 >> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_STREAMON >> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: set_vfe() - width = 768, height = 576 >> Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x41000000 >> Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: zoran_poll() - no buffers queued > > The TODO list of the zoran driver at > > http://mjpeg.sourceforge.net/driver-zoran/development.php > > contains the point > > * zoran_poll() is broken. > > . This would nicely (?) fit to this log ouput. First two buffers are > successfully queued and the grabbing is started. Then iceWing calls > select() to check if any grabbed images are available. The select call > calls zoran_poll(), which (perhaps after some time) thinks that there > are no more buffers although the two buffers should still be queued. > > If this is really the case perhaps the attached patch would fix or rather > by-pass the problem. The patch introduces a new option "noselect" for the > v4l2 driver. If used the driver does no longer call select() to check for > available images. So could you check if a patched iceWing will work if > called as > > icewing -sg v debug:noselect:input=1 > > ? And use > > debug=4 v4l_bufsize=900 v4l_nbufs=2 > > for the zoran kernel driver. > > Bye, > > > > ------------------------------------------------------------------------ > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share your > opinions on IT & business topics through brief surveys -- and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > > > ------------------------------------------------------------------------ > > _______________________________________________ > Icewing-user mailing list > Ice...@li... > https://lists.sourceforge.net/lists/listinfo/icewing-user - -- Int. Conference on Computer Vision Systems 07: Vision Systems in the Real World 21st-24th March, Bielefeld, Germany http://www.icvs2007.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iQEVAwUBRSwPdjuI6ru3QAHZAQqW0Qf+OfystO1j6xrNFg5lSwz0bREztQ2Q41Gh WiLBoJJaSf2BkGjQ7Emldv9Gy+Gz8HeRFoCFhEI/DFBm2Hf6E9lPRWa6o2ignUbs BNGMOrVxVwPCR1I3ysF1xL5oTO/SkzTIUo9uNhwXv5SQ0blgrNo1qF6vqGzBoU0w SUc6527IjFG6npSvrBvITFD5At3+qGUJRsw+LTznJlzRBQsJ+VeQPFOhEXguBfuQ Se7sLwI7Ot1XwjoVTkD8J5qMsu3s4zFr1IOuSEfY4xU8GMux1oJhjqaH47qeJP2/ 2jZw76HxuWUpYPj3IIx5BhNJpdP4vLZl6Ky9HR2sIpgfQa/gy79Yfw== =7T9I -----END PGP SIGNATURE----- |
|
From: Frank L. <flo...@Te...> - 2006-10-10 21:13:37
|
Hello,
On Tue, Oct 10, 2006 at 01:22:09PM +0200, Axel Haasch wrote:
> you're correct, I'm using the svideo input, not composite. Therefore I just
> tried out your proposals but neither doesn't work.
>
> In the following you see the output of the correct input=1: S-Video. I kicked out
> most of the redundant lines but nevertheless it looks as bad as with the wrong
> input used.
>
> ...
>
> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QBUF - type=1, index=0
> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QBUF - type=1, index=1
> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_STREAMON
> Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: set_vfe() - width = 768, height = 576
> Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x41000000
> Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: zoran_poll() - no buffers queued
The TODO list of the zoran driver at
http://mjpeg.sourceforge.net/driver-zoran/development.php
contains the point
* zoran_poll() is broken.
. This would nicely (?) fit to this log ouput. First two buffers are
successfully queued and the grabbing is started. Then iceWing calls
select() to check if any grabbed images are available. The select call
calls zoran_poll(), which (perhaps after some time) thinks that there
are no more buffers although the two buffers should still be queued.
If this is really the case perhaps the attached patch would fix or rather
by-pass the problem. The patch introduces a new option "noselect" for the
v4l2 driver. If used the driver does no longer call select() to check for
available images. So could you check if a patched iceWing will work if
called as
icewing -sg v debug:noselect:input=1
? And use
debug=4 v4l_bufsize=900 v4l_nbufs=2
for the zoran kernel driver.
Bye,
--
Frank Loemker
|
|
From: Axel H. <ah...@Te...> - 2006-10-10 11:22:20
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Hello, you're correct, I'm using the svideo input, not composite. Therefore I just tried out your proposals but neither doesn't work. In the following you see the output of the correct input=1: S-Video. I kicked out most of the redundant lines but nevertheless it looks as bad as with the wrong input used. Axel Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[13372]), users(-)=0 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: check_jpg_settings() - dec: 1, Hdcm: 1, Vdcm: 1, Tdcm: 1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: check_jpg_settings() - x: 0, y: 0, w: 768, y: 288 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: dc10plus_init() Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x00000000 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - sleep GPIO=0x81008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - wake GPIO=0x89008888 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x00000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x00000000 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - wake GPIO=0xc9008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - sleep GPIO=0xc1008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: UNKNOWN ioctl cmd: 0xc0685624 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: UNKNOWN ioctl cmd: 0xc0685624 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: UNKNOWN ioctl cmd: 0xc0385602 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_STREAMOFF Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: zoran_close(icewing, pid=[13372]), users(+)=1 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - wake GPIO=0xc9008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - sleep GPIO=0xc1008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: interrupts received: GIRQ1:6 queue_state=0/0/0/0 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: zoran_close() done Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[13372]), users(-)=0 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: check_jpg_settings() - dec: 1, Hdcm: 1, Vdcm: 1, Tdcm: 1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: check_jpg_settings() - x: 0, y: 0, w: 768, y: 288 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: dc10plus_init() Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - sleep GPIO=0x81008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - wake GPIO=0x89008888 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x00000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x00000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x00000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x00000000 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - wake GPIO=0xc9008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - sleep GPIO=0xc1008888 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYCTRL - id=9963776 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_G_CTRL - id=9963776 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYCTRL - id=9963777 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_G_CTRL - id=9963777 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYCTRL - id=9963778 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_G_CTRL - id=9963778 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYCTRL - id=9963779 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_G_CTRL - id=9963779 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYCTRL - id=9963780 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYCTRL - id=134217728 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUM_FMT - index=0 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUM_FMT - index=1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUM_FMT - index=2 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUM_FMT - index=3 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUM_FMT - index=4 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUM_FMT - index=5 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUM_FMT - index=6 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_TRY_FMT - type=1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_S_FMT - type=1, size=768x576, fmt=0x56595559 (YUYV) Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_S_CTRL - id=134217732 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUMINPUT - index=0 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUMINPUT - index=1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUMINPUT - index=2 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_ENUMINPUT - index=3 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_S_INPUT - input=1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_REQBUFS - type=1 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: v4l_fbuffer_alloc() - using 1800 KB high memory Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYBUF - index=0, type=1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: mmap(V4L) of 0x40816000-0x408f7000 (size=921600) Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QUERYBUF - index=1, type=1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: mmap(V4L) of 0x408f7000-0x409d8000 (size=921600) Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QBUF - type=1, index=0 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_QBUF - type=1, index=1 Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: VIDIOC_STREAMON Oct 10 13:15:16 Aphrodite kernel: DC10plus[0]: set_vfe() - width = 768, height = 576 Oct 10 13:15:16 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x41000000 Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: zoran_poll() - no buffers queued Oct 10 13:15:17 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x41000000 Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: zoran_poll() - no buffers queued Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: zoran_poll() - no buffers queued Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: VIDIOC_STREAMOFF Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: munmap(V4L) Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: munmap(V4L) Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: zoran_close(icewing, pid=[13371]), users(+)=1 Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: v4l_fbuffer_free() Oct 10 13:15:17 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:17 Aphrodite kernel: zoran_irq: astat: 0x40000000, mask: 0x01000000 Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - wake GPIO=0xc9008888 Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: jpeg_codec_sleep() - sleep GPIO=0xc1008888 Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: interrupts received: GIRQ1:60 queue_state=0/0/0/0 Oct 10 13:15:17 Aphrodite kernel: DC10plus[0]: zoran_close() done Frank Loemker wrote, On 10.10.2006 13:13: > Hello, > > On Tue, Oct 10, 2006 at 12:59:51PM +0200, Axel Haasch wrote: >> Ok, I changed the options for the kernel module like this and reloaded it: >> >> zr36067 debug=4 default_norm=0 lock_norm=0 pass_through=1 v4l_bufsize=900 v4l_nbufs=2 >> >> Attached, you'll find the output of >> >>> icewing -sg v debug >> and my /var/log/messages entries. So far I see nothing unusual except the >> lines >> >> UNKNOWN ioctl cmd: 0xc068562 > > That's ok. iceWing first tries to access the grabber via a slightly older V4L2 > interface. If this fails the current programming interface is used. As the > driver complies to the current interface, the old accessing method gives the > error messages. > >> ./icewing -sg v debug >> ... >> v4l2: Using format 4: YUYV Size: 768x576 >> v4l2: Unable to switch of AGC! >> v4l2: Available video inputs: >> 0: Composite >> 1: S-Video >> 2: Internal/comp >> v4l2: Using input 0 > > A short question before I check the complete /var/log/messages log: > Is input 0, i.e. composite, the correct input? If not, try to start iceWing > with the command: > > icewing -sg v debug:input=1 > or > icewing -sg v debug:input=2 > > depending on which input is the correct one. > Will check the log's next. > > Bye, > - -- Int. Conference on Computer Vision Systems 07: Vision Systems in the Real World 21st-24th March, Bielefeld, Germany http://www.icvs2007.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iQEVAwUBRSuCYDuI6ru3QAHZAQrJsQf/eptHajb1/fBMnxJBTBzJihhgx0ip691I WKAu9a5vyBcJ9glKstmSuX98ktbpPD3PxDx4xdJXiPlFEO2xiamUY2L1QvfRbg7S xXTl8mqjvVZRhZtDQzs2tkXPNtnAYpOnlxr+RQo/Fc2EXDT2IRkiW6IwjlbOCS6s 5bW94yvSm6j5IoZ0RKw9odI+sSkaKeIx0ZiDX+408kOepZo9V75bcJWne1HZnQ0M htvLQRuJT0ptpVDU+cCMcWfaT5gKaUf9dx7U28EPTiuFZXX3kcTYAOoIUjPIdKxE VXICdF49U1mYkOgGVOu8sdfLjzzd7foe+mHQfVDx+aaF+g9PlWlNjg== =I7Ey -----END PGP SIGNATURE----- |
|
From: Frank L. <flo...@Te...> - 2006-10-10 11:13:16
|
Hello,
On Tue, Oct 10, 2006 at 12:59:51PM +0200, Axel Haasch wrote:
> Ok, I changed the options for the kernel module like this and reloaded it:
>
> zr36067 debug=4 default_norm=0 lock_norm=0 pass_through=1 v4l_bufsize=900 v4l_nbufs=2
>
> Attached, you'll find the output of
>
> > icewing -sg v debug
>
> and my /var/log/messages entries. So far I see nothing unusual except the
> lines
>
> UNKNOWN ioctl cmd: 0xc068562
That's ok. iceWing first tries to access the grabber via a slightly older V4L2
interface. If this fails the current programming interface is used. As the
driver complies to the current interface, the old accessing method gives the
error messages.
> ./icewing -sg v debug
> ...
> v4l2: Using format 4: YUYV Size: 768x576
> v4l2: Unable to switch of AGC!
> v4l2: Available video inputs:
> 0: Composite
> 1: S-Video
> 2: Internal/comp
> v4l2: Using input 0
A short question before I check the complete /var/log/messages log:
Is input 0, i.e. composite, the correct input? If not, try to start iceWing
with the command:
icewing -sg v debug:input=1
or
icewing -sg v debug:input=2
depending on which input is the correct one.
Will check the log's next.
Bye,
--
Frank Loemker
|
|
From: Axel H. <ah...@Te...> - 2006-10-10 11:00:04
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Hello! Ok, I changed the options for the kernel module like this and reloaded it: zr36067 debug=4 default_norm=0 lock_norm=0 pass_through=1 v4l_bufsize=900 v4l_nbufs=2 Attached, you'll find the output of > icewing -sg v debug and my /var/log/messages entries. So far I see nothing unusual except the lines UNKNOWN ioctl cmd: 0xc068562 Shall I run iceWing in gdb? Cheers, Axel Frank Loemker wrote, On 10.10.2006 09:12: > Hello, > > On Mon, Oct 09, 2006 at 07:12:37PM +0200, Axel Haasch wrote: >> Ok, the 128kB buffer problem seems to be well known, according to >> >> http://www.linuxtv.org/v4lwiki/index.php/Zoran_devices_(zr36057,_zr36067) >> >> So I added the kernel option 'mem=496M' to my grub bootloader and restarted, >> as I also have 512MB. Furthermore, I set the zr36067-parameters to > > That's because of (from zoran_driver.c::v4l_fbuffer_alloc()): > -------------------------------------------------------------------------------- > if (fh->v4l_buffers.buffer_size <= MAX_KMALLOC_MEM) { > ... > } else { > #if defined(CONFIG_BIGPHYS_AREA) > /* Use bigphysarea_alloc_pages */ > ... > #elif defined(BUZ_USE_HIMEM) > /* Use high memory which has been left at boot time */ > ... > } > -------------------------------------------------------------------------------- > with > zoran.h:#define MAX_KMALLOC_MEM (128*1024) > > i.e. if more than 128k should be allocated the driver tries to use high memory > for the buffers. And that failed previously because of insufficient high > memory: > > Oct 9 18:35:28 Aphrodite kernel: ZORAN: get_high_mem() - requested 5308416, avail 16384 > > Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: v4l_fbuffer_alloc() - get_high_mem (size = 5184 KB) for V4L bufs failed > > Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: v4l2_buffer_status() - wrong number or buffers not allocated > > >> v4l_bufsize=512 v4l_nbufs=32 > > Try to use > > debug=4 v4l_bufsize=900 v4l_nbufs=2 > > to get > - the maximum debug output > (a lot of debugging messages are missing in your /var/log/messages output) > - enough memory for at least one image of size 768 x 576 and 2 bpp (864k) > - the least memory usage which should still provide a sensitive working > environment (2 buffers). 4 buffers might be better in the long run, > 32 are definitely a little bit overkill. > > and provide the complete stderr/stdout output as well as the complete output > from /var/log/messages (from the start of the loading of the zoran module > until the last (there are AT LEAST two) zoran_close from iceWing) from > > icewing -sg v debug > >> The memalloc error is gone, as you can see below, but iceWing still exits the >> same way like before. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iQEVAwUBRSt9JzuI6ru3QAHZAQoQbgf+Jhba9F01Ms+sapOZbjhU3fMky+90r1WF 80YPCaQmiY/ztIHr3YmEAL7CmdKO3Xb9c1s9uk2fJq7qqA+fXuefVOLx/Tvgzlwt IFGFh4fjT1sy4Up7fc4MkLRbMQWF0ua0S9F9Gmh3uIHSsynvwPqsH+Eu44Q3TZsq 2hchxV6x21/puZ+YRGCNbtjPSDWJWCB6deIUP4K6FILM9eZFSjutjNbW9JjfDrSG GhKLBgShsE7XkmfDVFAdc50OcP+6EXo/HGU1T4XLjzKofF/SsguxK6rAdSNH57Fn 9IaCbAETldDVMZ3N3zqxkTPobqFPjEk2i1CVTthDe/KrZTSvNxcLYA== =JKGA -----END PGP SIGNATURE----- |
|
From: Frank L. <flo...@Te...> - 2006-10-10 07:12:36
|
Hello, On Mon, Oct 09, 2006 at 07:12:37PM +0200, Axel Haasch wrote: > Ok, the 128kB buffer problem seems to be well known, according to > > http://www.linuxtv.org/v4lwiki/index.php/Zoran_devices_(zr36057,_zr36067) > > So I added the kernel option 'mem=496M' to my grub bootloader and restarted, > as I also have 512MB. Furthermore, I set the zr36067-parameters to That's because of (from zoran_driver.c::v4l_fbuffer_alloc()): -------------------------------------------------------------------------------- if (fh->v4l_buffers.buffer_size <= MAX_KMALLOC_MEM) { ... } else { #if defined(CONFIG_BIGPHYS_AREA) /* Use bigphysarea_alloc_pages */ ... #elif defined(BUZ_USE_HIMEM) /* Use high memory which has been left at boot time */ ... } -------------------------------------------------------------------------------- with zoran.h:#define MAX_KMALLOC_MEM (128*1024) i.e. if more than 128k should be allocated the driver tries to use high memory for the buffers. And that failed previously because of insufficient high memory: > Oct 9 18:35:28 Aphrodite kernel: ZORAN: get_high_mem() - requested 5308416, avail 16384 > Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: v4l_fbuffer_alloc() - get_high_mem (size = 5184 KB) for V4L bufs failed > Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: v4l2_buffer_status() - wrong number or buffers not allocated > v4l_bufsize=512 v4l_nbufs=32 Try to use debug=4 v4l_bufsize=900 v4l_nbufs=2 to get - the maximum debug output (a lot of debugging messages are missing in your /var/log/messages output) - enough memory for at least one image of size 768 x 576 and 2 bpp (864k) - the least memory usage which should still provide a sensitive working environment (2 buffers). 4 buffers might be better in the long run, 32 are definitely a little bit overkill. and provide the complete stderr/stdout output as well as the complete output from /var/log/messages (from the start of the loading of the zoran module until the last (there are AT LEAST two) zoran_close from iceWing) from icewing -sg v debug > The memalloc error is gone, as you can see below, but iceWing still exits the > same way like before. > > Seems to be a neverending story. :-( Don't think so, up to now we haven't tried a lot ;) Bye, -- Frank Loemker |
|
From: Axel H. <ah...@Te...> - 2006-10-09 17:12:53
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Hello! Ok, the 128kB buffer problem seems to be well known, according to http://www.linuxtv.org/v4lwiki/index.php/Zoran_devices_(zr36057,_zr36067) So I added the kernel option 'mem=496M' to my grub bootloader and restarted, as I also have 512MB. Furthermore, I set the zr36067-parameters to v4l_bufsize=512 v4l_nbufs=32 The memalloc error is gone, as you can see below, but iceWing still exits the same way like before. Seems to be a neverending story. :-( Oct 9 19:04:28 Aphrodite kernel: DC10plus[0]: i2c_client_unregister() Oct 9 19:04:28 Aphrodite kernel: MJPEG[0]: i2c_client_unregister() Oct 9 19:04:28 Aphrodite kernel: ACPI: PCI interrupt for device 0000:00:0d.0 disabled Oct 9 19:04:31 Aphrodite kernel: Zoran MJPEG board driver version 0.9.5 Oct 9 19:04:31 Aphrodite kernel: ACPI: PCI Interrupt 0000:00:0d.0[A] -> Link [LNKA] -> GSI 11 (level, low) -> IRQ 11 Oct 9 19:04:31 Aphrodite kernel: MJPEG[0]: Zoran ZR36067 (rev 2) irq: 11, memory: 0xad000000 Oct 9 19:04:31 Aphrodite kernel: MJPEG[0]: subsystem vendor=0x1031 id=0x7efe Oct 9 19:04:31 Aphrodite kernel: DC10plus[0]: Initializing i2c bus... Oct 9 19:04:31 Aphrodite kernel: DC10plus[0]: i2c_client_register() - driver id = 22 Oct 9 19:04:31 Aphrodite kernel: DC10plus[0]: i2c_client_register() - driver id = 48 Oct 9 19:04:31 Aphrodite kernel: DC10plus[0]: Initializing videocodec bus... Oct 9 19:04:31 Aphrodite kernel: ZORAN: 1 card(s) found Oct 9 19:04:31 Aphrodite kernel: DC10plus[0]: zr36057_init() - initializing card[0], zr=e03e6940 Oct 9 19:04:31 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 19:04:32 Aphrodite kernel: DC10plus[0]: interrupts received: GIRQ1:50 queue_state=0/0/0/0 Oct 9 19:04:32 Aphrodite kernel: DC10plus[0]: procfs entry /proc/zoran0 allocated. data=e03e6940 Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[6880]), users(-)=0 Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: UNKNOWN ioctl cmd: 0xc0385602 Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: zoran_close(icewing, pid=[6880]), users(+)=1 Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: interrupts received: GIRQ1:7 queue_state=0/0/0/0 Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[6880]), users(-)=0 Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 19:04:42 Aphrodite kernel: DC10plus[0]: set_vfe() - width = 256, height = 192 Cheers, Axel Frank Loemker wrote, On 09.10.2006 17:05: > Hello, > > On Mon, Oct 09, 2006 at 01:36:48PM +0200, Axel Haasch wrote: >> I am trying to get iceWing cooperating with my Pinnacle Studio >> DC10plus video capture board, but unfortuanetly I wasn't successful so far. >> Maybe you've got an idea? >> >> ... >> >> That wasn't successful either. So, what can I do next? > > Are there any messages in the file /var/log/messages concerning the capture > board? > > It might be necessary to increase the verbosity level of the syslog > system. If you have syslog-ng, something like > > destination messages { file("/var/log/messages"); }; > log { source(src); destination(messages); }; > > in the file > > /etc/syslog-ng/syslog-ng.conf > > might work (or something similar). Don't forget to restart the / send > a HUP to the syslog daemon. > >> Does it make sense to >> fetch a recent devel-version from the iceWing repository? > > To fix the problem -- no, nothing concerning the grabber interaction has > changed there. To see the latest and greatest features and help enhancing > iceWing -- yes of course :) > > Bye, > - -- Int. Conference on Computer Vision Systems 07: Vision Systems in the Real World 21st-24th March, Bielefeld, Germany http://www.icvs2007.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iQEVAwUBRSqDBTuI6ru3QAHZAQphBgf/foj9lzdCI0d1UCJCVqa4TUS/QkmgEGqr lFunn1PHrwpkdFa7CkqwYW8T//xRDjjFVV7Iv7TSWe7/RxpuS5CvdDOgoM5coXMR qYACmqBDY7l22oao3hCoewCxoTg+yPlzCMJ9Bs7SjRIjjkKFkzPmbXEcrH25j1Te fBVTY+gAFrBl7g3p/aF/E760HXaUdeLx7RfA5v2LCV5azrGasn7X1Wukfa8Xtx6A O7hdDMLtCz5WkKMRUIBfuO6pMkUuaCJNd8C1GsFrcnGEScx7TJiRrDFGt+yWQHmS tUi6oiVqhRd1cEXicYIpddzlanqB6uiowVfwrGiOtzHqz9T2mOzFnA== =TA+g -----END PGP SIGNATURE----- |
|
From: Axel H. <ah...@Te...> - 2006-10-09 16:43:45
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Hi! Ok, as soon as I added the params to my '/etc/modprobe.conf.local' it worked, but for 'v4l_bufsize=700' it is also reported that the size is too small, even with a downsampling of 3! Hence, according to http://usuario.cicese.mx/~mirsev/Linux/DC10plus/README I increased the bufsize value first to 1728 and then to 1296. In both cases iceWing reported: v4l2: QUERYBUF returned error (22: Invalid argument)! I also increased the debug level of the kernel module. Here's the output of the higher values. Oct 9 18:35:27 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[22101]), users(-)=0 Oct 9 18:35:27 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 18:35:27 Aphrodite kernel: DC10plus[0]: UNKNOWN ioctl cmd: 0xc0385602 Oct 9 18:35:27 Aphrodite kernel: DC10plus[0]: zoran_close(icewing, pid=[22101]), users(+)=1 Oct 9 18:35:27 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 18:35:27 Aphrodite kernel: DC10plus[0]: interrupts received: GIRQ1:6 queue_state=0/0/0/0 Oct 9 18:35:27 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[22101]), users(-)=0 Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 18:35:28 Aphrodite kernel: ZORAN: get_high_mem() - requested 5308416, avail 16384 Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: v4l_fbuffer_alloc() - get_high_mem (size = 5184 KB) for V4L bufs failed Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: v4l2_buffer_status() - wrong number or buffers not allocated Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: zoran_close(icewing, pid=[22101]), users(+)=1 Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: enable_jpg(IDLE) Oct 9 18:35:28 Aphrodite kernel: DC10plus[0]: interrupts received: GIRQ1:10 queue_state=0/0/0/0 If I understood the webpage described above, I have to restart the kernel with an adapted mem-Entry. Is that correct? Could that be the problem? I also tried to increase the number of buffers (v4l_nbufs=4). That didn't work either. Regards, Axel Frank Loemker wrote, On 09.10.2006 17:37: > Hello, > > On Mon, Oct 09, 2006 at 05:05:51PM +0200, Frank Loemker wrote: >> Hello, >> >> On Mon, Oct 09, 2006 at 01:36:48PM +0200, Axel Haasch wrote: >>> I am trying to get iceWing cooperating with my Pinnacle Studio >>> DC10plus video capture board, but unfortuanetly I wasn't successful so far. >>> Maybe you've got an idea? >>> >>> ... >>> >>> That wasn't successful either. So, what can I do next? >> Are there any messages in the file /var/log/messages concerning the capture >> board? >> >> It might be necessary to increase the verbosity level of the syslog >> system. If you have syslog-ng, something like >> >> destination messages { file("/var/log/messages"); }; >> log { source(src); destination(messages); }; >> >> in the file >> >> /etc/syslog-ng/syslog-ng.conf >> >> might work (or something similar). Don't forget to restart the / send >> a HUP to the syslog daemon. >> >>> Does it make sense to >>> fetch a recent devel-version from the iceWing repository? >> To fix the problem -- no, nothing concerning the grabber interaction has >> changed there. To see the latest and greatest features and help enhancing >> iceWing -- yes of course :) > > And a second possibility / try: > > What happens if you try to load the zoran kernel module with the > option "v4l_bufsize=700"? It seems, that the default buffer size is > set to 128k, which, if this is really right, is quite small. > > Or, alternatively, set the Downsampling in iceWing to 3 BEFORE iceWing > accesses the grabber for example by starting iceWing with > -rc "Other.Downsampling=3" > > Bye, > - -- Int. Conference on Computer Vision Systems 07: Vision Systems in the Real World 21st-24th March, Bielefeld, Germany http://www.icvs2007.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iQEVAwUBRSp8LTuI6ru3QAHZAQoTtwf/YH7uNAd1W3Itr9FEBDLIMUBZ+SEDRTnU 0syMZhp4R6WF/xky1hkv8nzLrln3W9ITS9VyST8vVtd76qOfoqOVElOhZYBKvywq b5oEcZ/QnO9XGkRt3s93D9BIrXO7LqY4nc9iZe3VI5t3OEB0MIK7gsaza0cL2e9k wMVKrqCI5J9Msz+XeYWRdMe9YOwhxcE5OYkeNxRYb+It2t2XxlrIJqDXlKd8S9Dp h+Lh/qE3e45aj7aXbBlaF6P9fhvfih58q0QJPQ01Zx4rD+Qs27SIkqR89v1HOqX3 aCyNoclvWm+t25GHTSgT84RS9CN0Rkxi1Bswrz0T422X6uXxEHLwOg== =R1rH -----END PGP SIGNATURE----- |
|
From: Axel H. <ah...@Te...> - 2006-10-09 16:16:18
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Hi! Sorry, I just saw in '/var/log/messages' that the parameter 'v4l_bufsize' doesn't seem to work although there was no error message while loading the kernel module. First, I did an 'rmmod zr36067' twice, so I am pretty sure, that the module was removed. Then, I loaded it again with modprobe zr36067 default_norm=0 lock_norm=0 pass_through=1 v4l_bufsize=700 But in the log, still the message v4l_set_format() - video buffer size (128 kB) is too small occurs, when I start iceWing. I'll try get more information on this, before you begin to despair. ;-) Cheers, Axel -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iQEVAwUBRSp1wDuI6ru3QAHZAQrBRAf+MuLwSP/weeEjOVEl6Y0+lzAMRysCisU2 pkv8VaWPlkpGihCIIhKaQaeFvh4vdCuNtkFhT47WilWAEa+slaX7WjU134loXvlT k9b2ewvyrTu6Nl8bUvIDSlRso/05YNs3XJ09bNS/CdS0Ibf84j3CKeMzOUMEKj/P lzXMc/m1aCKW0diBsOjRQiH8i5RRZZ0UjKgGzeBCFGKOyD1TsBj2+IY0HS8Quncm TSCgiqrC0JPAvz1Yv1peL3MS1kmRr8ZERWiLcPfHGPg7l79vtKMWmSytyD31SAFB CQS9QeiRaKab+B+tncq8LCRpL3QkzNMyHRWQEidFQ+GL1b43X7Bf2Q== =q/qH -----END PGP SIGNATURE----- |
|
From: Axel H. <ah...@Te...> - 2006-10-09 15:59:27
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
Hi!
Ok, I tried several combination with the v4l2 and unicap driver. Nothing worked.
Here's the most recent output:
=========
./icewing -sg -t 5 -rc "Other.Downsampling=3"
v4l2_old: No formats available!
v4l2: Unable to switch of AGC!
> grab.c (1261): Getting image...
select timeout
AVGetLastImg: Grabbing frame failed!
iceWing, Error in main/grab.c at 428: Can't grab image from input signal 1 !
=========
Without talklevel 5, the output is the same as before.
A couple of new lines are added to '/var/log/messages' as well. All like:
Oct 9 17:50:06 Aphrodite kernel: DC10plus[0]: zoran_poll() - no buffers queued
The camera is still available via 'xawtv'. So, at least the restart of the kernel module
with the increased buffer size has worked. :-)
Cheers,
Axel
Frank Loemker wrote, On 09.10.2006 17:37:
> Hello,
>
> On Mon, Oct 09, 2006 at 05:05:51PM +0200, Frank Loemker wrote:
>> Hello,
>>
>> On Mon, Oct 09, 2006 at 01:36:48PM +0200, Axel Haasch wrote:
>>> I am trying to get iceWing cooperating with my Pinnacle Studio
>>> DC10plus video capture board, but unfortuanetly I wasn't successful so far.
>>> Maybe you've got an idea?
>>>
>>> ...
>>>
>>> That wasn't successful either. So, what can I do next?
>> Are there any messages in the file /var/log/messages concerning the capture
>> board?
>>
>> It might be necessary to increase the verbosity level of the syslog
>> system. If you have syslog-ng, something like
>>
>> destination messages { file("/var/log/messages"); };
>> log { source(src); destination(messages); };
>>
>> in the file
>>
>> /etc/syslog-ng/syslog-ng.conf
>>
>> might work (or something similar). Don't forget to restart the / send
>> a HUP to the syslog daemon.
>>
>>> Does it make sense to
>>> fetch a recent devel-version from the iceWing repository?
>> To fix the problem -- no, nothing concerning the grabber interaction has
>> changed there. To see the latest and greatest features and help enhancing
>> iceWing -- yes of course :)
>
> And a second possibility / try:
>
> What happens if you try to load the zoran kernel module with the
> option "v4l_bufsize=700"? It seems, that the default buffer size is
> set to 128k, which, if this is really right, is quite small.
>
> Or, alternatively, set the Downsampling in iceWing to 3 BEFORE iceWing
> accesses the grabber for example by starting iceWing with
> -rc "Other.Downsampling=3"
>
> Bye,
>
- --
Int. Conference on Computer Vision Systems 07: Vision Systems in the Real World
21st-24th March, Bielefeld, Germany http://www.icvs2007.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
iQEVAwUBRSpx0TuI6ru3QAHZAQp32Qf/YKHyGzAe2x7jgJ9GWUdfNqSPUlFj1obR
QrHHFIKY5CShxNkytc4k/egqLUA4VVL5zEclA5EeGbP3XEgChkX9GvUX1HeLQuuP
jkTW862JxuwA/vIHTtlVWM6BzB5iATdBvSv+TuhRrOfpbfypAGGJrhizlky8Vv03
Qn4o2nfW133gyhGl2o39HAXql+AR2v3M7oW7ta/jIqGGYlOFlH956aglKM5SfPRG
JEJzNFjfEIa8VxVlts72gMKAM4zJ6zUrTSRJxvQg1ljWHI81HLJqLDNaPEQDZoYX
CfP8+FCAuhIB0BfuJLYx9gubQp5vvBSJGQCgCN55dhThbM7NwhnoXQ==
=9yxZ
-----END PGP SIGNATURE-----
|
|
From: Axel H. <ah...@Te...> - 2006-10-09 15:44:41
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
Hi!
Thank you very much for the quick response.
Frank, you're UNBELIEVABLE! :-)
I'll try immediately to increase the buffer size, as it seems indeed to small,
as the syslog showed me. The increasing of the log level hasn't provide more
useful information. Here's the fragment.
Oct 9 17:35:04 Aphrodite kernel: Kernel logging (proc) stopped.
Oct 9 17:35:04 Aphrodite kernel: Kernel log daemon terminating.
Oct 9 17:35:05 Aphrodite syslog-ng[19692]: syslog-ng version 1.6.8 going down
Oct 9 17:35:05 Aphrodite syslog-ng[19813]: syslog-ng version 1.6.8 starting
Oct 9 17:35:10 Aphrodite kernel: klogd 1.4.1, log source = /proc/kmsg started.
Oct 9 17:35:12 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[19830]), users(-)=0
Oct 9 17:35:12 Aphrodite kernel: DC10plus[0]: UNKNOWN ioctl cmd: 0xc0385602
Oct 9 17:35:12 Aphrodite kernel: DC10plus[0]: zoran_close(icewing, pid=[19830]), users(+)=1
Oct 9 17:35:12 Aphrodite kernel: DC10plus[0]: zoran_open(icewing, pid=[19830]), users(-)=0
Oct 9 17:35:12 Aphrodite syslog-ng[19813]: Changing permissions on special file /dev/xconsole
Oct 9 17:35:12 Aphrodite syslog-ng[19813]: Changing permissions on special file /dev/tty10
Oct 9 17:35:12 Aphrodite kernel: DC10plus[0]: v4l_set_format() - video buffer size (128 kB) is too small
Oct 9 17:35:12 Aphrodite kernel: DC10plus[0]: zoran_close(icewing, pid=[19830]), users(+)=1
Both doesn't seem to work, neither the increased buffer size nor the downsampling. I'll provide
more infos in a few minutes.
Cheers,
Axel
Frank Loemker wrote, On 09.10.2006 17:05:
> Hello,
>
> On Mon, Oct 09, 2006 at 01:36:48PM +0200, Axel Haasch wrote:
>> I am trying to get iceWing cooperating with my Pinnacle Studio
>> DC10plus video capture board, but unfortuanetly I wasn't successful so far.
>> Maybe you've got an idea?
>>
>> ...
>>
>> That wasn't successful either. So, what can I do next?
>
> Are there any messages in the file /var/log/messages concerning the capture
> board?
>
> It might be necessary to increase the verbosity level of the syslog
> system. If you have syslog-ng, something like
>
> destination messages { file("/var/log/messages"); };
> log { source(src); destination(messages); };
>
> in the file
>
> /etc/syslog-ng/syslog-ng.conf
>
> might work (or something similar). Don't forget to restart the / send
> a HUP to the syslog daemon.
>
>> Does it make sense to
>> fetch a recent devel-version from the iceWing repository?
>
> To fix the problem -- no, nothing concerning the grabber interaction has
> changed there. To see the latest and greatest features and help enhancing
> iceWing -- yes of course :)
>
> Bye,
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
iQEVAwUBRSpuXzuI6ru3QAHZAQpo9AgAjDWVkXu/PWvvysop71wqk1KPVere0WX2
QebcZ5LZiBCyDXun7FaDFGr64jKmy+uBNaT5+lnZrFCHCAL0eANMWp4/hm7tB4nh
m1zX2HSd3QgnQH8utKu0kkJyKkhuvmUd9hiP3mkU/Gvk/DXu40WfWm5nF/5CsT+i
Bl7WmsLbA4/TS+pPEAol2X6dqmLNOH/AO8SYvR3W9zExg88aOTAffvzi9WFsvdPJ
vkO6xyH63NvuvTL1QlOY4Q9BM/6D/RY74erozX3oO1dK1m1E/3MM1rVlhXpeAr/K
LU85Eteya+bjBsKh3FrJO0DUm0gDLwig/AdacfVPoVfOmX+KfwZaaQ==
=MzX9
-----END PGP SIGNATURE-----
|
|
From: Frank L. <flo...@Te...> - 2006-10-09 15:37:24
|
Hello,
On Mon, Oct 09, 2006 at 05:05:51PM +0200, Frank Loemker wrote:
> Hello,
>
> On Mon, Oct 09, 2006 at 01:36:48PM +0200, Axel Haasch wrote:
> > I am trying to get iceWing cooperating with my Pinnacle Studio
> > DC10plus video capture board, but unfortuanetly I wasn't successful so far.
> > Maybe you've got an idea?
> >
> >...
> >
> > That wasn't successful either. So, what can I do next?
>
> Are there any messages in the file /var/log/messages concerning the capture
> board?
>
> It might be necessary to increase the verbosity level of the syslog
> system. If you have syslog-ng, something like
>
> destination messages { file("/var/log/messages"); };
> log { source(src); destination(messages); };
>
> in the file
>
> /etc/syslog-ng/syslog-ng.conf
>
> might work (or something similar). Don't forget to restart the / send
> a HUP to the syslog daemon.
>
> > Does it make sense to
> > fetch a recent devel-version from the iceWing repository?
>
> To fix the problem -- no, nothing concerning the grabber interaction has
> changed there. To see the latest and greatest features and help enhancing
> iceWing -- yes of course :)
And a second possibility / try:
What happens if you try to load the zoran kernel module with the
option "v4l_bufsize=700"? It seems, that the default buffer size is
set to 128k, which, if this is really right, is quite small.
Or, alternatively, set the Downsampling in iceWing to 3 BEFORE iceWing
accesses the grabber for example by starting iceWing with
-rc "Other.Downsampling=3"
Bye,
--
Frank Loemker
|
|
From: Frank L. <flo...@Te...> - 2006-10-09 15:06:03
|
Hello,
On Mon, Oct 09, 2006 at 01:36:48PM +0200, Axel Haasch wrote:
> I am trying to get iceWing cooperating with my Pinnacle Studio
> DC10plus video capture board, but unfortuanetly I wasn't successful so far.
> Maybe you've got an idea?
>
>...
>
> That wasn't successful either. So, what can I do next?
Are there any messages in the file /var/log/messages concerning the capture
board?
It might be necessary to increase the verbosity level of the syslog
system. If you have syslog-ng, something like
destination messages { file("/var/log/messages"); };
log { source(src); destination(messages); };
in the file
/etc/syslog-ng/syslog-ng.conf
might work (or something similar). Don't forget to restart the / send
a HUP to the syslog daemon.
> Does it make sense to
> fetch a recent devel-version from the iceWing repository?
To fix the problem -- no, nothing concerning the grabber interaction has
changed there. To see the latest and greatest features and help enhancing
iceWing -- yes of course :)
Bye,
--
Frank Loemker
|
|
From: Axel H. <ah...@Te...> - 2006-10-09 11:37:04
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Dear Frank, dear colleagues, I am trying to get iceWing cooperating with my Pinnacle Studio DC10plus video capture board, but unfortuanetly I wasn't successful so far. Maybe you've got an idea? Here are my specs: - - SuSE 10.0 - - 2.6.13-15.12-default - - Kernel module: zr36067 default_norm=0 lock_norm=0 pass_through=1 - - iceWing V0.9 (May 30 2006) - - ffmpeg-20060526 - - unicap-0.1.16 I tried two version of iceWing. First, the installed version in university: /vol/vita/bin/icewing and second a self-compiled version of the latest stable release provided on the sf.net pages. Neither worked, but my video card works fine with 'xawtv'. On the other hand it does not work with 'xawtv4' or 'kwintv'. I haven't tried their parameters so far. My video-device '/dev/video0' links to '/dev/video'. Here's the xawtv output: > /usr/X11R6/bin/xawtv -hwscan > This is xawtv-3.94, running on Linux/i686 (2.6.13-15.12-default) > looking for available devices > port 69-69 [ -xvport 69 ] > type : Xvideo, video overlay > name : video4linux > > port 70-70 > type : Xvideo, image scaler > name : ATI Radeon Video Overlay > > /dev/video0: OK [ -device /dev/video0 ] > type : v4l2 > name : DC10plus[0] > flags: overlay capture I also attached a fine xawtv run with the parameter '-debug 2'. Maybe that'll help. What I did so far: ================== > /usr/X11R6/bin/v4l-conf > v4l-conf: using X11 display :0.0 > dga: version 2.0 > mode: 1280x1024, depth=24, bpp=32, bpl=5120, base=0xd0000000 > /dev/video0 [v4l2]: configuration done First I tried to use the default v4l2 driver '-sg' which reports the message 'v4l2: S_FMT returned error (22: Invalid argument)!' Then I enabled the debug mode. > ./icewing -sg V4L2 debug > v4l2_old: Available properties: > v4l2_old: Available formats: > v4l2_old: No formats available! > v4l2: Available properties: > 0: Brightness [0 - 1] 0.50 > 1: Contrast [0 - 1] 0.50 > 2: Saturation [0 - 1] 0.50 > 3: Hue [0 - 1] 0.50 > v4l2: Available formats: > 0: 15-bit RGB > 1: 16-bit RGB > 2: 24-bit RGB > 3: 32-bit RGB > 4: 4:2:2, packed, YUYV > 5: Hardware-encoded Motion-JPEG > v4l2: S_FMT returned error (22: Invalid argument)! > AVOpen: Capture init failed! > iceWing, Error in main/grab.c at 355: Unable to init AV-driver for input signal 2 ! Well, that doesn't seem to work. Nevertheless, I tried all 5 format options, the same way I did it for the unicap driver (e.g., ./icewing -sg unicap device=0:format=5) > ./icewing -sg unicap debug > unicap: Available devices: > 0: DC10plus[0] (PCI:0000:00:0d.0) [/dev/video0] > unicap: Using device 0: DC10plus[0] (PCI:0000:00:0d.0) [/dev/video0] > unicap: Available properties: > 0: Brightness [0 - 65535] 32768 > 1: Contrast [0 - 65535] 32768 > 2: Saturation [0 - 65535] 32768 > 3: Hue [0 - 65535] 32768 > 4: video source [Composite | *S-Video | Internal/comp] > 5: video norm [*PAL | NTSC | SECAM | Autodetect] > unicap: Available formats: > 0: 15-bit RGB [640x480] > 1: 16-bit RGB [640x480] > 2: 24-bit RGB [640x480] > 3: 32-bit RGB [640x480] > 4: 4:2:2, packed, YUYV [640x480] > 5: Hardware-encoded Motion-JPEG [640x480] > unicap: Failed to set video format 0 (22: Invalid argument)! > unicap: Failed to stop capture on device: DC10plus[0] (PCI:0000:00:0d.0)! > AVOpen: Capture init failed! > iceWing, Error in main/grab.c at 355: Unable to init AV-driver for input signal 4 ! That wasn't successful either. So, what can I do next? Does it make sense to fetch a recent devel-version from the iceWing repository? I also added an 'strace -f' run as attachment of a faulty 'icewing -sg' start. Last but not least here's the 'lspci' output, maybe it's interesting for the chipset used on the video board: 00:0d.0 Multimedia video controller: Zoran Corporation ZR36057PQC Video cutting chipset (rev 02) Subsystem: Miro Computer Products AG DC10 Plus Flags: fast devsel, IRQ 11 Memory at ad000000 (32-bit, non-prefetchable) [size=4K] Regards, Axel -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.2 (GNU/Linux) iQEVAwUBRSo0UDuI6ru3QAHZAQrW/ggAnf8lHxsjWZwkYSolktie4nDsMZrLP3uP oWLFoXkJkySleFCHL+/xhjUMHjx2pB99V7Qr8ulD24Cy4pj1BBnfxHLL8aLtB/2t DjPNOHGwN6lxBzL6W39DEoUYHtfYTcbRkATb5rWFJtmEIK/x2GHpb8tolWltLcEX pgvx2X5w1S0+9KKxO4IkfKzyLrDFE5oiQw2EJBD+V4X3weS2Chb69UYy56GOOXGY QsJuRvK4vsp+fWRjzEcKMKT09jEJYZi90oqLRcTIJIKLYFTgtL8Y1/JHD290Q74A LwAw8hIUkOZmwEkYI4xqlR0uzmP25KMi+ebU/Vf0vtJN+FW4RNmInw== =d/qh -----END PGP SIGNATURE----- |
