The #GstVideoOverlay interface is used for 2 main purposes :
To get a grab on the Window where the video sink element is going to render.
This is achieved by either being informed about the Window identifier that
the video sink element generated, or by forcing the video sink element to use
a specific Window identifier for rendering.
To force a redrawing of the latest video frame the video sink element
displayed on the Window. Indeed if the #GstPipeline is in #GST_STATE_PAUSED
state, moving the Window around will damage its content. Application
developers will want to handle the Expose events themselves and force the
video sink element to refresh the Window's content.
Using the Window created by the video sink is probably the simplest scenario,
in some cases, though, it might not be flexible enough for application
developers if they need to catch events such as mouse moves and button
clicks.
Setting a specific Window identifier on the video sink element is the most
flexible solution but it has some issues. Indeed the application needs to set
its Window identifier at the right time to avoid internal Window creation
from the video sink element. To solve this issue a #GstMessage is posted on
the bus to inform the application that it should set the Window identifier
immediately. Here is an example on how to do that correctly:
There are two basic usage scenarios: in the simplest case, the application
uses #playbin or #playsink or knows exactly what particular element is used
for video output, which is usually the case when the application creates
the videosink to use (e.g. #xvimagesink, #ximagesink, etc.) itself; in this
case, the application can just create the videosink element, create and
realize the window to render the video on and then
call gstvideo.video_overlay.VideoOverlay.setWindowHandle directly with the XID or native
window handle, before starting up the pipeline.
As #playbin and #playsink implement the video overlay interface and proxy
it transparently to the actual video sink even if it is created later, this
case also applies when using these elements.
In the other and more common case, the application does not know in advance
what GStreamer video sink element will be used for video output. This is
usually the case when an element such as #autovideosink is used.
In this case, the video sink element itself is created
asynchronously from a GStreamer streaming thread some time after the
pipeline has been started up. When that happens, however, the video sink
will need to know right then whether to render onto an already existing
application window or whether to create its own window. This is when it
posts a prepare-window-handle message, and that is also why this message needs
to be handled in a sync bus handler which will be called from the streaming
thread directly (because the video sink will need an answer right then).
As response to the prepare-window-handle element message in the bus sync
handler, the application may use gstvideo.video_overlay.VideoOverlay.setWindowHandle to tell
the video sink to render onto an existing window surface. At this point the
application should already have obtained the window handle / XID, so it
just needs to set it. It is generally not advisable to call any GUI toolkit
functions or window system functions from the streaming thread in which the
prepare-window-handle message is handled, because most GUI toolkits and
windowing systems are not thread-safe at all and a lot of care would be
required to co-ordinate the toolkit and window system calls of the
different threads (Gtk+ users please note: prior to Gtk+ 2.18
GDK_WINDOW_XID was just a simple structure access, so generally fine to do
within the bus sync handler; this macro was changed to a function call in
Gtk+ 2.18 and later, which is likely to cause problems when called from a
sync handler; see below for a better approach without GDK_WINDOW_XID
used in the callback).
#include <gst/video/videooverlay.h>
#include <gtk/gtk.h>
#ifdef GDK_WINDOWING_X11
#include <gdk/gdkx.h> // for GDK_WINDOW_XID
#endif
#ifdef GDK_WINDOWING_WIN32
#include <gdk/gdkwin32.h> // for GDK_WINDOW_HWND
#endif
...
static guintptr video_window_handle = 0;
...
static GstBusSyncReply
bus_sync_handler (GstBus * bus, GstMessage * message, gpointer user_data)
{
// ignore anything but 'prepare-window-handle' element messages
if (!gst_is_video_overlay_prepare_window_handle_message (message))
return GST_BUS_PASS;
if (video_window_handle != 0) {
GstVideoOverlay *overlay;
// GST_MESSAGE_SRC (message) will be the video sink element
overlay = GST_VIDEO_OVERLAY (GST_MESSAGE_SRC (message));
gst_video_overlay_set_window_handle (overlay, video_window_handle);
} else {
g_warning ("Should have obtained video_window_handle by now!");
}
gst_message_unref (message);
return GST_BUS_DROP;
}
...
static void
video_widget_realize_cb (GtkWidget * widget, gpointer data)
{
#if GTK_CHECK_VERSION(2,18,0)
// Tell Gtk+/Gdk to create a native window for this widget instead of
// drawing onto the parent widget.
// This is here just for pedagogical purposes, GDK_WINDOW_XID will call
// it as well in newer Gtk versions
if (!gdk_window_ensure_native (widget->window))
g_error ("Couldn't create native window needed for GstVideoOverlay!");
#endif
#ifdef GDK_WINDOWING_X11
{
gulong xid = GDK_WINDOW_XID (gtk_widget_get_window (video_window));
video_window_handle = xid;
}
#endif
#ifdef GDK_WINDOWING_WIN32
{
HWND wnd = GDK_WINDOW_HWND (gtk_widget_get_window (video_window));
video_window_handle = (guintptr) wnd;
}
#endif
}
...
int
main (int argc, char **argv)
{
GtkWidget *video_window;
GtkWidget *app_window;
...
app_window = gtk_window_new (GTK_WINDOW_TOPLEVEL);
...
video_window = gtk_drawing_area_new ();
g_signal_connect (video_window, "realize",
G_CALLBACK (video_widget_realize_cb), NULL);
gtk_widget_set_double_buffered (video_window, FALSE);
...
// usually the video_window will not be directly embedded into the
// application window like this, but there will be many other widgets
// and the video window will be embedded in one of them instead
gtk_container_add (GTK_CONTAINER (ap_window), video_window);
...
// show the GUI
gtk_widget_show_all (app_window);
// realize window now so that the video window gets created and we can
// obtain its XID/HWND before the pipeline is started up and the videosink
// asks for the XID/HWND of the window to render onto
gtk_widget_realize (video_window);
// we should have the XID/HWND now
g_assert (video_window_handle != 0);
...
// set up sync handler for setting the xid once the pipeline is started
bus = gst_pipeline_get_bus (GST_PIPELINE (pipeline));
gst_bus_set_sync_handler (bus, (GstBusSyncHandler) bus_sync_handler, NULL,
NULL);
gst_object_unref (bus);
...
gst_element_set_state (pipeline, GST_STATE_PLAYING);
...
}
The #GstVideoOverlay interface is used for 2 main purposes :
Using the Window created by the video sink is probably the simplest scenario, in some cases, though, it might not be flexible enough for application developers if they need to catch events such as mouse moves and button clicks.
Setting a specific Window identifier on the video sink element is the most flexible solution but it has some issues. Indeed the application needs to set its Window identifier at the right time to avoid internal Window creation from the video sink element. To solve this issue a #GstMessage is posted on the bus to inform the application that it should set the Window identifier immediately. Here is an example on how to do that correctly:
Two basic usage scenarios
There are two basic usage scenarios: in the simplest case, the application uses #playbin or #playsink or knows exactly what particular element is used for video output, which is usually the case when the application creates the videosink to use (e.g. #xvimagesink, #ximagesink, etc.) itself; in this case, the application can just create the videosink element, create and realize the window to render the video on and then call gstvideo.video_overlay.VideoOverlay.setWindowHandle directly with the XID or native window handle, before starting up the pipeline. As #playbin and #playsink implement the video overlay interface and proxy it transparently to the actual video sink even if it is created later, this case also applies when using these elements.
In the other and more common case, the application does not know in advance what GStreamer video sink element will be used for video output. This is usually the case when an element such as #autovideosink is used. In this case, the video sink element itself is created asynchronously from a GStreamer streaming thread some time after the pipeline has been started up. When that happens, however, the video sink will need to know right then whether to render onto an already existing application window or whether to create its own window. This is when it posts a prepare-window-handle message, and that is also why this message needs to be handled in a sync bus handler which will be called from the streaming thread directly (because the video sink will need an answer right then).
As response to the prepare-window-handle element message in the bus sync handler, the application may use gstvideo.video_overlay.VideoOverlay.setWindowHandle to tell the video sink to render onto an existing window surface. At this point the application should already have obtained the window handle / XID, so it just needs to set it. It is generally not advisable to call any GUI toolkit functions or window system functions from the streaming thread in which the prepare-window-handle message is handled, because most GUI toolkits and windowing systems are not thread-safe at all and a lot of care would be required to co-ordinate the toolkit and window system calls of the different threads (Gtk+ users please note: prior to Gtk+ 2.18 GDK_WINDOW_XID was just a simple structure access, so generally fine to do within the bus sync handler; this macro was changed to a function call in Gtk+ 2.18 and later, which is likely to cause problems when called from a sync handler; see below for a better approach without GDK_WINDOW_XID used in the callback).
GstVideoOverlay and Gtk+
GstVideoOverlay and Qt