Dear all, I want to mix (videomixer) a multimedia source (video or image) with a text (textoverlay) with gnlcomposition to insert the text in the proper time. 1) At the beginning I used a gnloperation to insert the textoverlay inside a bin with a textoverlay. But problems arised when I had a lot of images and texts. It was not possible to get a desired layout. Because to get a correct functionality the priority has to be low and unique (according to my tests). 2) So, I am trying to insert each text as a video with an alpha channel. I achieved it with a gnlsource with a bin that contains a videotestsrc with green pattern and an alpha element to remove the green part with an alpha channel but in some borders of the image there is a residual line. videotestsrc pattern=5 ! ffmpegcolorspace ! textoverlay text="HELLOWORLD" ! alpha prefer-passthrough=TRUE method=1 ! video/x-raw-yuv, framerate=\(fraction\)25/1, format=\(fourcc\)AYUV, width=640, height=480 ! ffmpegcolorspace And for example in this case: textoverlay text="HELLOWORLD" color=4278255360 (0xff00ff00) The text is also "alphaed". So not completely solved yet. Moreover, the alpha module inserts a residual line and computational waste. 3) So I want a pure alpha video with a text to mix it with an image. This is a test pipeline where the videotestsrc has directly a 0x00000000 (ARGB value = pure alpha black = transparent) with a text overlay with color 0xff00ff00 (ARGB value = solid green). gst-launch videomixer name=mix sink_0::zorder=3 sink_1::zorder=2 ! ffmpegcolorspace ! timeoverlay ! autovideosink videotestsrc pattern=17 foreground-color=0 ! textoverlay text="HELLOWORLD" color=4278255360 shaded-background=TRUE ! video/x-raw-yuv, framerate=\(fraction\)25/1, format=\(fourcc\)AYUV, width=640, height=480 ! ffmpegcolorspace ! mix.sink_0 filesrc location=~/image.jpg ! decodebin ! queue ! ffmpegcolorspace ! videoscale ! imagefreeze2 ! alpha prefer-passthrough=TRUE ! video/x-raw-yuv, framerate=\(fraction\)25/1, format=\(fourcc\)AYUV, width=640, height=480 ! ffmpegcolorspace ! mix.sink_1 However the textoverlay plugin doesn't overwrite its alpha value and the text doesn't appear anyplace. It seems the text inherits the alpha value of the videotestsrc. If a change the videotestsrc to get semitransparent color=2147483648 (0x80000000) (ARGB semitransparent black) the text appears but again It seems the text inherits the alpha value of the videotestsrc. gst-launch videomixer name=mix sink_0::zorder=3 sink_1::zorder=2 ! ffmpegcolorspace ! timeoverlay ! autovideosink videotestsrc pattern=17 foreground-color=2147483648 ! textoverlay text="HELLOWORLD" color=4278255360 shaded-background=TRUE ! video/x-raw-yuv, framerate=\(fraction\)25/1, format=\(fourcc\)AYUV, width=640, height=480 ! ffmpegcolorspace ! mix.sink_0 filesrc location=~/image.jpg ! decodebin ! queue ! ffmpegcolorspace ! videoscale ! imagefreeze2 ! alpha prefer-passthrough=TRUE ! video/x-raw-yuv, framerate=\(fraction\)25/1, format=\(fourcc\)AYUV, width=640, height=480 ! ffmpegcolorspace ! mix.sink_1 I tryed to insert alphacolor elements and introduce different schemes of ffmpegcolorspaces and capsfilter, and I also checked that the zorder of the videomixer is right, but I haven't achieved the expected result. Any ideas? Thank you in advance. Best Regards.