Dynamic textoverlay from network
osm-lists at mazdermind.de
Tue Jun 9 03:19:01 PDT 2015
On 08.06.2015 20:20, uler3161 wrote:
> Unfortunately this appears to put in quite a lot of lag. Maybe
> a second or two. Not so great for video conferencing. I'm pretty sure this
> is due to the h264 decode and/or encode.
While you could tweak the encoder to produce less lag (disable b-frames,
reduce number of p-pframes etc.) reencoding the stream is not a nice way
to do this.
You could however mux your text, as a subtitle, into the stream, but
this also depends on the container format you're using. This way the
test would be synced to the particular frames visible at that time.
> So my thought was to use the PC to
> generate the countdown and transmit it over the network to the Pis. If I
> could do this, then I could also probably use the same method to send any
> message to both Pis to display.
You could try to embed the Text into the Container like you would with a
Subtitle. This can be done on the PC because muxing another stream in
the container does not require de/reencoding the video/audio is should
be nearly lag-free.
If you don't want to mess around with the container, you can always open
a separate tcp-connection and feed your messages in that way. The
program running on the pi can set the message overlays by textoverlay by
setting the text-property dynamicly when a packet comes in.
> I'm using gst-launch-1.0 version 1.4.5. I'm not afraid to write code, but
> would prefer to have something I could just use without code. I see there is
> a text_sink on the textoverlay. Can that be of use, maybe with the tcp
gst-launch is really onyl a test utility. It's very easy to write small
GStreamer-Programs in Python:
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 473 bytes
Desc: OpenPGP digital signature
More information about the gstreamer-devel