[gstreamer-bugs] [Bug 626857] New: MJPEG frame decode element should not generate an error if decoder fails to decode.

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=626857
  GStreamer | gst-plugins-good | 0.10.24

           Summary: MJPEG frame decode element should not generate an
                    error if decoder fails to decode.
    Classification: Desktop
           Product: GStreamer
           Version: 0.10.24
        OS/Version: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: Normal
         Component: gst-plugins-good
        AssignedTo: gstreamer-bugs at lists.sourceforge.net
        ReportedBy: GStreamer-Bugs at tycosp.com
         QAContact: gstreamer-bugs at lists.sourceforge.net
      GNOME target: ---
     GNOME version: ---


Currently, the gstreamer MJPEG element
(gst-plugins-good.XXX/ext/jpeg/gstjpegdec.c) will generate an error
(GST_ELEMENT_ERROR/GST_FLOW_ERROR) if the decoder fails to decode for a couple
of specific anomalies.  We would like for these “anomalies” to actually
generate a (GST_ELEMENT_WARNING/GST_FLOW_OK) instead of errors.  We attempt to
obey the documentation: (GST_MESSAGE_ERROR
an error occured. When the application receives an error message it should stop
playback of the pipeline and not assume that more data will be played. 
).  So in our application, the pipeline will ultimately be torn-down and
rebuilt thus possibly disturbing important video footage.  A copy of an IRC
chat with feedback from some gstreamer representatives has been attached.

<GstMatt> Greetings.
<GstMatt> Can I get an opinion?
<GstMatt> Should it be an error if a MJPEG frame fails to decode?
<__tim> probably not
<MikeS-tp> ideally not; perhaps a warning
<GstMatt> It was bothering me, because I'm tearing down the pipeline &
rebuilding, when I get an error.
<GstMatt> But the next frame might be (is probably) just fine...
<GstMatt> I'm using decodebin, but I think it's using the ffmpeg jpeg decoder,
inside
<GstMatt> Would you consider that a bug, then?
<MikeS-tp> GstMatt: if you're using the ffmpeg plugin, then.. don't?
<MikeS-tp> GstMatt: if the jpegdec element behaves the same, feel free to file
a bug
<GstMatt> You mean if the jpegdec element also treats it as an error, then file
a bug on the jpegdec element?
<MikeS-tp> Yes. It's generally a bad idea to use the ffmpeg elements where
there are alternatives.
<GstMatt> Ah.
<GstMatt> But ffmpeg's JPEG decoder uses MMX
<GstMatt> Is it worth trying  to make the ffmpeg element behave better?
<GstMatt> (elements, I mean)

== MikeS-tp [~msmith at c-67-188-38-75.hsd1.ca.comcast.net]
==  realname : Michael Smith
==  channels : #gstreamer 
==  server   : leguin.freenode.net [Ume?, SE, EU]
==  account  : MikeS-tp
== End of WHOIS

== __tim [~timm at dsl-217-155-195-89.zen.co.uk]
==  realname : Tim-Philipp Müller
==  channels : #gstreamer 
==  server   : kornbluth.freenode.net [Frankfurt, Germany]
==           : is using a secure connection
==  account  : __tim
== End of WHOIS

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.