[vlc-devel] Re: Re : VLC behavior

hornsby adrian_hornsby at yahoo.co.uk
Wed Nov 9 07:40:59 CET 2005


On Tuesday 08 November 2005 22:57, Jean-Paul Saman wrote:
> adrian hornsby wrote:
> >>>hornsby wrote:
> >>>Hi everyone,
> >>>What would be VLC behavior if in a MPEG2 TS stream,
> >
> > one full GOP is
> >
> >>>missing
> >>>?? would it be able to resyncro without any
> >
> > artifacts on the screen ??
> >
> >>If the first frame is an I-frame, then there won't
> >
> > be any
> > artefacts.
> >
> >>Kind greetings,
> >>Jean-Paul Saman.
> >
> > Hi Jean-Paul,
> > Thanks for your answer. The problem is that a GOP ,in
> > display order, never start with I frame but usually
> > with a B frame. So VLC receive first an I frame then
> > the B frame,  but reorder it, right ? Can vlc lose a
> > bit of info ??
>
> I think you are wrong. An I-frame is a selfcontained frame, a B-frame
> depends on other frames in 2 directions, a P-frame depends on an other
> frame only in one direction. The typical sequence a GOP or series of
> frames is sent is:
>
> IPBBBPBBBBPBBP I
>
> The B-frames are always sent before they are actually needed, but after
> the first I-frame of a GOP.
>
So, if I their is an I frame, no risk to have anything before it,being related 
to it ?? 
I agree with you on the sending order, it makes sense. I have to check my 
streams because I think I didn't have that sending order. I also was checking 
the time_reference field and it suggested something else to !!
Is their some risk that the sending order can be different than the decoding 
order ??? (not the display order that is always different than decoding order 
for the reason you explain)

Or should the sending order respect the decoding order ??

thanks again


> > The thing is that I did the test, and some very little
> > artifact appears on the screen, few scares.
> > My attention is to test my splicing tool with VLC. I
> > use beginning and end of GOP as splicing points. But
> > of course I had to try with the exact same stream
> > first. So i only remove one GOP from the stream and
> > see VLC's behavior.
> >
> > You can use my tool to see yourself. Of course, VLC is
> > so good that I tend to think the problem is in my tool
> > but I checked everything and I couldn't find any
> > explanation.
> >
> > My tool is build on live555.com library. You can
> > download my tool from here.
> > Just compile it and play the "testRelay" file in
> > /testProgs/ directory.
> >
> > testRelay work on command line as follow :
> > ./testRelay <input_multicast_address1> <port1>
> > <input_multicast_address2> <port2>
> > <output_multicast_address1> <port1>
> >
> > (i.e:./testRelay 239.252.40.1 4444 239.252.40.1 4444
> > 239.252.50.1 1234)
> > to ask splicing, press the key "s", to stop "q"
> >
> > The splicing tool is in the liveMedia directory :
> > -MPEG2TransportStreamSplicer.cpp
> >
> > Thanks so much for helping me.
> > Best Regards,
> >
> > I want to release a Real Time splicing tool on GPL, if
> > you want to help me with little thing, please email
> > me. I would really like having experts helping me.
> >
> > adrian
>
> Grtz,
> Jean-Paul Saman.

-- 
##################################
http://thisisallaround.freezee.org
##################################

		
___________________________________________________________ 
To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com

-- 
This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
To unsubscribe, please read http://developers.videolan.org/lists.html



More information about the vlc-devel mailing list