Actually, I set fPresentationTime to the application time (time got using QPC, which I saw is also used in gettimeofday()) at which the initial frame was captured. So, it should be consistent and increasing. All NALUs composing a single frame have the same presentation time, is that incorrect ? I will try to call gettimeofday() and use the value return when sending each single NALUs instead of my "saved time"
This is basically what happens. In the video, you can see the two clients (two different PCs), getting the stream from a third PC. When playing the stream on the first client, no problem (once buffering is diong its job). When starting the stream on the second client, one can see the first one pauses: https://youtu.be/gLqlbUjH3Y4 Some further investigation shows that playing time shown by VLC briefly jumps from 0:00:xx to 5:xx:xx on the first client when starting the second client, and then goes back to the original value. https://youtu.be/NmfqCkLP1PM This is weird :/ -----Original Message----- From: live-devel <live-devel-boun...@us.live555.com> On Behalf Of Ross Finlayson Sent: Monday, February 14, 2022 3:17 PM To: LIVE555 Streaming Media - development & use <live-de...@us.live555.com> Subject: Re: [Live-devel] compilation issue / vcpkg / windows > On Feb 15, 2022, at 2:44 AM, g.jaegy <g.ja...@imagine3d.fr> wrote: > > This all works, until a second client connect. In that case, the first client > stop playing. > > I must be doing something wrong, but I don't know what. > > Would you have any advice or hint that would allow me to find out what is > wrong ? Not really, except to repeat something that I said way back on February 5: > On Feb 5, 2022, at 5:52 AM, Ross Finlayson <finlay...@live555.com> wrote: > >> Only issue I have, when connecting a second client, the video on the first >> client stops one second later, basically the picture is not updated anymore. >> VLC statistics still shows newly decoded blocks and displayed frame, time >> increases, but the picture doesn't change on that first client. I'm using >> multicast sockets (same as in example). Maybe you have an idea ? > > Not really - except to check the “fPresentationTime” values that you’re > setting for each NAL unit (in your “FramedSource” subclass). They must be > aligned with ‘wall clock’ time - i.e., the times that you’d get by calling > “gettimeofday()”. I notice that your code is setting “fPresentationTime” from your “dNALUTime” variable, which is *not* set by calling “gettimeofday()”. That’s fine, but the “fPresentationTime” values that you set here MUST be the same as the values that you’d get by calling “gettimeofday()”. This is the first thing I’d check. Ross Finlayson Live Networks, Inc. http://www.live555.com/ _______________________________________________ live-devel mailing list live-devel@lists.live555.com http://lists.live555.com/mailman/listinfo/live-devel _______________________________________________ live-devel mailing list live-devel@lists.live555.com http://lists.live555.com/mailman/listinfo/live-devel
