The VideoLAN Forums

Hi everybody,

Does somebody know a way to comine a video input with invmem (or imem) and an audio input (e.g dshow input) to a single network stream?
If I use invmem, I've to create a media with MRU fake:// but then I'm not able to use MRU dshow:// for audio input and vice versa.

Does somebody know a way to get video and audio combined to one stream by VLC if video has to be provided with invmem? The way of getting the audio signal is not predefined, it only has to capture the audio from an available audio source (e.g. microphone).

Any help would be much appreciated.

Or explained a little bit simpler: Can libVLC handle two inputs (video over invmem, audio over something else) for one player?

http://en.wikipedia.org/wiki/VLC_media_ ... te_note-14 might be interesting to you.

Short answer is if it's a directshow video input and a directshow audio input, it should work out of the box. If not, then I'm not sure entirely how you would combine inputs. Maybe using mosaic? Maybe write your own directshow input?

It's currently not a DShow video input but I'll create on. Should work with the DShow access module from VLC. Thx for your help and the link you've provided.

Use a slave input. But it might be difficult to keep the time sources consistent so that the synchronization works

Hi Remi,

I tried to use dshow:// for audio as slave input with invmem for video as primary input but it does not work. Also if I switch them (dshow with audio as primary and invmem with video as slave) I only get a stream with a picture but no sound. I see both, the invmem and dshow input modules, get initialized and successfully opened but when constructing my output stream, mux_ts never gets an audio stream added so it's clear that the stream is only constructed from the transcoded h264 video stream provided from invmem.

Should input slave actually work that way?