The Fine Print: The following comments
are owned by whomever posted them.
( Reply )
|
Re: PulseAudio and Phonon?
by Skeithy on Wednesday 31/Oct/2007, @17:54
|
To add another question, what is the relation of a sound server and a multimedia framework? Multimedia framework provides an API to play audio/video and a sound server manages audio devices, but what is their relation to each other? Do multimedia frameworks depend on a specific sound server?
|
[
Reply To This | View ]
|
Re: PulseAudio and Phonon?
by bonybrown on Wednesday 31/Oct/2007, @19:38
|
Phonon provides an API to applications for playing sound and video.
It in turn has backends that actually do the decoding and playing.
One of the backends that exists uses xine and xine-lib to do the decoding and playing....so when used with this backend, Phonon will support any file type that xine can play.
There's a backend for gStreamer that's in progress. I presume that for KDE apps on Mac and Windows there will need to be an appropriate backend built for those too (perhaps the xine-lib version will just work on those platforms...I don't know)
Each of the backends will have it's own configuration as to how it gets the data to your ears - they will support whatever sound-servers or devices that the backend is built on....so if you want to use pulseaudio, and the backend you're using supports pulseaudio for output, it should be possible.
I know that xine supports pulseaudio (because I use it with an LTSP setup ) so everything should "just work"!
|
[
Reply To This | View ]
|
Re: PulseAudio and Phonon?
by Thomas on Wednesday 31/Oct/2007, @19:41
|
Phonon provides an api to standard applications. Then depending on the system, phonon can use a variety of backends to decode the audio e.g. xine,gstreamer... Xine/gstreamrer then output to a soundserver which mixes it and sends to the soundcard.
|
[
Reply To This | View ]
|
Re: PulseAudio and Phonon?
by Kevin Krammer on Thursday 01/Nov/2007, @05:00
|
Sound servers are usually used for either mixing sounds from different source into one, probably applying effects to each source, or for sending the data somewhere else, e.g. over the network to a partner sound server.
Multimedia frameworks do not need a soundserver, i.e. it is not a technical requirement.
The mixing can also be done in the audio driver or the audio hardware itself, things like redirecting sound can also be handled by the framework itself.
Usually multimedia frameworks therefore provide several output options, e.g. directly using the sound device and a couple of sound servers.
|
[
Reply To This | View ]
|
|
Re: PulseAudio and Phonon?
by Sutoka on Wednesday 31/Oct/2007, @20:56
|
Not directly. PulseAudio doesn't do any decoding, but PulseAudio can be used as an output device for things like xinelib, gstreamer, and mplayer (I think gstreamer has a pulseaudio backend, maybe xine, not sure about mplayer).
|
[
Reply To This | View ]
|
Re: PulseAudio and Phonon?
by Darryl Wheatley on Thursday 01/Nov/2007, @00:32
|
I'm confused. If you look at the screenshot of the various pulseaudio config utilities at http://0pointer.de/public/pulse-screenshot.png (warning: biggish image), you see stuff like PulseAudio Manager which lets you "view & modify the Daemon's internals", a device chooser (I guess phonon already covers this?) and Volume Control, which lets you configure "sinks" and have different volume levels for the various applications and computer speakers. I'm guessing Phonon will be doing this anyway?
In an interview with the Fedora man responsible (http://fedoraproject.org/wiki/Interviews/LennartPoettering), he talked about it being all about "ear candy" and a "compiz for sound" e.g. active window has sound at 100%, other window's sounds are at 20%; clicking left button makes sound come out of left speaker and same with right. Will Phonon offer similar features or will it need PulseAudio to offer this functionality?
What I don't understand is, if pulseaudio is a sound server, won't KDE 4 need one too? Or is Phonon going to let us choose both the media engines (xine/gstreamer/mplayer/nmm) and sound servers used?
|
[
Reply To This | View ]
|
Re: PulseAudio and Phonon?
by Morty on Thursday 01/Nov/2007, @03:07
|
"active window has sound at 100%, other window's sounds are at 20%; clicking left button makes sound come out of left speaker and same with right"
This will also need support by the windowmanager etc, but I don't think there should be any problem for Phonon to support this. As long as the underlying media engine/soundsystem support mixing there is nothing special about it.
That said, I'm not particularly impressed with what I read abut PulseAudio in that link. Actually It does not offer anything new comparing to what aRts does(A better design perhaps, but that should be expected compared to something as old as aRts).
|
[
Reply To This | View ]
|
Re: PulseAudio and Phonon?
by Colin Guthrie on Friday 02/Nov/2007, @06:34
|
Forgive me for not understanding, but does Phonon actually use the backend (xine, gstreamer etc.) only for decoding and then bring the sound back in and output it to the final destination (e.g. alsa)? Or does it let the backend do the actual sound output too? It was my understanding that the latter was the case.
If this was provided then surely Pulse adds a lot to the table? Using a Pulse as the final output device would allow for easy hotplugging of audio devices and seamless output. With pulse you could easily start playing a tune in Amarok and have it come out your laptop's speakers, plug in a USB soundcard and then move the playing stream across to it. No need to stop the music and reconfigure things and play it again? Perhaps phonon can do this without pulse (e.g. it has it's own callback based API) and if so then I guess it doesn't add too much other than network device transparency, simultaneous output to all your local sound devices and simultaneous output to all your network sounds devices... which although very cool, it not of interest to most average joe users.
|
[
Reply To This | View ]
|
Re: PulseAudio and Phonon?
by Kevin Kofler on Sunday 04/Nov/2007, @18:34
|
The callgraph actually looks like:
Phonon -> xine-lib/GStreamer -> PulseAudio -> ALSA
Xine-lib or GStreamer are used to implement the Phonon backend. (KDE 4.0 will come with Phonon-Xine. As far as I know, Trolltech is working on a GStreamer backend for Phonon.) They handle decoding using a plugin framework, then hand off the decoded stream to either the hardware or a sound server such as PulseAudio through an output plugin.
PulseAudio can also work as an ALSA plugin, in which case the applications think they're talking to a native ALSA device, but what they get is not a hardware device, but a connection to PulseAudio. This is very similar to how dmix works. In a Phonon context, this makes the callgraph look like this:
Phonon -> xine-lib/GStreamer -> ALSA (PulseAudio plugin) -> PulseAudio -> ALSA (hardware device)
I've read reports that xine-lib's native PulseAudio output plugin currently isn't very stable, so with xine-lib using the PulseAudio ALSA plugin is more reliable. Fixing this in xine-lib is on the PulseAudio TODO list, but in a KDE 4 context it might become moot with the Trolltech-backed Phonon-GStreamer anyway.
|
[
Reply To This | View ]
|
|
Re: PulseAudio and Phonon?
by Ben on Thursday 01/Nov/2007, @07:03
|
In Phonon you select what backend you want to use, the backend will turn your mp3 and ogg into actual sound.
You can also select where you want that sound to end up, providing the backend supports it, soundcard, headphones, Pulse, JACK.
Screenshot here: http://dot.kde.org/1170773239/1170778900/1170862970/1170863051/kcmphonon5.png
|
[
Reply To This | View ]
|
The Fine Print: The previous
comments are owned by whomever posted them.
( Reply )
|
|
![[KDE Dot News]](http://www.kde.org/dot/Images/kdedotnews.gif)
