On Thu, 2005-04-21 at 22:26 +0200, Kyrre Ness Sjobak wrote: > tor, 21.04.2005 kl. 19.18 skrev Pete Zaitcev: > > On Fri, 22 Apr 2005 00:10:23 +1200, David Mohring <heretic@xxxxxxxxxx> wrote: > > > > > On Thu, 2005-04-21 at 13:39 +0300, Marius Andreiana wrote: > > > > Looks like infrastructure for removing esd is in place. > > > > The following prevent it: > > > > > > The esound demon ESD, clunky as it, is still currently the best way to > > > serve sound to remote X11/LTSP terminals. [...] > > > > This should not prevent us from removing dependencies on ESD, so that > > we do not have to have it installed unconditionally. The vast majority of > > users do not need it. Currently, gnome-libs and libgnomeui depend on esound. > > David wants to remove it completely, which may be extreme. > > > > -- Pete > > What about making some (new?) protocol for sound-over-network - > something that *just is there* when you connect over XDMCP? Maybe > something hooked into ALSA just to make it more complicated - like a > "viritual soundcard device"? > See Proposal: Rationalizing Fedora Audio https://www.redhat.com/archives/fedora-devel-list/2004-October/msg00899.htm But as for removing ESD or even arts as application sound output options, you have to remember that each layer adds more latency and processing overhead. ALSA -> Jack -> ESD -> Network ( yes it is possible) works if your only serving one or two users, but ESD -> Network does not use anything like the required overhead. -- David Mohring <heretic@xxxxxxxxxx>
