--- Michel D�nzer <[EMAIL PROTECTED]> wrote: > On Die, 2002-09-17 at 17:33, Ondrej Certik wrote: > > You are right, that the problem is endiannes, > because > > I wanted to know what is going on, so I switched > back > > to ALSA and > > modified the Alsa output plugin for xmms, so that > it > > swaps > > the endiannes of the data, which goes through it > and > > ---- > > the ogg/mp3 now plays very well, but wav causes > the > > noise. > > Oh dear, like that it probably gets swapped at least > twice in > software...
I know. I just wanted to know if endiannes really is the problem. > > > So there is only one question: which code should I > > modify > > to swap the endiannes? Probably the input plugin > for > > ogg/mp3? > > Or both input plugin for wav and alsa output > plugin? > > Or even alsa driver? Everything should work. > > > > I don't exactly understand which fault it is. If I > > understand > > it well, than Alsa expects input in native > endiannes. > > So > > on ppc, the alsa-xmms plugin must send the output > in > > big-endian. > > So the change has to be in ogg plugin. But I don't > > know > > which endiannes xmms expect its data. So maybe > > changing ogg plugin > > isn't the solution. > > Ideally, the input plugins would provide data in > native endianness, and > the other stages would simply pass it through. I > don't know how many > interfaces there are between input plugin and output > device, you'd have > to make sure that it's labeled correctly at each of > them. I know. But I thought that maybe you know where the fault is. Because I don't like the idea to spend a week browsing xmms code just to know how many interfaces there are. If we know, that sometimes the endiannes is wrong, why just not to ask it on Alsa and xmms lists? I think that the problem can be easily solved. They must know better where exactly in the code the swap has to be done. Should I ask on Alsa and/or xmms list? __________________________________________________ Do you Yahoo!? Yahoo! News - Today's headlines http://news.yahoo.com
