2002.07.30 tprado@charter.net * Changed default ALSA device to be plughw:# instead of hw:#. (Martijn) * Changed -d parameter (when ALSA compiled in) to be passed directly to the sound card open call instead of being converted to hw:# so the parameter needs to be specified as hw:0 instead of 0 for sound card 0. * Removed call to snd_pcm_hw_params_set_periods_integer in ALSA9 card init. (Martijn) * esound.ebuild is created dynamically from esound.ebuild.in for Gentoo Linux. Gentoo's awesome, give it a try! 2002.06.23 tprado@charter.net * Merged with 0.2.28. * -unlocked option replaced by -promiscuous option now included with esound. * Replaced alsa routines with my own ones so my laptop will work. 2002.06.15 tprado@charter.net * Merged with 0.2.27, since alot had changed between 0.2.24 and this one, and it happened so rapidly, I manually added all my patches back into the tree where needed. Several were dropped since there were similar implementations in the new esound. * Added back in -unlock option and start_unlocked parameter to esd.conf * Rewrote esd_audio_write in audio_alsa09.c to be similar to the way I had it in my patch set because it wouldn't work with my laptop as it was. * Enhanced alsa init routines to reset/close the card after failures. * Change snd_pcm_hw_params_set_rate_near to snd_pcm_hw_params_set_rate in alsa init. * Add back in support to specify device in alsa driver. * Rewrote esd_audio_flush in audio_alsa09.c to be like my patch was before - includes error checks and a snd_pcm_prepare step. * Abandoned maintaining alsa04 and alsa05 seperate files, we're never going back, right?! * Enhance alsa detection in configure.in * Add back in configure setting summary in configure.in * Add back -v option in instructions * Add back alsa card detection of available devices when -h is passed to esd. * Add back display of esd version when -h is passed to esd. * Removed /tmp/.esd-/socket stuff, it's doing it the original way again. No dependancy on glib. * Added .ebuild file for gentoo linux. 2002.06.02 tprado@charter.net * Merged with 0.2.26 * Removed mini_sox stuff * Removed Alt-mix stuff 2002.04.22 tprado@charter.net Added bigendian to alsa support. 2002.03.19 tprado@charter.net Fixed several sound anomalies when compiled for alsa. Removed alsa_buffer queuing logic I had added previously, it wasn't really needed. Added code to symlink /tmp/.esd-/socket to /tmp/.esd-socket so clients owned by other users can connect. Added new esd option -unlock to be able to start the sound server unlocked so that other users can connect to it (via unix domain sockets). Added start_unlocked parameter to esd.conf to be able to do the same that way (It's currently defaulted to 0). 2002.03.14 tprado@charter.net I found a bug in players.c where the actual_length was getting set to twice the buffer_length so I was getting segfaults. Since the original code worked for me I changed the read_player code back to the way I had it before. 2002.03.13 tprado@charter.net Merged with 0.2.24. The ChangeLog for 0.2.24 mentions fixes to player.c and clients.c which my patch-set had already included. However; there was also changes to mix.c that fixes the 44.1khz to 48khz metallic sound. This is very cool. I've fixed the trailing sounds in the alsa code (again). 2002.02.28 tprado@charter.net Lots of changes here. I wanted to play around with the SoX interfaces so I've started a rewrite of the mixer functionality to use SoX. There's also another rewrite of the stereo16s to stereo32s conversions that I put in. These are both disabled by default but you can use --enable-sox-mix and --enable-alt-mix to enable them. They're exclusive though so if both are enabled, the sox mix will be the one used (and it's broken right now). The alt mix seems to work pretty good for me though. I've found and squished various alsa related bugs. The most important being one where it would find the best suitable frequency rate to use but esd was getting updated if it was different than the one requested. I.e. if a card only supported 48khz then alsa would use 48 but esd would be set for 44.1. This would cause the samples to play too fast. I've found that the mixer functionality doesn't translate 44.1 to 48 very well so it's kinda scratchy at the moment if you try to use 48. There's some cleanups in the autostandby code as well. 2001.12.19 tprado@charter.net It turns out some sound cards don't like it when you manually set the period size. The period size and alsa buffer size is now determined by the alsa API. The esd internal buffer size is still 4096 so clients like xmms and mpg123 won't break. The esd output buffer will get cached into the alsa buffer for playback. 2001.11.13 tprado@charter.net I now have esound 0.2.23 working natively with alsa 0.9.0beta9. I finally had a chance to really dive into my alsa patch code to try to figure out why I was getting trailing sounds in my output. It turned out that the buffer size in alsa was getting set to 16K yet esound uses only 4K. I had the period size of alsa set to 4K instead of the buffer size and after monitoring the alsa mailing lists, it turned out that's not what I wanted to do. I've changed the period size to 1K and the buffer size to 4K and viola, no extra blips at the end of samples! Most excellent. 2001.10.21 tprado@charter.net I now have esound 0.2.23 working natively with alsa 0.9.8a. The patch includes many tweaks from the debian distribution as well. Also, instead of using /tmp/.esd/socket, it now uses /tmp/.esd-/socket. So stale sockets/directories from a previous user's login won't break the next user that uses esd. I'm using g_get_user_name() so GLIB is now required. I also cleaned up the configure.in for alsa detection a bit. I fixed a few bugs in esdfile.c which fixes playback problems I experienced when using esdplay. There are many more changes and tweaks as well. 2001.09.02 tprado@charter.net Wow, it's been a while since my last patch posting. New in my patch is: ALSA 0.9.x support! Yippee. ALSA 0.4 support is back (though not tested at all) The source for each version of alsa is in it's own .c file now. At the time of this writing, I've tested against and ALSA 0.9.0beta7 and ALSA 0.5.10b. Interlaced sound cards are not supported in ALSA9 at the moment though. I changed the version extension to ztp5alsa9 so Ximian Gnome won't try to over write it. Have fun! 2001.01.15 tprado@charter.net Applied patches to esound 0.2.22. Using .spec file from Mandrake Cooker so the packaging is a little different than what normally comes on a RedHat or Mandrake CD. 2000.11.23 tprado@charter.net Repackaged and applied some of Linux-Mandrake's patches. Separated my patches from the original gnome tarball in the SRPMS. Linux-Mandrake also has a 10 second delay fix but I prefer the one submitted by Juergen Keil at bugs.gnome.org so that's the one included. 2000.11.20 tprado@charter.net Fixed 10 second delay when starting a program that uses esd when the esd daemon is not yet loaded. 2000.10.14 tprado@charter.net ALSA 0.5.0 or greater is now required if compiling for ALSA. esd -h now shows the ALSA cards available and you can now specify which one to use with esd -d ALSA manages the cards so you have to specify the ALSA card number, not the dsp file. If this is not specified, the first available card found will be used. 2000.10.12 tprado@charter.net Create this web page and make my patches available to all. 2000.03.13 tprado@charter.net Sound garbled when played though a GUS Classic due to it being interleaved. ALSA not using ESD set fragment size Number of ALSA cards still determined with 0.4.x ALSA function when using 0.5.x ALSA. Used ALSA recommended API calls when interacting with ALSA (applies only to 0.5.x.) various other tweaks. 2000.02.29 tprado@charter.net Original 0.5.0 patch