 Mike Brady
|
95a2dec9af
|
Fix BB bug in player_put_packet flush code.
|
2014-09-27 08:53:06 +01:00 |
|
|
Mike Brady
|
df70427a22
|
Supress multiple sequencial resync activity in some situations.
|
2014-09-27 08:38:02 +01:00 |
|
|
Mike Brady
|
c641dba3a0
|
Add flush logic to the point of adding packets to the queue.
|
2014-09-26 15:12:30 +01:00 |
|
|
Mike Brady
|
f2a54dd00c
|
Send SIGUSR2 to shairport sync and it will send a pause request to the source. Also keep checking buffers.
|
2014-09-22 18:31:30 +01:00 |
|
|
Mike Brady
|
d5fb35cd83
|
Still on the trail of that silencing bug -- fixed a silencing bug that could actually have caused it! Flush is a bit cruder, but I can't tell.
|
2014-09-15 18:14:01 +01:00 |
|
|
Mike Brady
|
f88cc5ad99
|
Fix a bug that sometimes led to the periodic status messages not being sent. Duh.
|
2014-09-15 08:24:36 +01:00 |
|
|
Mike Brady
|
9c4ed54522
|
Detab the code, turn off othe 'Resend ...' message, delete some commented-out code and update a few comments.
|
2014-09-14 21:11:11 +00:00 |
|
|
Mike Brady
|
c7e49a6327
|
Improve sequence checking and packet resending requesting. May fix a bug that silences shairport after about three hours on a poor connection.
|
2014-09-14 17:30:49 +00:00 |
|
|
Mike Brady
|
c87b78cfaf
|
One brown-bag and extra checking on buffer management.
|
2014-09-11 09:07:07 +01:00 |
|
|
Mike Brady
|
0820749207
|
Add debug messages after resend requests; disable last-ditch resend request -- it seems to be bust.
|
2014-09-10 18:55:02 +01:00 |
|
|
Mike Brady
|
266e4b29eb
|
Fix a probably-silent bug in makeing resend requests.
|
2014-09-10 08:41:05 +01:00 |
|
|
Mke Brady
|
edd99037c2
|
Merge osxmods into 2.1
|
2014-09-09 00:14:31 +01:00 |
|
|
Mke Brady
|
22495a0a3b
|
Suppress redundant packet-out-of-sequence messages when missing frames are inserted
|
2014-09-08 23:43:28 +01:00 |
|
|
Mike Brady
|
0462ead319
|
Add seq_t PREDECSSOR and seq_sum functions, use them in place of add and subtract. Make play loop record plays and stats even when given silent frames.
|
2014-09-08 13:28:26 +01:00 |
|
|
Mike Brady
|
caf359befb
|
1. Stop Mac OS X compiler complaining about seq_diff being “inline”
2. Use get_absolute_time_in_fp rather than get_absolute_time_in_ns
3. Remove useless and buggy jitter count stuff
|
2014-09-07 12:43:08 +01:00 |
|
|
Mike Brady
|
385f31b09a
|
Add extra option -t/--timeout to vary the time to wait for more audio before giving up. Zero to disable timeout, default 120 seconds.
|
2014-09-03 13:18:58 +01:00 |
|
|
Mike Brady
|
b21fafe482
|
shorten the duration of the Gibbs phenomenon overlay -- sounds a little better...
|
2014-08-29 12:39:51 +01:00 |
|
|
Mike Brady
|
88fc18ad49
|
Add Gibss phenomenon hack to end of the packet (as well as the start).
|
2014-08-29 12:00:30 +01:00 |
|
|
Mike Brady
|
f6df7e9232
|
Add a simple pseudo-random throttle to keep the rate of corrections below 1:1000 on average -- that is, about 1 packet in three may be interpolated.
|
2014-08-27 18:34:24 +00:00 |
|
|
Mke Brady
|
d713591df8
|
Move the time_of_last_packet timeout check to where it can run even if playback never started or was aborted, duh.
|
2014-08-26 19:23:54 +01:00 |
|
|
Mike Brady
|
771aa1a130
|
Add support for PolarSSL, libsoxr, fix ab_* and seq_* definitions and utilites, remove FANCY_RESAMPLING stuff, fix many many bugs.
|
2014-08-24 16:32:34 +01:00 |
|
|
Mike Brady
|
115991ca24
|
Make the timeout for unexplained loss of audio packets 120 seconds, up from 6 seconds. Too short a delay can make the source think shaiport has gone away. There might be a better way to do this.
|
2014-06-12 11:11:17 +01:00 |
|
|
Mike Brady
|
3f121c2b05
|
Improve the misquotation of Yeats' 'Easter, 1916'
|
2014-06-07 17:45:44 +00:00 |
|
|
Mike Brady
|
fab829cfde
|
Change no-audio timeout from 4 to 6 seconds. An then, only ask to teardown once.
|
2014-06-07 17:38:43 +00:00 |
|
|
Mike Brady
|
32448afc4a
|
If there is a gap in audio for more than 4 seconds, it tears down the connection.
|
2014-06-04 22:12:30 +01:00 |
|
|
Mike Brady
|
d6a69d6f23
|
Delay for 0.1 seconds if first packet is late.
|
2014-05-27 08:16:43 +01:00 |
|
|
Mike Brady
|
f14221698e
|
Flush if first packet is late.
|
2014-05-25 18:48:19 +01:00 |
|
|
Mike Brady
|
86f911c2fd
|
ignore silent frames substituted for missing frames.
|
2014-05-23 16:55:28 +01:00 |
|
|
Mike Brady
|
93e247b182
|
Stop complaint about losing sysnc when asked to stop.
|
2014-05-22 16:33:35 +00:00 |
|
|
Mike Brady
|
38a249624c
|
Flush and resync if player gets more than 30 ms out of sync. Can happen if iTunes machine is very busy or network is bad.
|
2014-05-22 13:39:46 +00:00 |
|
|
Mike Brady
|
2c01e9b927
|
Correct sign of divergence, clean up formatting a little.
|
2014-05-22 13:02:04 +01:00 |
|
|
Mike Brady
|
3c351d7b40
|
Make DAC buffer 6615 frames (0.15 sec.) for Debian 7 on VMWare Fusion.
|
2014-05-21 21:10:58 +01:00 |
|
|
Mike Brady
|
74ba4f3da4
|
Turned off a few debug messages.
|
2014-05-20 16:12:08 +00:00 |
|
|
Mike Brady
|
f1bd0373c7
|
Remove a stupid error -- introduced in the last few updates -- causing it to ignore the contents of the buffer. Duh.
|
2014-05-20 15:55:58 +00:00 |
|
|
Mike Brady
|
049ab78641
|
Check for dealy error return code.
|
2014-05-20 15:19:59 +01:00 |
|
|
Mike Brady
|
1a0932d92a
|
Remove extra returns and generally clean up some debug messages.
|
2014-05-19 19:12:13 +01:00 |
|
|
Mike Brady
|
26992acf95
|
Remove the biquad filter code.
|
2014-05-19 17:48:08 +01:00 |
|
 Mike Brady
|
08199b9910
|
Add copyright and edit a 'too late' debug message.
|
2014-05-12 17:17:19 +01:00 |
|
|
Mike Brady
|
dcf0b80f70
|
Oops -- the sign on the drift is wrong. Fixed.
|
2014-05-09 18:15:21 +01:00 |
|
|
Mike Brady
|
82cef38bb3
|
Make use of vol2attn in software and hardware mixers.
|
2014-05-09 17:54:30 +01:00 |
|
|
Mike Brady
|
2d24a265a0
|
As before.
|
2014-05-07 15:53:35 +01:00 |
|
|
Mike Brady
|
3ac57d2cb2
|
Allow DAC to have 11025 frames; add flushing if we are too late to start the first frame.
|
2014-05-07 15:50:50 +01:00 |
|
|
Mike Brady
|
a4d3e115db
|
Add some extra diagnosrics output.
|
2014-05-07 10:19:23 +01:00 |
|
|
Mike Brady
|
75c15ba333
|
Fix initial silence calculation that could cause it to wait for about 100,000 seconds.
|
2014-05-06 23:21:56 +01:00 |
|
|
Mike Brady
|
e464e99337
|
Add some diagnostics, fix 252 frame error -> should be 352. Remove chatty debug messages.
|
2014-05-05 21:31:20 +01:00 |
|
|
Mike Brady
|
9a5e313d93
|
Improve error detection.
|
2014-05-05 21:19:58 +01:00 |
|
|
Mike Brady
|
b00238eccb
|
Make initial latency report more truthful
|
2014-05-02 18:02:21 +01:00 |
|
|
Mike Brady
|
f531194e7a
|
Improve seriously broken flush facility.
|
2014-05-02 17:43:08 +01:00 |
|
|
Mike Brady
|
7187b63eb6
|
too numerous to mention: watch timing of everything, add pthread_cond_wait signalling, monitor latency of DAC, try to guess initial delay for first packet, trim or stuff packets to keep latency within bounds, much much more...
|
2014-04-30 20:54:52 +01:00 |
|
 James Laird
|
8259ae60e8
|
die/warn functions now have implicit newline at the end
|
2013-06-04 00:30:17 +10:00 |
|