13
0
Commit Graph

12759 Commits

Author SHA1 Message Date
fbe94f4470 pixel-align canvas PolyItems with single-pixel outlines, for crispness 2014-01-13 09:39:50 -05:00
d202fd9ad9 delete 'big' 2d panner window, if panner changes to mono|stereo 2014-01-13 09:39:50 -05:00
2f3830eddb fix delivery: check if panshell exists on session-load 2014-01-13 09:39:50 -05:00
6b7fd786aa ProcessorWindowProxy::processor_going_away takes care of deletion 2014-01-13 09:39:50 -05:00
2edc3d1b74 misc panning related UI tweaks:
* connect 2D panner "edit" to big window
* disconnect 2D-panner GUI when it's visible but panner-type changes
* ignore mixer-strip level-meter context-menu for Aux-sends
2014-01-13 09:39:50 -05:00
113f28b6fe independent panning for external sends 2014-01-13 09:39:50 -05:00
533e0091b7 fix output metering for Sends (Aux and External) 2014-01-13 09:39:50 -05:00
7e1e7e7629 add independent panner for internal (Aux) sends 2014-01-13 09:39:50 -05:00
0da58eeebf update internal-send port-count when target port-count changes 2014-01-13 09:39:50 -05:00
b4dea5309d ignore additional channels for AFL, PFL:
If the monitor-section has fewer-channels than the solo-listen point:
ignore additional channels.
2014-01-13 09:39:50 -05:00
1e4e6499e3 fix deadlock when removing monitoring-section 2014-01-13 09:39:50 -05:00
1b24aad933 fix processor -> reconfigure I/O || process concurrency
Add a ReaderLock to Route::process_output_buffers().

But process_output_buffers() is always called with processor-lock
held. To avoid deadlocks, a processor WriterLock must always imply
a process-lock (IFF reconfigure-I/O is called with _processor_lock).

Otherwise: e.g.
*  add_processor() -> takes processor-lock. set up and activate processor.
*  simult. audio-engine process, process-lock -> call process_output_buffers() -> wait for processor-lock
*  add_processor() continues -> calls reconfigure-io -> take process-lock -> deadlock.
2014-01-13 09:39:50 -05:00
ac8eb0581d take process lock when adding processors:
fixes possible crash if a processor modifies port-count

1. a processor is inserted and activated with processor-lock held
2. only after that the process_lock() is taken, configure_processors() is called which reconfigures-IO

BUT if the processor that is inserted changes the channel count AND audio is processed before IOs are reconfigured
 -> possible crash (invalid port-buffers)

To reproduce: Bus1 (2in, 3out), Bus2 (2in, 3out)
- add a send from Bus1 to Bus2,
- then add a processor to Bus1, just before the send which
  increases the channel-count to 4 -> occasional crash or assert.
2014-01-13 09:39:18 -05:00
ac8e258e15 VBAP panner UI tweaks:
* fix mouse-grab of sentinel
* make GUI more hemisphere like (circles at latitude)
* change alpha slightly to show signal overlap
2014-01-13 09:39:18 -05:00
a919ac8a86 aubio 3+4 compat 2014-01-13 09:38:24 -05:00
75e21d3ab4 cosmetic change - properly use lilv API 2014-01-13 09:38:24 -05:00
39d5ef5034 don't display elevation if panner does not support it. 2014-01-13 09:38:24 -05:00
28470af5f7 swap channel VBAP channel 2014-01-13 09:38:24 -05:00
5d8e392892 credit where credit is due for speaker-config and VBAP fixes inspiration 2014-01-13 09:38:24 -05:00
cc34174652 VBAP GUI object-grab & position 2014-01-13 09:38:24 -05:00
97758dcbea hide speaker-config in menu (it's not implemented yet) 2014-01-13 09:38:24 -05:00
a5ff738b87 add spinbox to control width in VBAP GUI,.. 2014-01-13 09:37:35 -05:00
dd4c0e040f VBAP rework (part III): fix position computation backend & GUI 2014-01-13 09:37:35 -05:00
431babc272 update default speaker position list 2014-01-13 09:37:35 -05:00
60fe9606f4 VBAP GUI depends on signal-position (not parameter changes) 2014-01-13 09:37:35 -05:00
1bf9c4c990 VBAP backend re-work (part two): speaker positioning
* clean up source (whitespace)
* fix speaker 3x3 matrix iteration
* update math to go along with Ardour Cartesian -- fixes rounding errors
* fix division by zero in cross_prod()
* disable old debug output

(NB PBD::spherical_to_cartesian() returns
  3.7494e-33, 6.12323e-17, 1 for azimuth 90 elevation 90 distance 1
while it should return
  0.000000, 0.000000, 1 for azimuth 90 elevation 90 distance 1
IOW  cos(90.0 * 2.0 * M_PI / 360.0) != 0
Cause unknown. This is currently worked around check in vec_length()
)
2014-01-13 09:37:34 -05:00
112de00841 VBAP backend re-work (part one):
* fix azimuth, don't clamp but map to [0,1]
* prepare elevation (10+ speakers)
2014-01-13 09:35:57 -05:00
5d8da11f4b Add provision for building 'libtimecode' as a DLL (MSVC build) 2014-01-13 14:27:56 +00:00
1fb9437e26 'libs/timecode' - Change some instances of 'rint' to be 'lrint' (better compatibility with MSVC) 2014-01-13 14:26:19 +00:00
291618fe71 allow to custom select panner-type for each delivery. 2014-01-13 15:13:37 +01:00
5da8c2b476 'libs/evoral' - and add 'Note.impl' to the project 2014-01-13 12:47:21 +00:00
1d69f31266 'libs/evoral' - #include 'Note.impl' when building with MSVC (this is necessary to ensure that Evoral::Note gets exported from the evoral DLL) 2014-01-13 12:16:02 +00:00
9c595f8131 delete 'big' 2d panner window, if panner changes to mono|stereo 2014-01-13 12:02:00 +01:00
316973d7c7 Merge branch 'windows+cc' into cairocanvas 2014-01-13 10:49:18 +00:00
346310c60b fix delivery: check if panshell exists on session-load 2014-01-13 11:49:09 +01:00
8e51a6c84f Check in some minor modifications (to MSVC specific headers) 2014-01-13 10:22:31 +00:00
35aed0efab ProcessorWindowProxy::processor_going_away takes care of deletion 2014-01-13 11:08:51 +01:00
1146d58d21 misc panning related UI tweaks:
* connect 2D panner "edit" to big window
* disconnect 2D-panner GUI when it's visible but panner-type changes
* ignore mixer-strip level-meter context-menu for Aux-sends
2014-01-13 11:08:37 +01:00
bc88203ef5 independent panning for external sends 2014-01-13 10:50:49 +01:00
d9cf6880b6 fix output metering for Sends (Aux and External) 2014-01-13 10:50:22 +01:00
0559c1babb add independent panner for internal (Aux) sends 2014-01-13 10:48:37 +01:00
0c384b7c21 update internal-send port-count when target port-count changes 2014-01-13 10:48:23 +01:00
5e2a145cdd ignore additional channels for AFL, PFL:
If the monitor-section has fewer-channels than the solo-listen point:
ignore additional channels.
2014-01-13 10:47:17 +01:00
35ab234e5e fix deadlock when removing monitoring-section 2014-01-13 10:46:11 +01:00
9c753dbaec use LIBEVORAL_TEMPLATE_API with Evoral::Note 2014-01-12 20:23:45 -05:00
2fc2df8c5d fix 64 bit template types 2014-01-12 20:01:12 -05:00
cdfc094b9f add new file (required by MSVC) to sources list 2014-01-12 20:00:55 -05:00
491bc4e865 use LIBEVORAL_TEMPLATE_API
Also clean up unnecessary comments - git blame provides complete per-line history, no need to leave comments behind
2014-01-12 20:00:26 -05:00
a94dadc759 add LIBEVORAL_TEMPLATE_API to deal with gcc/msvc difference of opinion 2014-01-12 19:59:40 -05:00
a08e7c0016 fix processor -> reconfigure I/O || process concurrency
Add a ReaderLock to Route::process_output_buffers().

But process_output_buffers() is always called with processor-lock
held. To avoid deadlocks, a processor WriterLock must always imply
a process-lock (IFF reconfigure-I/O is called with _processor_lock).

Otherwise: e.g.
*  add_processor() -> takes processor-lock. set up and activate processor.
*  simult. audio-engine process, process-lock -> call process_output_buffers() -> wait for processor-lock
*  add_processor() continues -> calls reconfigure-io -> take process-lock -> deadlock.
2014-01-12 23:13:35 +01:00