Paul Davis [Wed, 15 Jan 2014 20:00:07 +0000 (15:00 -0500)]
improve comment explaining the use of alloca()
Paul Davis [Wed, 15 Jan 2014 19:58:59 +0000 (14:58 -0500)]
add back comment explaining the use of alloca()
Paul Davis [Wed, 15 Jan 2014 19:29:32 +0000 (14:29 -0500)]
merge with master
Robin Gareus [Wed, 15 Jan 2014 19:03:20 +0000 (20:03 +0100)]
fix pan-width automation
Robin Gareus [Wed, 15 Jan 2014 16:59:16 +0000 (17:59 +0100)]
don't pan send/return inserts
Robin Gareus [Wed, 15 Jan 2014 13:55:26 +0000 (14:55 +0100)]
rework LXVST port assignments - fixes #5827
Robin Gareus [Wed, 15 Jan 2014 13:07:02 +0000 (14:07 +0100)]
stereo-panner: clamp width during processing to valid range
Robin Gareus [Wed, 15 Jan 2014 10:30:54 +0000 (11:30 +0100)]
Midi-track fader & pan automation
NB. Fader automation lane always shows 'dB' value and is logarithmic.
OK for pre-fader synths, but not for raw-midi data.
Robin Gareus [Wed, 15 Jan 2014 10:23:38 +0000 (11:23 +0100)]
update pan automation-track display
Robin Gareus [Wed, 15 Jan 2014 09:38:32 +0000 (10:38 +0100)]
prevent stackoverflow when pannable changes to fewer params
endless loop: e.g. 2in2out -> balance (or 1in1out)
#23 0xb7ab5c17 in ARDOUR::Pannable::value_as_string
#24 0xb2ebb206 in ARDOUR::Pannerbalance::value_as_string
#25 0xb7ab5c17 in ARDOUR::Pannable::value_as_string
#26 0xb2ebb206 in ARDOUR::Pannerbalance::value_as_string
ad infinitum
Robin Gareus [Wed, 15 Jan 2014 09:37:25 +0000 (10:37 +0100)]
re-allow panners for monitoring-section (for now)
Robin Gareus [Wed, 15 Jan 2014 08:45:02 +0000 (09:45 +0100)]
fix invalid width when swiching to 2in2out
Robin Gareus [Wed, 15 Jan 2014 08:37:20 +0000 (09:37 +0100)]
VBAP GUI convention: top == front ^= azimuth == .5
This allows to move from stereo,mono panners to VBAP and back
and also facilitates sharing pannables of all currently
existing panners with semantically similar results.
(somewhat dirty solution, this retains PBD::spherical_to_cartesian
and maps angles pretty much everywhere else)
Paul Davis [Wed, 15 Jan 2014 02:46:41 +0000 (21:46 -0500)]
the vampplugins lib that we build needs to use VAMPSDK
Paul Davis [Wed, 15 Jan 2014 02:13:21 +0000 (21:13 -0500)]
merge with master
Robin Gareus [Wed, 15 Jan 2014 01:51:09 +0000 (02:51 +0100)]
NOOP, semantic update, don't pannable's panner for direct access.
Robin Gareus [Tue, 14 Jan 2014 22:08:34 +0000 (23:08 +0100)]
update panGUI connections (panner-shell vs pannable vs panner)
Robin Gareus [Tue, 14 Jan 2014 20:11:41 +0000 (21:11 +0100)]
NOOP, clean up whitespace/TAB mess
Robin Gareus [Wed, 15 Jan 2014 01:50:17 +0000 (02:50 +0100)]
cont'd backend work on panner-linking
Robin Gareus [Tue, 14 Jan 2014 15:19:49 +0000 (16:19 +0100)]
remove debug code
Paul Davis [Tue, 14 Jan 2014 22:27:21 +0000 (17:27 -0500)]
fix compilation problem on OS X caused by a missing implementation of a virtual method
Paul Davis [Tue, 14 Jan 2014 20:11:32 +0000 (15:11 -0500)]
add back required extern
Paul Davis [Tue, 14 Jan 2014 17:43:27 +0000 (12:43 -0500)]
new MIDI binding map for Roland V-Studio 20, from Keith (Milner?)
Paul Davis [Tue, 14 Jan 2014 17:22:18 +0000 (12:22 -0500)]
revert incorrect include path changes and removal of export macros for two libardour symbols (revision and ardour_config_info)
John Emmas [Tue, 14 Jan 2014 15:01:13 +0000 (15:01 +0000)]
'libs/ardour' - Non-visibility stuff changed for MSVC (mostly alterations to #include paths)
John Emmas [Tue, 14 Jan 2014 11:18:09 +0000 (11:18 +0000)]
'libs/ardour' - Remove the old 'visibility.h' (so as not to merge it into cairocanvas)
John Emmas [Tue, 14 Jan 2014 10:13:12 +0000 (10:13 +0000)]
'libs/ardour' - DLL visibility stuff and associated changes needed for building with MSVC. Currently includes debugging information and things that are just commented out until we have known compatibility with the other platforms (i.e. contains stuff to be removed at a later date)
Paul Davis [Tue, 14 Jan 2014 15:56:17 +0000 (10:56 -0500)]
merge (with conflict fixes) with master (even against rgareus' recommendation)
Paul Davis [Tue, 14 Jan 2014 14:33:02 +0000 (09:33 -0500)]
fix broken build after addition of normalizer.cc
Robin Gareus [Mon, 13 Jan 2014 22:22:02 +0000 (23:22 +0100)]
add preference -- link-send-and-route-panner default
Robin Gareus [Mon, 13 Jan 2014 22:21:30 +0000 (23:21 +0100)]
first stab at send+route panner link
Robin Gareus [Mon, 13 Jan 2014 18:16:10 +0000 (19:16 +0100)]
center VBAP UI crosshair
John Emmas [Mon, 13 Jan 2014 15:05:30 +0000 (15:05 +0000)]
Merge branch 'windows+cc' into cairocanvas
John Emmas [Mon, 13 Jan 2014 15:01:14 +0000 (15:01 +0000)]
'libs/audiographer' - Add a new source file (to accommodate the stuff removed from 'normalizer.h')
John Emmas [Mon, 13 Jan 2014 14:58:04 +0000 (14:58 +0000)]
'libs/audiographer' - DLL visibility stuff and associated changes needed for building with MSVC. Currently includes debugging information and things that are just commented out until we have known compatibility with the other platforms (i.e. contains stuff to be removed at a later date)
Paul Davis [Mon, 13 Jan 2014 14:49:09 +0000 (09:49 -0500)]
clamp marker name text appropriately
Paul Davis [Mon, 13 Jan 2014 14:47:20 +0000 (09:47 -0500)]
fix botched merge of vbap changes
Robin Gareus [Sat, 11 Jan 2014 21:59:25 +0000 (22:59 +0100)]
VBAP backend re-work (part one):
* fix azimuth, don't clamp but map to [0,1]
* prepare elevation (10+ speakers)
Paul Davis [Mon, 13 Jan 2014 14:28:44 +0000 (09:28 -0500)]
define marker height in one location, to avoid duplicated magic numbers and allow easy changing
Paul Davis [Mon, 13 Jan 2014 14:28:06 +0000 (09:28 -0500)]
pixel-align canvas PolyItems with single-pixel outlines, for crispness
Robin Gareus [Mon, 13 Jan 2014 11:02:00 +0000 (12:02 +0100)]
delete 'big' 2d panner window, if panner changes to mono|stereo
Robin Gareus [Mon, 13 Jan 2014 10:49:09 +0000 (11:49 +0100)]
fix delivery: check if panshell exists on session-load
Robin Gareus [Mon, 13 Jan 2014 10:08:51 +0000 (11:08 +0100)]
ProcessorWindowProxy::processor_going_away takes care of deletion
Robin Gareus [Mon, 13 Jan 2014 10:08:37 +0000 (11:08 +0100)]
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
Robin Gareus [Mon, 13 Jan 2014 09:50:49 +0000 (10:50 +0100)]
independent panning for external sends
Robin Gareus [Mon, 13 Jan 2014 09:50:22 +0000 (10:50 +0100)]
fix output metering for Sends (Aux and External)
Robin Gareus [Mon, 13 Jan 2014 09:48:37 +0000 (10:48 +0100)]
add independent panner for internal (Aux) sends
Robin Gareus [Mon, 13 Jan 2014 09:48:23 +0000 (10:48 +0100)]
update internal-send port-count when target port-count changes
Robin Gareus [Mon, 13 Jan 2014 09:47:17 +0000 (10:47 +0100)]
ignore additional channels for AFL, PFL:
If the monitor-section has fewer-channels than the solo-listen point:
ignore additional channels.
Robin Gareus [Mon, 13 Jan 2014 09:46:11 +0000 (10:46 +0100)]
fix deadlock when removing monitoring-section
Robin Gareus [Sun, 12 Jan 2014 22:13:35 +0000 (23:13 +0100)]
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.
Robin Gareus [Sun, 12 Jan 2014 20:47:15 +0000 (21:47 +0100)]
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.
Robin Gareus [Sun, 12 Jan 2014 20:41:02 +0000 (21:41 +0100)]
VBAP panner UI tweaks:
* fix mouse-grab of sentinel
* make GUI more hemisphere like (circles at latitude)
* change alpha slightly to show signal overlap
Robin Gareus [Sun, 12 Jan 2014 17:11:14 +0000 (18:11 +0100)]
aubio 3+4 compat
Robin Gareus [Sun, 12 Jan 2014 16:14:14 +0000 (17:14 +0100)]
cosmetic change - properly use lilv API
Robin Gareus [Sun, 12 Jan 2014 02:56:35 +0000 (03:56 +0100)]
don't display elevation if panner does not support it.
Robin Gareus [Sun, 12 Jan 2014 02:36:20 +0000 (03:36 +0100)]
swap channel VBAP channel
Robin Gareus [Sun, 12 Jan 2014 01:05:01 +0000 (02:05 +0100)]
credit where credit is due for speaker-config and VBAP fixes inspiration
Robin Gareus [Sat, 11 Jan 2014 23:57:27 +0000 (00:57 +0100)]
VBAP GUI object-grab & position
Robin Gareus [Sat, 11 Jan 2014 22:34:19 +0000 (23:34 +0100)]
hide speaker-config in menu (it's not implemented yet)
Robin Gareus [Sat, 11 Jan 2014 22:30:13 +0000 (23:30 +0100)]
add spinbox to control width in VBAP GUI,..
Robin Gareus [Sat, 11 Jan 2014 22:30:01 +0000 (23:30 +0100)]
VBAP rework (part III): fix position computation backend & GUI
Robin Gareus [Sat, 11 Jan 2014 22:29:46 +0000 (23:29 +0100)]
update default speaker position list
Robin Gareus [Sat, 11 Jan 2014 22:29:36 +0000 (23:29 +0100)]
VBAP GUI depends on signal-position (not parameter changes)
Robin Gareus [Sat, 11 Jan 2014 22:29:23 +0000 (23:29 +0100)]
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()
)
Robin Gareus [Sat, 11 Jan 2014 21:59:25 +0000 (22:59 +0100)]
VBAP backend re-work (part one):
* fix azimuth, don't clamp but map to [0,1]
* prepare elevation (10+ speakers)
John Emmas [Mon, 13 Jan 2014 14:27:56 +0000 (14:27 +0000)]
Add provision for building 'libtimecode' as a DLL (MSVC build)
John Emmas [Mon, 13 Jan 2014 14:26:19 +0000 (14:26 +0000)]
'libs/timecode' - Change some instances of 'rint' to be 'lrint' (better compatibility with MSVC)
Robin Gareus [Mon, 13 Jan 2014 14:13:37 +0000 (15:13 +0100)]
allow to custom select panner-type for each delivery.
John Emmas [Mon, 13 Jan 2014 12:47:21 +0000 (12:47 +0000)]
'libs/evoral' - and add 'Note.impl' to the project
John Emmas [Mon, 13 Jan 2014 12:16:02 +0000 (12:16 +0000)]
'libs/evoral' - #include 'Note.impl' when building with MSVC (this is necessary to ensure that Evoral::Note gets exported from the evoral DLL)
Robin Gareus [Mon, 13 Jan 2014 11:02:00 +0000 (12:02 +0100)]
delete 'big' 2d panner window, if panner changes to mono|stereo
John Emmas [Mon, 13 Jan 2014 10:49:18 +0000 (10:49 +0000)]
Merge branch 'windows+cc' into cairocanvas
Robin Gareus [Mon, 13 Jan 2014 10:49:09 +0000 (11:49 +0100)]
fix delivery: check if panshell exists on session-load
John Emmas [Mon, 13 Jan 2014 10:22:31 +0000 (10:22 +0000)]
Check in some minor modifications (to MSVC specific headers)
Robin Gareus [Mon, 13 Jan 2014 10:08:51 +0000 (11:08 +0100)]
ProcessorWindowProxy::processor_going_away takes care of deletion
Robin Gareus [Mon, 13 Jan 2014 10:08:37 +0000 (11:08 +0100)]
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
Robin Gareus [Mon, 13 Jan 2014 09:50:49 +0000 (10:50 +0100)]
independent panning for external sends
Robin Gareus [Mon, 13 Jan 2014 09:50:22 +0000 (10:50 +0100)]
fix output metering for Sends (Aux and External)
Robin Gareus [Mon, 13 Jan 2014 09:48:37 +0000 (10:48 +0100)]
add independent panner for internal (Aux) sends
Robin Gareus [Mon, 13 Jan 2014 09:48:23 +0000 (10:48 +0100)]
update internal-send port-count when target port-count changes
Robin Gareus [Mon, 13 Jan 2014 09:47:17 +0000 (10:47 +0100)]
ignore additional channels for AFL, PFL:
If the monitor-section has fewer-channels than the solo-listen point:
ignore additional channels.
Robin Gareus [Mon, 13 Jan 2014 09:46:11 +0000 (10:46 +0100)]
fix deadlock when removing monitoring-section
Paul Davis [Mon, 13 Jan 2014 01:23:45 +0000 (20:23 -0500)]
use LIBEVORAL_TEMPLATE_API with Evoral::Note
Paul Davis [Mon, 13 Jan 2014 01:01:12 +0000 (20:01 -0500)]
fix 64 bit template types
Paul Davis [Mon, 13 Jan 2014 01:00:55 +0000 (20:00 -0500)]
add new file (required by MSVC) to sources list
Paul Davis [Mon, 13 Jan 2014 01:00:26 +0000 (20:00 -0500)]
use LIBEVORAL_TEMPLATE_API
Also clean up unnecessary comments - git blame provides complete per-line history, no need to leave comments behind
Paul Davis [Mon, 13 Jan 2014 00:59:40 +0000 (19:59 -0500)]
add LIBEVORAL_TEMPLATE_API to deal with gcc/msvc difference of opinion
Robin Gareus [Sun, 12 Jan 2014 22:13:35 +0000 (23:13 +0100)]
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.
Robin Gareus [Sun, 12 Jan 2014 20:47:15 +0000 (21:47 +0100)]
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.
Robin Gareus [Sun, 12 Jan 2014 20:41:02 +0000 (21:41 +0100)]
VBAP panner UI tweaks:
* fix mouse-grab of sentinel
* make GUI more hemisphere like (circles at latitude)
* change alpha slightly to show signal overlap
Paul Davis [Sun, 12 Jan 2014 20:39:58 +0000 (15:39 -0500)]
remove VAMP and Rubberband from source tree
Paul Davis [Sun, 12 Jan 2014 20:39:34 +0000 (15:39 -0500)]
add -fvisibility=default back to cxx/cflags when building QM DSP and VAMP plugins libs, because ... ugh
Paul Davis [Sun, 12 Jan 2014 20:38:49 +0000 (15:38 -0500)]
add export visibility macro
Paul Davis [Sun, 12 Jan 2014 20:38:29 +0000 (15:38 -0500)]
add new file
Paul Davis [Sun, 12 Jan 2014 19:50:23 +0000 (14:50 -0500)]
add missing export/visibility macro
Paul Davis [Sun, 12 Jan 2014 17:47:02 +0000 (12:47 -0500)]
remove taglib from source tree (require external provision)
Paul Davis [Sun, 12 Jan 2014 16:34:37 +0000 (11:34 -0500)]
move -fvisibility=hidden to the top of the source tree, and remove its internal use; use libtimecode as a shared lib again
Paul Davis [Sun, 12 Jan 2014 16:33:58 +0000 (11:33 -0500)]
add export/visibility macros to libtimecode and build as a shared library again
Paul Davis [Sun, 12 Jan 2014 16:33:16 +0000 (11:33 -0500)]
build libsmf as a static library, since it is private to evoral and would otherwise require export visibility control
This also removes -fvisibility=hidden, a change that is taking place across the source tree and will show up in a later commit elsewhere