ardour/libs/ardour/midi_track.cc

956 lines
25 KiB
C++
Raw Normal View History

/*
Copyright (C) 2006 Paul Davis
Author: David Robillard
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
#include <cmath>
#ifdef COMPILER_MSVC
#include <float.h>
// 'std::isinf()' and 'std::isnan()' are not available in MSVC.
#define isinf_local(val) !((bool)_finite((double)val))
#define isnan_local(val) (bool)_isnan((double)val)
#else
#define isinf_local std::isinf
#define isnan_local std::isnan
#endif
2013-07-17 06:03:34 -04:00
#include "pbd/ffs.h"
#include "pbd/enumwriter.h"
#include "pbd/convert.h"
#include "evoral/midi_util.h"
#include "ardour/buffer_set.h"
#include "ardour/debug.h"
#include "ardour/delivery.h"
#include "ardour/event_type_map.h"
#include "ardour/meter.h"
#include "ardour/midi_diskstream.h"
#include "ardour/midi_playlist.h"
#include "ardour/midi_port.h"
#include "ardour/midi_track.h"
#include "ardour/parameter_types.h"
#include "ardour/port.h"
#include "ardour/processor.h"
#include "ardour/session.h"
#include "ardour/session_playlists.h"
#include "ardour/utils.h"
#include "i18n.h"
namespace ARDOUR {
class InterThreadInfo;
class MidiSource;
class Region;
class SMFSource;
}
using namespace std;
using namespace ARDOUR;
using namespace PBD;
MidiTrack::MidiTrack (Session& sess, string name, Route::Flag flag, TrackMode mode)
: Track (sess, name, flag, mode, DataType::MIDI)
, _immediate_events(1024) // FIXME: size?
, _step_edit_ring_buffer(64) // FIXME: size?
, _note_mode(Sustained)
, _step_editing (false)
, _input_active (true)
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
, _playback_channel_mask(0x0000ffff)
, _capture_channel_mask(0x0000ffff)
{
}
MidiTrack::~MidiTrack ()
{
}
int
MidiTrack::init ()
{
if (Track::init ()) {
return -1;
}
_input->changed.connect_same_thread (*this, boost::bind (&MidiTrack::track_input_active, this, _1, _2));
return 0;
}
boost::shared_ptr<Diskstream>
MidiTrack::create_diskstream ()
{
2014-01-20 08:14:20 -05:00
MidiDiskstream::Flag dflags = MidiDiskstream::Flag (MidiDiskstream::Recordable);
assert(_mode != Destructive);
return boost::shared_ptr<Diskstream> (new MidiDiskstream (_session, name(), dflags));
}
void
MidiTrack::set_record_enabled (bool yn, void *src)
{
if (_step_editing) {
return;
}
Track::set_record_enabled (yn, src);
}
void
MidiTrack::set_diskstream (boost::shared_ptr<Diskstream> ds)
{
/* We have to do this here, as Track::set_diskstream will cause a buffer refill,
and the diskstream must be set up to fill its buffers using the correct _note_mode.
*/
boost::shared_ptr<MidiDiskstream> mds = boost::dynamic_pointer_cast<MidiDiskstream> (ds);
mds->set_note_mode (_note_mode);
Track::set_diskstream (ds);
mds->reset_tracker ();
_diskstream->set_track (this);
_diskstream->set_destructive (_mode == Destructive);
_diskstream->set_record_enabled (false);
_diskstream_data_recorded_connection.disconnect ();
mds->DataRecorded.connect_same_thread (
_diskstream_data_recorded_connection,
boost::bind (&MidiTrack::diskstream_data_recorded, this, _1));
DiskstreamChanged (); /* EMIT SIGNAL */
}
boost::shared_ptr<MidiDiskstream>
MidiTrack::midi_diskstream() const
{
return boost::dynamic_pointer_cast<MidiDiskstream>(_diskstream);
}
int
MidiTrack::set_state (const XMLNode& node, int version)
{
const XMLProperty *prop;
/* This must happen before Track::set_state(), as there will be a buffer
fill during that call, and we must fill buffers using the correct
_note_mode.
*/
if ((prop = node.property (X_("note-mode"))) != 0) {
_note_mode = NoteMode (string_2_enum (prop->value(), _note_mode));
} else {
_note_mode = Sustained;
}
if (Track::set_state (node, version)) {
return -1;
}
// No destructive MIDI tracks (yet?)
_mode = Normal;
if ((prop = node.property ("input-active")) != 0) {
set_input_active (string_is_affirmative (prop->value()));
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
ChannelMode playback_channel_mode = AllChannels;
ChannelMode capture_channel_mode = AllChannels;
if ((prop = node.property ("playback-channel-mode")) != 0) {
playback_channel_mode = ChannelMode (string_2_enum(prop->value(), playback_channel_mode));
}
if ((prop = node.property ("capture-channel-mode")) != 0) {
capture_channel_mode = ChannelMode (string_2_enum(prop->value(), capture_channel_mode));
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
}
if ((prop = node.property ("channel-mode")) != 0) {
/* 3.0 behaviour where capture and playback modes were not separated */
playback_channel_mode = ChannelMode (string_2_enum(prop->value(), playback_channel_mode));
capture_channel_mode = playback_channel_mode;
}
unsigned int playback_channel_mask = 0xffff;
unsigned int capture_channel_mask = 0xffff;
if ((prop = node.property ("playback-channel-mask")) != 0) {
sscanf (prop->value().c_str(), "0x%x", &playback_channel_mask);
}
if ((prop = node.property ("capture-channel-mask")) != 0) {
sscanf (prop->value().c_str(), "0x%x", &capture_channel_mask);
}
if ((prop = node.property ("channel-mask")) != 0) {
sscanf (prop->value().c_str(), "0x%x", &playback_channel_mask);
capture_channel_mask = playback_channel_mask;
}
set_playback_channel_mode (playback_channel_mode, playback_channel_mask);
set_capture_channel_mode (capture_channel_mode, capture_channel_mask);
pending_state = const_cast<XMLNode*> (&node);
if (_session.state_of_the_state() & Session::Loading) {
_session.StateReady.connect_same_thread (
*this, boost::bind (&MidiTrack::set_state_part_two, this));
} else {
set_state_part_two ();
}
return 0;
}
XMLNode&
MidiTrack::state(bool full_state)
{
XMLNode& root (Track::state(full_state));
XMLNode* freeze_node;
char buf[64];
if (_freeze_record.playlist) {
XMLNode* inode;
freeze_node = new XMLNode (X_("freeze-info"));
freeze_node->add_property ("playlist", _freeze_record.playlist->name());
freeze_node->add_property ("state", enum_2_string (_freeze_record.state));
for (vector<FreezeRecordProcessorInfo*>::iterator i = _freeze_record.processor_info.begin(); i != _freeze_record.processor_info.end(); ++i) {
inode = new XMLNode (X_("processor"));
(*i)->id.print (buf, sizeof(buf));
inode->add_property (X_("id"), buf);
inode->add_child_copy ((*i)->state);
freeze_node->add_child_nocopy (*inode);
}
root.add_child_nocopy (*freeze_node);
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
root.add_property("playback_channel-mode", enum_2_string(get_playback_channel_mode()));
root.add_property("capture_channel-mode", enum_2_string(get_capture_channel_mode()));
snprintf (buf, sizeof(buf), "0x%x", get_playback_channel_mask());
root.add_property("playback-channel-mask", buf);
snprintf (buf, sizeof(buf), "0x%x", get_capture_channel_mask());
root.add_property("capture-channel-mask", buf);
root.add_property ("note-mode", enum_2_string (_note_mode));
root.add_property ("step-editing", (_step_editing ? "yes" : "no"));
root.add_property ("input-active", (_input_active ? "yes" : "no"));
return root;
}
void
MidiTrack::set_state_part_two ()
{
XMLNode* fnode;
XMLProperty* prop;
LocaleGuard lg (X_("C"));
/* This is called after all session state has been restored but before
have been made ports and connections are established.
*/
if (pending_state == 0) {
return;
}
if ((fnode = find_named_node (*pending_state, X_("freeze-info"))) != 0) {
_freeze_record.state = Frozen;
for (vector<FreezeRecordProcessorInfo*>::iterator i = _freeze_record.processor_info.begin(); i != _freeze_record.processor_info.end(); ++i) {
delete *i;
}
_freeze_record.processor_info.clear ();
if ((prop = fnode->property (X_("playlist"))) != 0) {
boost::shared_ptr<Playlist> pl = _session.playlists->by_name (prop->value());
if (pl) {
_freeze_record.playlist = boost::dynamic_pointer_cast<MidiPlaylist> (pl);
} else {
_freeze_record.playlist.reset();
_freeze_record.state = NoFreeze;
return;
}
}
if ((prop = fnode->property (X_("state"))) != 0) {
_freeze_record.state = FreezeState (string_2_enum (prop->value(), _freeze_record.state));
}
XMLNodeConstIterator citer;
XMLNodeList clist = fnode->children();
for (citer = clist.begin(); citer != clist.end(); ++citer) {
if ((*citer)->name() != X_("processor")) {
continue;
}
if ((prop = (*citer)->property (X_("id"))) == 0) {
continue;
}
FreezeRecordProcessorInfo* frii = new FreezeRecordProcessorInfo (*((*citer)->children().front()),
boost::shared_ptr<Processor>());
frii->id = prop->value ();
_freeze_record.processor_info.push_back (frii);
}
}
if (midi_diskstream ()) {
midi_diskstream()->set_block_size (_session.get_block_size ());
}
return;
}
/** @param need_butler to be set to true if this track now needs the butler, otherwise it can be left alone
* or set to false.
*/
int
MidiTrack::roll (pframes_t nframes, framepos_t start_frame, framepos_t end_frame, int declick, bool& need_butler)
{
Glib::Threads::RWLock::ReaderLock lm (_processor_lock, Glib::Threads::TRY_LOCK);
if (!lm.locked()) {
boost::shared_ptr<MidiDiskstream> diskstream = midi_diskstream();
framecnt_t playback_distance = diskstream->calculate_playback_distance(nframes);
if (can_internal_playback_seek(llabs(playback_distance))) {
/* TODO should declick, and/or note-off */
internal_playback_seek(playback_distance);
}
return 0;
}
boost::shared_ptr<MidiDiskstream> diskstream = midi_diskstream();
if (n_outputs().n_total() == 0 && _processors.empty()) {
return 0;
}
if (!_active) {
silence (nframes);
2013-07-16 13:36:57 -04:00
if (_meter_point == MeterInput && (_monitoring & MonitorInput || _diskstream->record_enabled())) {
_meter->reset();
}
return 0;
}
framepos_t transport_frame = _session.transport_frame();
int dret;
framecnt_t playback_distance;
if ((nframes = check_initial_delay (nframes, transport_frame)) == 0) {
/* need to do this so that the diskstream sets its
playback distance to zero, thus causing diskstream::commit
to do nothing.
*/
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
BufferSet bufs; /* empty set - is OK, since nothing will happen */
dret = diskstream->process (bufs, transport_frame, 0, playback_distance, false);
need_butler = diskstream->commit (playback_distance);
return dret;
}
2013-07-31 08:21:40 -04:00
BufferSet& bufs = _session.get_route_buffers (n_process_buffers());
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
fill_buffers_with_input (bufs, _input, nframes);
/* filter captured data before meter sees it */
filter_channels (bufs, get_capture_channel_mode(), get_capture_channel_mask());
if (_meter_point == MeterInput && (_monitoring & MonitorInput || _diskstream->record_enabled())) {
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
_meter->run (bufs, start_frame, end_frame, nframes, true);
}
_silent = false;
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
if ((dret = diskstream->process (bufs, transport_frame, nframes, playback_distance, (monitoring_state() == MonitoringDisk))) != 0) {
need_butler = diskstream->commit (playback_distance);
silence (nframes);
return dret;
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
/* filter playback data before we do anything else */
filter_channels (bufs, get_playback_channel_mode(), get_playback_channel_mask ());
if (monitoring_state() == MonitoringInput) {
/* not actually recording, but we want to hear the input material anyway,
at least potentially (depending on monitoring options)
*/
/* because the playback buffer is event based and not a
* continuous stream, we need to make sure that we empty
* it of events every cycle to avoid it filling up with events
* read from disk, while we are actually monitoring input
*/
diskstream->flush_playback (start_frame, end_frame);
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
/* append immediate messages to the first MIDI buffer (thus sending it to the first output port) */
write_out_of_band_data (bufs, start_frame, end_frame, nframes);
/* final argument: don't waste time with automation if we're not recording or rolling */
process_output_buffers (bufs, start_frame, end_frame, nframes,
declick, (!diskstream->record_enabled() && !_session.transport_stopped()));
for (ProcessorList::iterator i = _processors.begin(); i != _processors.end(); ++i) {
boost::shared_ptr<Delivery> d = boost::dynamic_pointer_cast<Delivery> (*i);
if (d) {
d->flush_buffers (nframes);
}
}
need_butler = diskstream->commit (playback_distance);
return 0;
}
int
MidiTrack::no_roll (pframes_t nframes, framepos_t start_frame, framepos_t end_frame, bool state_changing)
{
int ret = Track::no_roll (nframes, start_frame, end_frame, state_changing);
if (ret == 0 && _step_editing) {
push_midi_input_to_step_edit_ringbuffer (nframes);
}
return ret;
}
void
MidiTrack::realtime_locate ()
{
Glib::Threads::RWLock::ReaderLock lm (_processor_lock, Glib::Threads::TRY_LOCK);
if (!lm.locked ()) {
return;
}
for (ProcessorList::iterator i = _processors.begin(); i != _processors.end(); ++i) {
(*i)->realtime_locate ();
}
midi_diskstream()->reset_tracker ();
}
void
MidiTrack::realtime_handle_transport_stopped ()
{
Glib::Threads::RWLock::ReaderLock lm (_processor_lock, Glib::Threads::TRY_LOCK);
if (!lm.locked ()) {
return;
}
for (ProcessorList::iterator i = _processors.begin(); i != _processors.end(); ++i) {
(*i)->realtime_handle_transport_stopped ();
}
}
void
MidiTrack::push_midi_input_to_step_edit_ringbuffer (framecnt_t nframes)
{
PortSet& ports (_input->ports());
for (PortSet::iterator p = ports.begin(DataType::MIDI); p != ports.end(DataType::MIDI); ++p) {
Buffer& b (p->get_buffer (nframes));
const MidiBuffer* const mb = dynamic_cast<MidiBuffer*>(&b);
assert (mb);
for (MidiBuffer::const_iterator e = mb->begin(); e != mb->end(); ++e) {
const Evoral::MIDIEvent<framepos_t> ev(*e, false);
/* note on, since for step edit, note length is determined
elsewhere
*/
if (ev.is_note_on()) {
/* we don't care about the time for this purpose */
_step_edit_ring_buffer.write (0, ev.type(), ev.size(), ev.buffer());
}
}
}
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
void
MidiTrack::filter_channels (BufferSet& bufs, ChannelMode mode, uint32_t mask)
{
if (mode == AllChannels) {
return;
}
MidiBuffer& buf (bufs.get_midi (0));
for (MidiBuffer::iterator e = buf.begin(); e != buf.end(); ) {
Evoral::MIDIEvent<framepos_t> ev(*e, false);
if (ev.is_channel_event()) {
switch (mode) {
case FilterChannels:
if (0 == ((1<<ev.channel()) & mask)) {
e = buf.erase (e);
} else {
++e;
}
break;
case ForceChannel:
2013-07-17 06:03:34 -04:00
ev.set_channel (PBD::ffs (mask) - 1);
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
++e;
break;
case AllChannels:
/* handled by the opening if() */
++e;
break;
}
} else {
++e;
}
}
}
void
MidiTrack::write_out_of_band_data (BufferSet& bufs, framepos_t /*start*/, framepos_t /*end*/, framecnt_t nframes)
{
MidiBuffer& buf (bufs.get_midi (0));
// Append immediate events
if (_immediate_events.read_space()) {
DEBUG_TRACE (DEBUG::MidiIO, string_compose ("%1 has %2 of immediate events to deliver\n",
name(), _immediate_events.read_space()));
/* write as many of the immediate events as we can, but give "true" as
* the last argument ("stop on overflow in destination") so that we'll
* ship the rest out next time.
*
* the Port::port_offset() + (nframes-1) argument puts all these events at the last
* possible position of the output buffer, so that we do not
* violate monotonicity when writing. Port::port_offset() will
* be non-zero if we're in a split process cycle.
*/
_immediate_events.read (buf, 0, 1, Port::port_offset() + nframes - 1, true);
}
}
int
2014-12-17 21:48:09 -05:00
MidiTrack::export_stuff (BufferSet& buffers,
framepos_t start,
framecnt_t nframes,
boost::shared_ptr<Processor> endpoint,
bool include_endpoint,
bool for_export,
bool for_freeze)
{
2014-12-17 21:48:09 -05:00
if (buffers.count().n_midi() == 0) {
return -1;
}
boost::shared_ptr<MidiDiskstream> diskstream = midi_diskstream();
Glib::Threads::RWLock::ReaderLock rlock (_processor_lock);
boost::shared_ptr<MidiPlaylist> mpl = boost::dynamic_pointer_cast<MidiPlaylist>(diskstream->playlist());
if (!mpl) {
return -2;
}
buffers.get_midi(0).clear();
if (mpl->read(buffers.get_midi(0), start, nframes, 0) != nframes) {
return -1;
}
//bounce_process (buffers, start, nframes, endpoint, include_endpoint, for_export, for_freeze);
return 0;
}
boost::shared_ptr<Region>
2014-12-17 21:48:09 -05:00
MidiTrack::bounce (InterThreadInfo& itt)
{
2014-12-17 21:48:09 -05:00
return bounce_range (_session.current_start_frame(), _session.current_end_frame(), itt, main_outs(), false);
}
boost::shared_ptr<Region>
2014-12-17 21:48:09 -05:00
MidiTrack::bounce_range (framepos_t start,
framepos_t end,
InterThreadInfo& itt,
boost::shared_ptr<Processor> endpoint,
bool include_endpoint)
{
2014-12-17 21:48:09 -05:00
vector<boost::shared_ptr<Source> > srcs;
return _session.write_one_track (*this, start, end, false, srcs, itt, endpoint, include_endpoint, false, false);
}
void
MidiTrack::freeze_me (InterThreadInfo& /*itt*/)
{
std::cerr << "MIDI freeze currently unsupported" << std::endl;
}
void
MidiTrack::unfreeze ()
{
_freeze_record.state = UnFrozen;
FreezeChange (); /* EMIT SIGNAL */
}
void
MidiTrack::set_note_mode (NoteMode m)
{
_note_mode = m;
midi_diskstream()->set_note_mode(m);
}
std::string
MidiTrack::describe_parameter (Evoral::Parameter param)
{
const std::string str(instrument_info().get_controller_name(param));
return str.empty() ? Automatable::describe_parameter(param) : str;
}
void
MidiTrack::midi_panic()
{
DEBUG_TRACE (DEBUG::MidiIO, string_compose ("%1 delivers panic data\n", name()));
for (uint8_t channel = 0; channel <= 0xF; channel++) {
uint8_t ev[3] = { ((uint8_t) (MIDI_CMD_CONTROL | channel)), ((uint8_t) MIDI_CTL_SUSTAIN), 0 };
write_immediate_event(3, ev);
ev[1] = MIDI_CTL_ALL_NOTES_OFF;
write_immediate_event(3, ev);
ev[1] = MIDI_CTL_RESET_CONTROLLERS;
write_immediate_event(3, ev);
}
}
/** \return true on success, false on failure (no buffer space left)
*/
bool
MidiTrack::write_immediate_event(size_t size, const uint8_t* buf)
{
if (!Evoral::midi_event_is_valid(buf, size)) {
cerr << "WARNING: Ignoring illegal immediate MIDI event" << endl;
return false;
}
const uint32_t type = midi_parameter_type(buf[0]);
return (_immediate_events.write (0, type, size, buf) == size);
}
void
MidiTrack::set_parameter_automation_state (Evoral::Parameter param, AutoState state)
{
switch (param.type()) {
case MidiCCAutomation:
case MidiPgmChangeAutomation:
case MidiPitchBenderAutomation:
case MidiChannelPressureAutomation:
case MidiSystemExclusiveAutomation:
/* The track control for MIDI parameters is for immediate events to act
as a control surface, write/touch for them is not currently
supported. */
return;
default:
Automatable::set_parameter_automation_state(param, state);
}
}
void
MidiTrack::MidiControl::set_value(double val)
{
const Evoral::Parameter &parameter = _list ? _list->parameter() : Control::parameter();
const Evoral::ParameterDescriptor &desc = EventTypeMap::instance().descriptor(parameter);
bool valid = false;
if (isinf_local(val)) {
cerr << "MIDIControl value is infinity" << endl;
} else if (isnan_local(val)) {
cerr << "MIDIControl value is NaN" << endl;
} else if (val < desc.lower) {
cerr << "MIDIControl value is < " << desc.lower << endl;
} else if (val > desc.upper) {
cerr << "MIDIControl value is > " << desc.upper << endl;
} else {
valid = true;
}
if (!valid) {
return;
}
assert(val <= desc.upper);
if ( ! _list || ! automation_playback()) {
size_t size = 3;
uint8_t ev[3] = { parameter.channel(), uint8_t (val), 0 };
switch(parameter.type()) {
case MidiCCAutomation:
ev[0] += MIDI_CMD_CONTROL;
ev[1] = parameter.id();
ev[2] = int(val);
break;
case MidiPgmChangeAutomation:
size = 2;
ev[0] += MIDI_CMD_PGM_CHANGE;
ev[1] = int(val);
break;
case MidiChannelPressureAutomation:
size = 2;
ev[0] += MIDI_CMD_CHANNEL_PRESSURE;
ev[1] = int(val);
break;
case MidiPitchBenderAutomation:
ev[0] += MIDI_CMD_BENDER;
ev[1] = 0x7F & int(val);
ev[2] = 0x7F & (int(val) >> 7);
break;
default:
assert(false);
}
_route->write_immediate_event(size, ev);
}
AutomationControl::set_value(val);
}
void
MidiTrack::set_step_editing (bool yn)
{
if (_session.record_status() != Session::Disabled) {
return;
}
if (yn != _step_editing) {
_step_editing = yn;
StepEditStatusChange (yn);
}
}
boost::shared_ptr<SMFSource>
MidiTrack::write_source (uint32_t)
{
return midi_diskstream()->write_source ();
}
void
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
MidiTrack::set_playback_channel_mode(ChannelMode mode, uint16_t mask)
{
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
ChannelMode old = get_playback_channel_mode ();
uint16_t old_mask = get_playback_channel_mask ();
if (old != mode || mask != old_mask) {
_set_playback_channel_mode (mode, mask);
PlaybackChannelModeChanged ();
_session.set_dirty ();
}
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
void
MidiTrack::set_capture_channel_mode(ChannelMode mode, uint16_t mask)
{
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
ChannelMode old = get_capture_channel_mode ();
uint16_t old_mask = get_capture_channel_mask ();
if (old != mode || mask != old_mask) {
_set_capture_channel_mode (mode, mask);
CaptureChannelModeChanged ();
_session.set_dirty ();
}
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
void
MidiTrack::set_playback_channel_mask (uint16_t mask)
{
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
uint16_t old = get_playback_channel_mask();
if (old != mask) {
_set_playback_channel_mask (mask);
PlaybackChannelMaskChanged ();
_session.set_dirty ();
}
}
void
MidiTrack::set_capture_channel_mask (uint16_t mask)
{
uint16_t old = get_capture_channel_mask();
if (old != mask) {
_set_capture_channel_mask (mask);
CaptureChannelMaskChanged ();
_session.set_dirty ();
}
}
boost::shared_ptr<MidiPlaylist>
MidiTrack::midi_playlist ()
{
return midi_diskstream()->midi_playlist ();
}
void
MidiTrack::diskstream_data_recorded (boost::weak_ptr<MidiSource> src)
{
DataRecorded (src); /* EMIT SIGNAL */
}
bool
MidiTrack::input_active () const
{
return _input_active;
}
void
MidiTrack::set_input_active (bool yn)
{
if (yn != _input_active) {
_input_active = yn;
map_input_active (yn);
InputActiveChanged (); /* EMIT SIGNAL */
}
}
void
MidiTrack::map_input_active (bool yn)
{
if (!_input) {
return;
}
PortSet& ports (_input->ports());
for (PortSet::iterator p = ports.begin(DataType::MIDI); p != ports.end(DataType::MIDI); ++p) {
boost::shared_ptr<MidiPort> mp = boost::dynamic_pointer_cast<MidiPort> (*p);
if (yn != mp->input_active()) {
mp->set_input_active (yn);
}
}
}
void
MidiTrack::track_input_active (IOChange change, void* /* src */)
{
if (change.type & IOChange::ConfigurationChanged) {
map_input_active (_input_active);
}
}
boost::shared_ptr<Diskstream>
MidiTrack::diskstream_factory (XMLNode const & node)
{
return boost::shared_ptr<Diskstream> (new MidiDiskstream (_session, node));
}
boost::shared_ptr<MidiBuffer>
MidiTrack::get_gui_feed_buffer () const
{
return midi_diskstream()->get_gui_feed_buffer ();
}
void
MidiTrack::act_on_mute ()
{
/* this is called right after our mute status has changed.
if we are now muted, send suitable output to shutdown
all our notes.
XXX we should should also stop all relevant note trackers.
*/
/* If we haven't got a diskstream yet, there's nothing to worry about,
and we can't call get_channel_mask() anyway.
*/
if (!midi_diskstream()) {
return;
}
if (muted() || _mute_master->muted_by_others_at(MuteMaster::AllPoints)) {
/* only send messages for channels we are using */
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00
uint16_t mask = get_playback_channel_mask();
for (uint8_t channel = 0; channel <= 0xF; channel++) {
if ((1<<channel) & mask) {
DEBUG_TRACE (DEBUG::MidiIO, string_compose ("%1 delivers mute message to channel %2\n", name(), channel+1));
uint8_t ev[3] = { ((uint8_t) (MIDI_CMD_CONTROL | channel)), MIDI_CTL_SUSTAIN, 0 };
write_immediate_event (3, ev);
/* Note we do not send MIDI_CTL_ALL_NOTES_OFF here, since this may
silence notes that came from another non-muted track. */
}
}
/* Resolve active notes. */
midi_diskstream()->resolve_tracker(_immediate_events, 0);
}
}
void
MidiTrack::set_monitoring (MonitorChoice mc)
{
if (mc != _monitoring) {
Track::set_monitoring (mc);
/* monitoring state changed, so flush out any on notes at the
* port level.
*/
PortSet& ports (_output->ports());
for (PortSet::iterator p = ports.begin(); p != ports.end(); ++p) {
boost::shared_ptr<MidiPort> mp = boost::dynamic_pointer_cast<MidiPort> (*p);
if (mp) {
mp->require_resolve ();
}
}
boost::shared_ptr<MidiDiskstream> md (midi_diskstream());
if (md) {
md->reset_tracker ();
}
}
}
MonitorState
MidiTrack::monitoring_state () const
{
MonitorState ms = Track::monitoring_state();
if (ms == MonitoringSilence) {
return MonitoringInput;
}
return ms;
}
Squashed commit of the following: commit fdbae82077db53add90df7448a06869dac89acc6 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 21:45:28 2013 -0400 mammoth changes in basic signal flow, total redesign of MIDI channel filtering and more. commit 59343a8283698e02bc0f622313b29e98f449e4c8 Author: Paul Davis <paul@linuxaudiosystems.com> Date: Wed Mar 27 01:58:53 2013 -0400 initial working version after changes to MIDI channel filtering. may affect metering input too. testing not yet finished this commit merges many deep changes in ardour's internal architecture, combined with a total redesign of how MIDI channel filtering works. data in a track used to flow from JACK port buffers to diskstream's ringbuffers and was then copied from the ringbuffers into a BufferSet for use during Route::process_output_buffers(). The butler thread would handle the movement of data between the ringbuffers and disk. with this commit, data now flows from JACK port buffers into the BufferSet used for Route processing, and is copied from the BufferSet into the diskstream's ringbuffers (the butler thread continues to handle interactions with disk as usual). this change allowed a dramatic consolidation of code and simplification of most aspects of Track/Route::roll() and Track/Route::no_roll(). in particular, see Route::fill_buffers_with_input() which now concisely describes how we move data from JACK port buffers into the BufferSet for all Route types (including Tracks). this work was initially motivated by changing MIDI channel filtering so that we can process capture and playback independently. there is now a very clean pathway for this - see MidiTrack::roll() (NOTE: This needs implementing in the no-roll case too - a TODO item). the channel selector for MIDI tracks has been moved out of the track header and is now accessible via the context menu. more work is likely here, to make it (more) obvious to the user when filtering is going on.
2013-03-27 21:50:18 -04:00