mirror of
https://github.com/cookiengineer/audacity
synced 2025-05-04 17:49:45 +02:00
908 lines
30 KiB
C++
908 lines
30 KiB
C++
/**********************************************************************
|
|
|
|
Audacity: A Digital Audio Editor
|
|
|
|
NoteTrack.cpp
|
|
|
|
Dominic Mazzoni
|
|
|
|
*******************************************************************//*!
|
|
|
|
\class NoteTrack
|
|
\brief A Track that is used for Midi notes. (Somewhat old code).
|
|
|
|
*//*******************************************************************/
|
|
|
|
|
|
#include "Audacity.h"
|
|
#include "NoteTrack.h"
|
|
|
|
#include <wx/dc.h>
|
|
#include <wx/brush.h>
|
|
#include <wx/pen.h>
|
|
#include <wx/intl.h>
|
|
|
|
#if defined(USE_MIDI)
|
|
#include <sstream>
|
|
|
|
#define ROUND(x) ((int) ((x) + 0.5))
|
|
|
|
#include "AColor.h"
|
|
#include "DirManager.h"
|
|
#include "Internat.h"
|
|
#include "Prefs.h"
|
|
#include "effects/TimeWarper.h"
|
|
|
|
#include "Experimental.h"
|
|
|
|
#ifdef SONIFY
|
|
#include "portmidi.h"
|
|
|
|
#define SON_PROGRAM 0
|
|
#define SON_AutoSave 67
|
|
#define SON_ModifyState 60
|
|
#define SON_NoteBackground 72
|
|
#define SON_NoteForeground 74
|
|
#define SON_Measures 76 /* "bar line" */
|
|
#define SON_Serialize 77
|
|
#define SON_Unserialize 79
|
|
#define SON_VEL 100
|
|
|
|
|
|
PmStream *sonMidiStream;
|
|
bool sonificationStarted = false;
|
|
|
|
void SonifyBeginSonification()
|
|
{
|
|
PmError err = Pm_OpenOutput(&sonMidiStream, Pm_GetDefaultOutputDeviceID(),
|
|
NULL, 0, NULL, NULL, 0);
|
|
if (err) sonMidiStream = NULL;
|
|
if (sonMidiStream)
|
|
Pm_WriteShort(sonMidiStream, 0, Pm_Message(0xC0, SON_PROGRAM, 0));
|
|
sonificationStarted = true;
|
|
}
|
|
|
|
|
|
void SonifyEndSonification()
|
|
{
|
|
if (sonMidiStream) Pm_Close(sonMidiStream);
|
|
sonificationStarted = false;
|
|
}
|
|
|
|
|
|
|
|
|
|
void SonifyNoteOnOff(int p, int v)
|
|
{
|
|
if (!sonificationStarted)
|
|
SonifyBeginSonification();
|
|
if (sonMidiStream)
|
|
Pm_WriteShort(sonMidiStream, 0, Pm_Message(0x90, p, v));
|
|
}
|
|
|
|
#define SONFNS(name) \
|
|
void SonifyBegin ## name() { SonifyNoteOnOff(SON_ ## name, SON_VEL); } \
|
|
void SonifyEnd ## name() { SonifyNoteOnOff(SON_ ## name, 0); }
|
|
|
|
SONFNS(NoteBackground)
|
|
SONFNS(NoteForeground)
|
|
SONFNS(Measures)
|
|
SONFNS(Serialize)
|
|
SONFNS(Unserialize)
|
|
SONFNS(ModifyState)
|
|
SONFNS(AutoSave)
|
|
|
|
#undef SONFNS
|
|
|
|
#endif
|
|
|
|
|
|
|
|
NoteTrack::Holder TrackFactory::NewNoteTrack()
|
|
{
|
|
return std::make_unique<NoteTrack>(mDirManager);
|
|
}
|
|
|
|
NoteTrack::NoteTrack(const std::shared_ptr<DirManager> &projDirManager):
|
|
Track(projDirManager)
|
|
{
|
|
SetDefaultName(_("Note Track"));
|
|
SetName(GetDefaultName());
|
|
|
|
mSeq = NULL;
|
|
mSerializationBuffer = NULL;
|
|
mSerializationLength = 0;
|
|
|
|
#ifdef EXPERIMENTAL_MIDI_OUT
|
|
mGain = 0;
|
|
#endif
|
|
mBottomNote = 24;
|
|
mPitchHeight = 5;
|
|
|
|
mVisibleChannels = ALL_CHANNELS;
|
|
mLastMidiPosition = 0;
|
|
}
|
|
|
|
NoteTrack::~NoteTrack()
|
|
{
|
|
if (mSerializationBuffer) {
|
|
delete [] mSerializationBuffer;
|
|
}
|
|
}
|
|
|
|
Track::Holder NoteTrack::Duplicate() const
|
|
{
|
|
auto duplicate = std::make_unique<NoteTrack>(mDirManager);
|
|
duplicate->Init(*this);
|
|
// Duplicate on NoteTrack moves data from mSeq to mSerializationBuffer
|
|
// and from mSerializationBuffer to mSeq on alternate calls. Duplicate
|
|
// to the undo stack and Duplicate back to the project should result
|
|
// in serialized blobs on the undo stack and traversable data in the
|
|
// project object.
|
|
if (mSeq) {
|
|
SonifyBeginSerialize();
|
|
assert(!mSerializationBuffer);
|
|
// serialize from this to duplicate's mSerializationBuffer
|
|
mSeq->serialize((void**)&duplicate->mSerializationBuffer,
|
|
&duplicate->mSerializationLength);
|
|
SonifyEndSerialize();
|
|
} else if (mSerializationBuffer) {
|
|
SonifyBeginUnserialize();
|
|
assert(!mSeq);
|
|
std::unique_ptr<Alg_track> alg_track{ Alg_seq::unserialize(mSerializationBuffer,
|
|
mSerializationLength) };
|
|
assert(alg_track->get_type() == 's');
|
|
duplicate->mSeq.reset(static_cast<Alg_seq*>(alg_track.release()));
|
|
SonifyEndUnserialize();
|
|
} else assert(false); // bug if neither mSeq nor mSerializationBuffer
|
|
// copy some other fields here
|
|
duplicate->SetBottomNote(mBottomNote);
|
|
duplicate->SetPitchHeight(mPitchHeight);
|
|
duplicate->mLastMidiPosition = mLastMidiPosition;
|
|
duplicate->mVisibleChannels = mVisibleChannels;
|
|
duplicate->SetOffset(GetOffset());
|
|
#ifdef EXPERIMENTAL_MIDI_OUT
|
|
duplicate->SetGain(GetGain());
|
|
#endif
|
|
// This std::move is needed to "upcast" the pointer type
|
|
return std::move(duplicate);
|
|
}
|
|
|
|
|
|
double NoteTrack::GetOffset() const
|
|
{
|
|
return mOffset;
|
|
}
|
|
|
|
double NoteTrack::GetStartTime() const
|
|
{
|
|
return GetOffset();
|
|
}
|
|
|
|
double NoteTrack::GetEndTime() const
|
|
{
|
|
return GetStartTime() + (mSeq ? mSeq->get_real_dur() : 0.0);
|
|
}
|
|
|
|
|
|
void NoteTrack::WarpAndTransposeNotes(double t0, double t1,
|
|
const TimeWarper &warper,
|
|
double semitones)
|
|
{
|
|
// Since this is a duplicate and duplicates convert mSeq to
|
|
// a text string for saving as XML, we probably have to
|
|
// duplicate again to get back an mSeq
|
|
double offset = this->GetOffset(); // track is shifted this amount
|
|
if (!mSeq) { // replace saveme with an (unserialized) duplicate
|
|
Track::Holder unt{ Duplicate() };
|
|
const auto nt = static_cast<NoteTrack*>(unt.get());
|
|
wxASSERT(!mSeq && nt->mSeq && !nt->mSerializationBuffer);
|
|
// swap mSeq and Buffer between this and nt
|
|
nt->mSerializationBuffer = mSerializationBuffer;
|
|
nt->mSerializationLength = mSerializationLength;
|
|
mSerializationBuffer = NULL;
|
|
mSerializationLength = 0;
|
|
mSeq = std::move(nt->mSeq);
|
|
}
|
|
mSeq->convert_to_seconds(); // make sure time units are right
|
|
t1 -= offset; // adjust time range to compensate for track offset
|
|
t0 -= offset;
|
|
if (t1 > mSeq->get_dur()) { // make sure t0, t1 are within sequence
|
|
t1 = mSeq->get_dur();
|
|
if (t0 >= t1) return;
|
|
}
|
|
Alg_iterator iter(mSeq.get(), false);
|
|
iter.begin();
|
|
Alg_event_ptr event;
|
|
while (0 != (event = iter.next()) && event->time < t1) {
|
|
if (event->is_note() && event->time >= t0 &&
|
|
// Allegro data structure does not restrict channels to 16.
|
|
// Since there is not way to select more than 16 channels,
|
|
// map all channel numbers mod 16. This will have no effect
|
|
// on MIDI files, but it will allow users to at least select
|
|
// all channels on non-MIDI event sequence data.
|
|
IsVisibleChan(event->chan % 16)) {
|
|
event->set_pitch(event->get_pitch() + semitones);
|
|
}
|
|
}
|
|
iter.end();
|
|
// now, use warper to warp the tempo map
|
|
mSeq->convert_to_beats(); // beats remain the same
|
|
Alg_time_map_ptr map = mSeq->get_time_map();
|
|
map->insert_beat(t0, map->time_to_beat(t0));
|
|
map->insert_beat(t1, map->time_to_beat(t1));
|
|
int i, len = map->length();
|
|
for (i = 0; i < len; i++) {
|
|
Alg_beat &beat = map->beats[i];
|
|
beat.time = warper.Warp(beat.time + offset) - offset;
|
|
}
|
|
// about to redisplay, so might as well convert back to time now
|
|
mSeq->convert_to_seconds();
|
|
}
|
|
|
|
|
|
|
|
int NoteTrack::DrawLabelControls(wxDC & dc, wxRect & r)
|
|
{
|
|
int wid = 23;
|
|
int ht = 16;
|
|
|
|
if (r.height < ht * 4) {
|
|
return r.y + 5 + ht * 4;
|
|
}
|
|
|
|
int x = r.x + (r.width / 2 - wid * 2) + 2;
|
|
int y = r.y + 5;
|
|
|
|
wxRect box;
|
|
for (int row = 0; row < 4; row++) {
|
|
for (int col = 0; col < 4; col++) {
|
|
// chanName is the "external" channel number (1-16)
|
|
// used by AColor and button labels
|
|
int chanName = row * 4 + col + 1;
|
|
|
|
box.x = x + col * wid;
|
|
box.y = y + row * ht;
|
|
box.width = wid;
|
|
box.height = ht;
|
|
|
|
if (IsVisibleChan(chanName - 1)) {
|
|
AColor::MIDIChannel(&dc, chanName);
|
|
dc.DrawRectangle(box);
|
|
// two choices: channel is enabled (to see and play) when button is in
|
|
// "up" position (original Audacity style) or in "down" position
|
|
//
|
|
#define CHANNEL_ON_IS_DOWN 1
|
|
#if CHANNEL_ON_IS_DOWN
|
|
AColor::DarkMIDIChannel(&dc, chanName);
|
|
#else
|
|
AColor::LightMIDIChannel(&dc, chanName);
|
|
#endif
|
|
AColor::Line(dc, box.x, box.y, box.x + box.width - 1, box.y);
|
|
AColor::Line(dc, box.x, box.y, box.x, box.y + box.height - 1);
|
|
|
|
#if CHANNEL_ON_IS_DOWN
|
|
AColor::LightMIDIChannel(&dc, chanName);
|
|
#else
|
|
AColor::DarkMIDIChannel(&dc, chanName);
|
|
#endif
|
|
AColor::Line(dc,
|
|
box.x + box.width - 1, box.y,
|
|
box.x + box.width - 1, box.y + box.height - 1);
|
|
AColor::Line(dc,
|
|
box.x, box.y + box.height - 1,
|
|
box.x + box.width - 1, box.y + box.height - 1);
|
|
} else {
|
|
AColor::MIDIChannel(&dc, 0);
|
|
dc.DrawRectangle(box);
|
|
#if CHANNEL_ON_IS_DOWN
|
|
AColor::LightMIDIChannel(&dc, 0);
|
|
#else
|
|
AColor::DarkMIDIChannel(&dc, 0);
|
|
#endif
|
|
AColor::Line(dc, box.x, box.y, box.x + box.width - 1, box.y);
|
|
AColor::Line(dc, box.x, box.y, box.x, box.y + box.height - 1);
|
|
|
|
#if CHANNEL_ON_IS_DOWN
|
|
AColor::DarkMIDIChannel(&dc, 0);
|
|
#else
|
|
AColor::LightMIDIChannel(&dc, 0);
|
|
#endif
|
|
AColor::Line(dc,
|
|
box.x + box.width - 1, box.y,
|
|
box.x + box.width - 1, box.y + box.height - 1);
|
|
AColor::Line(dc,
|
|
box.x, box.y + box.height - 1,
|
|
box.x + box.width - 1, box.y + box.height - 1);
|
|
|
|
}
|
|
|
|
wxString t;
|
|
wxCoord w;
|
|
wxCoord h;
|
|
|
|
t.Printf(wxT("%d"), chanName);
|
|
dc.GetTextExtent(t, &w, &h);
|
|
|
|
dc.DrawText(t, box.x + (box.width - w) / 2, box.y + (box.height - h) / 2);
|
|
}
|
|
}
|
|
AColor::MIDIChannel(&dc, 0); // always return with gray color selected
|
|
return box.GetBottom();
|
|
}
|
|
|
|
bool NoteTrack::LabelClick(wxRect & r, int mx, int my, bool right)
|
|
{
|
|
int wid = 23;
|
|
int ht = 16;
|
|
|
|
if (r.height < ht * 4)
|
|
return false;
|
|
|
|
int x = r.x + (r.width / 2 - wid * 2);
|
|
int y = r.y + 1;
|
|
// after adding Mute and Solo buttons, mapping is broken, so hack in the offset
|
|
y += 12;
|
|
|
|
int col = (mx - x) / wid;
|
|
int row = (my - y) / ht;
|
|
|
|
if (row < 0 || row >= 4 || col < 0 || col >= 4)
|
|
return false;
|
|
|
|
int channel = row * 4 + col;
|
|
|
|
if (right) {
|
|
if (mVisibleChannels == CHANNEL_BIT(channel))
|
|
mVisibleChannels = ALL_CHANNELS;
|
|
else
|
|
mVisibleChannels = CHANNEL_BIT(channel);
|
|
} else
|
|
ToggleVisibleChan(channel);
|
|
|
|
return true;
|
|
}
|
|
|
|
void NoteTrack::SetSequence(std::unique_ptr<Alg_seq> &&seq)
|
|
{
|
|
mSeq = std::move(seq);
|
|
}
|
|
|
|
Alg_seq* NoteTrack::GetSequence()
|
|
{
|
|
return mSeq.get();
|
|
}
|
|
|
|
void NoteTrack::PrintSequence()
|
|
{
|
|
FILE *debugOutput;
|
|
|
|
debugOutput = fopen("debugOutput.txt", "wt");
|
|
fprintf(debugOutput, "Importing MIDI...\n");
|
|
|
|
if (mSeq) {
|
|
int i = 0;
|
|
|
|
while(i < mSeq->length()) {
|
|
fprintf(debugOutput, "--\n");
|
|
fprintf(debugOutput, "type: %c\n",
|
|
((Alg_event_ptr)mSeq->track_list.tracks[i])->get_type());
|
|
fprintf(debugOutput, "time: %f\n",
|
|
((Alg_event_ptr)mSeq->track_list.tracks[i])->time);
|
|
fprintf(debugOutput, "channel: %li\n",
|
|
((Alg_event_ptr)mSeq->track_list.tracks[i])->chan);
|
|
|
|
if(((Alg_event_ptr)mSeq->track_list.tracks[i])->get_type() == wxT('n'))
|
|
{
|
|
fprintf(debugOutput, "pitch: %f\n",
|
|
((Alg_note_ptr)mSeq->track_list.tracks[i])->pitch);
|
|
fprintf(debugOutput, "duration: %f\n",
|
|
((Alg_note_ptr)mSeq->track_list.tracks[i])->dur);
|
|
fprintf(debugOutput, "velocity: %f\n",
|
|
((Alg_note_ptr)mSeq->track_list.tracks[i])->loud);
|
|
}
|
|
else if(((Alg_event_ptr)mSeq->track_list.tracks[i])->get_type() == wxT('n'))
|
|
{
|
|
fprintf(debugOutput, "key: %li\n", ((Alg_update_ptr)mSeq->track_list.tracks[i])->get_identifier());
|
|
fprintf(debugOutput, "attribute type: %c\n", ((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.attr_type());
|
|
fprintf(debugOutput, "attribute: %s\n", ((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.attr_name());
|
|
|
|
if(((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.attr_type() == wxT('r'))
|
|
{
|
|
fprintf(debugOutput, "value: %f\n", ((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.r);
|
|
}
|
|
else if(((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.attr_type() == wxT('i')) {
|
|
fprintf(debugOutput, "value: %li\n", ((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.i);
|
|
}
|
|
else if(((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.attr_type() == wxT('s')) {
|
|
fprintf(debugOutput, "value: %s\n", ((Alg_update_ptr)mSeq->track_list.tracks[i])->parameter.s);
|
|
}
|
|
else {}
|
|
}
|
|
|
|
i++;
|
|
}
|
|
}
|
|
else {
|
|
fprintf(debugOutput, "No sequence defined!\n");
|
|
}
|
|
|
|
fclose(debugOutput);
|
|
}
|
|
|
|
int NoteTrack::GetVisibleChannels()
|
|
{
|
|
return mVisibleChannels;
|
|
}
|
|
|
|
Track::Holder NoteTrack::Cut(double t0, double t1)
|
|
{
|
|
if (t1 <= t0)
|
|
return{};
|
|
double len = t1-t0;
|
|
|
|
auto newTrack = std::make_unique<NoteTrack>(mDirManager);
|
|
|
|
newTrack->Init(*this);
|
|
|
|
mSeq->convert_to_seconds();
|
|
newTrack->mSeq.reset(mSeq->cut(t0 - GetOffset(), len, false));
|
|
newTrack->SetOffset(GetOffset());
|
|
|
|
// What should be done with the rest of newTrack's members?
|
|
//(mBottomNote, mDirManager, mLastMidiPosition,
|
|
// mSerializationBuffer, mSerializationLength, mVisibleChannels)
|
|
|
|
// This std::move is needed to "upcast" the pointer type
|
|
return std::move(newTrack);
|
|
}
|
|
|
|
Track::Holder NoteTrack::Copy(double t0, double t1) const
|
|
{
|
|
if (t1 <= t0)
|
|
return{};
|
|
double len = t1-t0;
|
|
|
|
auto newTrack = std::make_unique<NoteTrack>(mDirManager);
|
|
|
|
newTrack->Init(*this);
|
|
|
|
mSeq->convert_to_seconds();
|
|
newTrack->mSeq.reset(mSeq->copy(t0 - GetOffset(), len, false));
|
|
newTrack->SetOffset(GetOffset());
|
|
|
|
// What should be done with the rest of newTrack's members?
|
|
//(mBottomNote, mDirManager, mLastMidiPosition,
|
|
// mSerializationBuffer, mSerializationLength, mVisibleChannels)
|
|
|
|
// This std::move is needed to "upcast" the pointer type
|
|
return std::move(newTrack);
|
|
}
|
|
|
|
bool NoteTrack::Trim(double t0, double t1)
|
|
{
|
|
if (t1 <= t0)
|
|
return false;
|
|
mSeq->convert_to_seconds();
|
|
// DELETE way beyond duration just in case something is out there:
|
|
mSeq->clear(t1 - GetOffset(), mSeq->get_dur() + 10000.0, false);
|
|
// Now that stuff beyond selection is cleared, clear before selection:
|
|
mSeq->clear(0.0, t0 - GetOffset(), false);
|
|
// want starting time to be t0
|
|
SetOffset(t0);
|
|
return true;
|
|
}
|
|
|
|
bool NoteTrack::Clear(double t0, double t1)
|
|
{
|
|
// If t1 = t0, should Clear return true?
|
|
if (t1 <= t0)
|
|
return false;
|
|
double len = t1-t0;
|
|
|
|
if (mSeq)
|
|
mSeq->clear(t0 - GetOffset(), len, false);
|
|
|
|
return true;
|
|
}
|
|
|
|
bool NoteTrack::Paste(double t, const Track *src)
|
|
{
|
|
// Paste inserts src at time t. If src has a positive offset,
|
|
// the offset is treated as silence which is also inserted. If
|
|
// the offset is negative, the offset is ignored and the ENTIRE
|
|
// src is inserted (otherwise, we would either lose data from
|
|
// src by not inserting things at negative times, or inserting
|
|
// things at negative times could overlap things already in
|
|
// the destination track).
|
|
|
|
//Check that src is a non-NULL NoteTrack
|
|
if (src == NULL || src->GetKind() != Track::Note)
|
|
return false;
|
|
|
|
NoteTrack* other = (NoteTrack*)src;
|
|
if (other->mSeq == NULL)
|
|
return false;
|
|
|
|
if(!mSeq)
|
|
mSeq = std::make_unique<Alg_seq>();
|
|
|
|
if (other->GetOffset() > 0) {
|
|
mSeq->convert_to_seconds();
|
|
mSeq->insert_silence(t - GetOffset(), other->GetOffset());
|
|
t += other->GetOffset();
|
|
}
|
|
mSeq->paste(t - GetOffset(), other->mSeq.get());
|
|
|
|
return true;
|
|
}
|
|
|
|
// Call this function to manipulate the underlying sequence data. This is
|
|
// NOT the function that handles horizontal dragging.
|
|
bool NoteTrack::Shift(double t) // t is always seconds
|
|
{
|
|
if (t > 0) {
|
|
// insert an even number of measures
|
|
mSeq->convert_to_beats();
|
|
// get initial tempo
|
|
double tempo = mSeq->get_tempo(0.0);
|
|
double beats_per_measure = mSeq->get_bar_len(0.0);
|
|
int m = ROUND(t * tempo / beats_per_measure);
|
|
// need at least 1 measure, so if we rounded down to zero, fix it
|
|
if (m == 0) m = 1;
|
|
// compute NEW tempo so that m measures at NEW tempo take t seconds
|
|
tempo = beats_per_measure * m / t; // in beats per second
|
|
mSeq->insert_silence(0.0, beats_per_measure * m);
|
|
mSeq->set_tempo(tempo * 60.0 /* bpm */, 0.0, beats_per_measure * m);
|
|
mSeq->write("afterShift.gro");
|
|
} else if (t < 0) {
|
|
mSeq->convert_to_seconds();
|
|
mSeq->clear(0, t, true);
|
|
} else { // offset is zero, no modifications
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
double NoteTrack::NearestBeatTime(double time, double *beat)
|
|
{
|
|
wxASSERT(mSeq);
|
|
// Alg_seq knows nothing about offset, so remove offset time
|
|
double seq_time = time - GetOffset();
|
|
seq_time = mSeq->nearest_beat_time(seq_time, beat);
|
|
// add the offset back in to get "actual" audacity track time
|
|
return seq_time + GetOffset();
|
|
}
|
|
|
|
bool NoteTrack::StretchRegion(double t0, double t1, double dur)
|
|
{
|
|
wxASSERT(mSeq);
|
|
// Alg_seq::stretch_region uses beats, so we translate time
|
|
// to beats first:
|
|
t0 -= GetOffset();
|
|
t1 -= GetOffset();
|
|
double b0 = mSeq->get_time_map()->time_to_beat(t0);
|
|
double b1 = mSeq->get_time_map()->time_to_beat(t1);
|
|
bool result = mSeq->stretch_region(b0, b1, dur);
|
|
if (result) {
|
|
mSeq->convert_to_seconds();
|
|
mSeq->set_dur(mSeq->get_dur() + dur - (t1 - t0));
|
|
}
|
|
return result;
|
|
}
|
|
|
|
namespace
|
|
{
|
|
void swap(std::unique_ptr<Alg_seq> &a, std::unique_ptr<Alg_seq> &b)
|
|
{
|
|
std::unique_ptr<Alg_seq> tmp = std::move(a);
|
|
a = std::move(b);
|
|
b = std::move(tmp);
|
|
}
|
|
}
|
|
|
|
Alg_seq *NoteTrack::MakeExportableSeq(std::unique_ptr<Alg_seq> &cleanup)
|
|
{
|
|
cleanup.reset();
|
|
double offset = GetOffset();
|
|
if (offset == 0)
|
|
return mSeq.get();
|
|
// make a copy, deleting events that are shifted before time 0
|
|
double start = -offset;
|
|
if (start < 0) start = 0;
|
|
// notes that begin before "start" are not included even if they
|
|
// extend past "start" (because "all" parameter is set to false)
|
|
cleanup.reset( mSeq->copy(start, mSeq->get_dur() - start, false) );
|
|
auto seq = cleanup.get();
|
|
if (offset > 0) {
|
|
// swap cleanup and mSeq so that Shift operates on the NEW copy
|
|
swap(mSeq, cleanup);
|
|
Shift(offset);
|
|
swap(mSeq, cleanup); // undo the swap
|
|
#ifdef OLD_CODE
|
|
// now shift events by offset. This must be done with an integer
|
|
// number of measures, so first, find the beats-per-measure
|
|
double beats_per_measure = 4.0;
|
|
Alg_time_sig_ptr tsp = NULL;
|
|
if (seq->time_sig.length() > 0 && seq->time_sig[0].beat < ALG_EPS) {
|
|
// there is an initial time signature
|
|
tsp = &(seq->time_sig[0]);
|
|
beats_per_measure = (tsp->num * 4) / tsp->den;
|
|
}
|
|
// also need the initial tempo
|
|
double bps = ALG_DEFAULT_BPM / 60;
|
|
Alg_time_map_ptr map = seq->get_time_map();
|
|
Alg_beat_ptr bp = &(map->beats[0]);
|
|
if (bp->time < ALG_EPS) { // tempo change at time 0
|
|
if (map->beats.len > 1) { // compute slope to get tempo
|
|
bps = (map->beats[1].beat - map->beats[0].beat) /
|
|
(map->beats[1].time - map->beats[0].time);
|
|
} else if (seq->get_time_map()->last_tempo_flag) {
|
|
bps = seq->get_time_map()->last_tempo;
|
|
}
|
|
}
|
|
// find closest number of measures to fit in the gap
|
|
// number of measures is offset / measure_time
|
|
double measure_time = beats_per_measure / bps; // seconds per measure
|
|
int n = ROUND(offset / measure_time);
|
|
if (n == 0) n = 1;
|
|
// we will insert n measures. Compute the desired duration of each.
|
|
measure_time = offset / n;
|
|
bps = beats_per_measure / measure_time;
|
|
// insert integer multiple of measures at beginning
|
|
seq->convert_to_beats();
|
|
seq->insert_silence(0, beats_per_measure * n);
|
|
// make sure time signature at 0 is correct
|
|
if (tsp) {
|
|
seq->set_time_sig(0, tsp->num, tsp->den);
|
|
}
|
|
// adjust tempo to match offset
|
|
seq->set_tempo(bps * 60.0, 0, beats_per_measure * n);
|
|
#endif
|
|
} else {
|
|
// if offset is negative, it might not be a multiple of beats, but
|
|
// we want to preserve the relative positions of measures. I.e. we
|
|
// should shift barlines and time signatures as well as notes.
|
|
// Insert a time signature at the first bar-line if necessary.
|
|
|
|
// Translate start from seconds to beats and call it beat:
|
|
double beat = mSeq->get_time_map()->time_to_beat(start);
|
|
// Find the time signature in mSeq in effect at start (beat):
|
|
int i = mSeq->time_sig.find_beat(beat);
|
|
// i is where you would insert a NEW time sig at beat,
|
|
// Case 1: beat coincides with a time sig at i. Time signature
|
|
// at beat means that there is a barline at beat, so when beat
|
|
// is shifted to 0, the relative barline positions are preserved
|
|
if (mSeq->time_sig.length() > 0 &&
|
|
within(beat, mSeq->time_sig[i].beat, ALG_EPS)) {
|
|
// beat coincides with time signature change, so offset must
|
|
// be a multiple of beats
|
|
/* do nothing */ ;
|
|
// Case 2: there is no time signature before beat.
|
|
} else if (i == 0 && (mSeq->time_sig.length() == 0 ||
|
|
mSeq->time_sig[i].beat > beat)) {
|
|
// If beat does not fall on an implied barline, we need to
|
|
// insert a time signature.
|
|
double measures = beat / 4.0;
|
|
double imeasures = ROUND(measures);
|
|
if (!within(measures, imeasures, ALG_EPS)) {
|
|
double bar_offset = (int(measures) + 1) * 4.0 - beat;
|
|
seq->set_time_sig(bar_offset, 4, 4);
|
|
}
|
|
// This case should never be true because if i == 0, either there
|
|
// are no time signatures before beat (Case 2),
|
|
// or there is one time signature at beat (Case 1)
|
|
} else if (i == 0) {
|
|
/* do nothing (might be good to assert(false)) */ ;
|
|
// Case 3: i-1 must be the effective time sig position
|
|
} else {
|
|
i -= 1; // index the time signature in effect at beat
|
|
Alg_time_sig_ptr tsp = &(mSeq->time_sig[i]);
|
|
double beats_per_measure = (tsp->num * 4) / tsp->den;
|
|
double measures = (beat - tsp->beat) / beats_per_measure;
|
|
int imeasures = ROUND(measures);
|
|
if (!within(measures, imeasures, ALG_EPS)) {
|
|
// beat is not on a measure, so we need to insert a time sig
|
|
// to force a bar line at the first measure location after
|
|
// beat
|
|
double bar = tsp->beat + beats_per_measure * (int(measures) + 1);
|
|
double bar_offset = bar - beat;
|
|
// insert NEW time signature at bar_offset in NEW sequence
|
|
// It will have the same time signature, but the position will
|
|
// force a barline to match the barlines in mSeq
|
|
seq->set_time_sig(bar_offset, tsp->num, tsp->den);
|
|
}
|
|
// else beat coincides with a barline, so no need for an extra
|
|
// time signature to force barline alignment
|
|
}
|
|
}
|
|
return seq;
|
|
}
|
|
|
|
|
|
bool NoteTrack::ExportMIDI(const wxString &f)
|
|
{
|
|
std::unique_ptr<Alg_seq> cleanup;
|
|
auto seq = MakeExportableSeq(cleanup);
|
|
bool rslt = seq->smf_write(f.mb_str());
|
|
return rslt;
|
|
}
|
|
|
|
bool NoteTrack::ExportAllegro(const wxString &f)
|
|
{
|
|
double offset = GetOffset();
|
|
bool in_seconds;
|
|
gPrefs->Read(wxT("/FileFormats/AllegroStyle"), &in_seconds, true);
|
|
if (in_seconds) {
|
|
mSeq->convert_to_seconds();
|
|
} else {
|
|
mSeq->convert_to_beats();
|
|
}
|
|
return mSeq->write(f.mb_str(), offset);
|
|
}
|
|
|
|
|
|
bool NoteTrack::HandleXMLTag(const wxChar *tag, const wxChar **attrs)
|
|
{
|
|
if (!wxStrcmp(tag, wxT("notetrack"))) {
|
|
while (*attrs) {
|
|
const wxChar *attr = *attrs++;
|
|
const wxChar *value = *attrs++;
|
|
if (!value)
|
|
break;
|
|
const wxString strValue = value;
|
|
long nValue;
|
|
double dblValue;
|
|
if (!wxStrcmp(attr, wxT("name")) && XMLValueChecker::IsGoodString(strValue))
|
|
mName = strValue;
|
|
else if (!wxStrcmp(attr, wxT("offset")) &&
|
|
XMLValueChecker::IsGoodString(strValue) &&
|
|
Internat::CompatibleToDouble(strValue, &dblValue))
|
|
SetOffset(dblValue);
|
|
else if (!wxStrcmp(attr, wxT("visiblechannels"))) {
|
|
if (!XMLValueChecker::IsGoodInt(strValue) ||
|
|
!strValue.ToLong(&nValue) ||
|
|
!XMLValueChecker::IsValidVisibleChannels(nValue))
|
|
return false;
|
|
mVisibleChannels = nValue;
|
|
}
|
|
else if (!wxStrcmp(attr, wxT("height")) &&
|
|
XMLValueChecker::IsGoodInt(strValue) && strValue.ToLong(&nValue))
|
|
mHeight = nValue;
|
|
else if (!wxStrcmp(attr, wxT("minimized")) &&
|
|
XMLValueChecker::IsGoodInt(strValue) && strValue.ToLong(&nValue))
|
|
mMinimized = (nValue != 0);
|
|
else if (!wxStrcmp(attr, wxT("isSelected")) &&
|
|
XMLValueChecker::IsGoodInt(strValue) && strValue.ToLong(&nValue))
|
|
this->SetSelected(nValue != 0);
|
|
#ifdef EXPERIMENTAL_MIDI_OUT
|
|
else if (!wxStrcmp(attr, wxT("velocity")) &&
|
|
XMLValueChecker::IsGoodString(strValue) &&
|
|
Internat::CompatibleToDouble(strValue, &dblValue))
|
|
mGain = (float) dblValue;
|
|
#endif
|
|
else if (!wxStrcmp(attr, wxT("bottomnote")) &&
|
|
XMLValueChecker::IsGoodInt(strValue) && strValue.ToLong(&nValue))
|
|
SetBottomNote(nValue);
|
|
else if (!wxStrcmp(attr, wxT("data"))) {
|
|
std::string s(strValue.mb_str(wxConvUTF8));
|
|
std::istringstream data(s);
|
|
mSeq = std::make_unique<Alg_seq>(data, false);
|
|
}
|
|
} // while
|
|
return true;
|
|
}
|
|
return false;
|
|
}
|
|
|
|
XMLTagHandler *NoteTrack::HandleXMLChild(const wxChar * WXUNUSED(tag))
|
|
{
|
|
return NULL;
|
|
}
|
|
|
|
void NoteTrack::WriteXML(XMLWriter &xmlFile)
|
|
{
|
|
std::ostringstream data;
|
|
// Normally, Duplicate is called in pairs -- once to put NoteTrack
|
|
// on the Undo stack, and again to move from the Undo stack to an
|
|
// "active" editable state. For efficiency, we do not do a "real"
|
|
// Duplicate followed by serialization into a binary blob. Instead,
|
|
// we combine the Duplicate with serialization or unserialization.
|
|
// Serialization and Unserialization happen on alternate calls to
|
|
// Duplicate and (usually) produce the right results at the right
|
|
// time.
|
|
// It turns out that this optimized Duplicate is a little too
|
|
// clever. There is at least one case where a track can be duplicated
|
|
// and then AutoSave'd. (E.g. do an "Insert Silence" effect on a
|
|
// NoteTrack.) In this case, mSeq will be NULL. To avoid a crash
|
|
// and perform WriteXML, we may need to restore NoteTracks from binary
|
|
// blobs to regular data structures (with an Alg_seq member).
|
|
Track::Holder holder;
|
|
NoteTrack *saveme = this;
|
|
if (!mSeq) { // replace saveme with an (unserialized) duplicate
|
|
holder = Duplicate();
|
|
saveme = static_cast<NoteTrack*>(holder.get());
|
|
assert(saveme->mSeq);
|
|
}
|
|
saveme->mSeq->write(data, true);
|
|
xmlFile.StartTag(wxT("notetrack"));
|
|
xmlFile.WriteAttr(wxT("name"), saveme->mName);
|
|
xmlFile.WriteAttr(wxT("offset"), saveme->GetOffset());
|
|
xmlFile.WriteAttr(wxT("visiblechannels"), saveme->mVisibleChannels);
|
|
xmlFile.WriteAttr(wxT("height"), saveme->GetActualHeight());
|
|
xmlFile.WriteAttr(wxT("minimized"), saveme->GetMinimized());
|
|
xmlFile.WriteAttr(wxT("isSelected"), this->GetSelected());
|
|
|
|
#ifdef EXPERIMENTAL_MIDI_OUT
|
|
xmlFile.WriteAttr(wxT("velocity"), (double) saveme->mGain);
|
|
#endif
|
|
xmlFile.WriteAttr(wxT("bottomnote"), saveme->mBottomNote);
|
|
xmlFile.WriteAttr(wxT("data"), wxString(data.str().c_str(), wxConvUTF8));
|
|
xmlFile.EndTag(wxT("notetrack"));
|
|
}
|
|
|
|
void NoteTrack::StartVScroll()
|
|
{
|
|
mStartBottomNote = mBottomNote;
|
|
}
|
|
|
|
void NoteTrack::VScroll(int start, int end)
|
|
{
|
|
int ph = GetPitchHeight();
|
|
int delta = ((end - start) + ph / 2) / ph;
|
|
SetBottomNote(mStartBottomNote + delta);
|
|
}
|
|
|
|
// Zoom the note track, centering the pitch at centerY,
|
|
// amount is 1 for zoom in, and -1 for zoom out
|
|
void NoteTrack::Zoom(int centerY, int amount)
|
|
{
|
|
// Construct track rectangle to map pitch to screen coordinates
|
|
// Only y and height are needed:
|
|
wxRect trackRect(0, GetY(), 1, GetHeight());
|
|
PrepareIPitchToY(trackRect);
|
|
int centerPitch = YToIPitch(centerY);
|
|
// zoom out by changing the pitch height -- a small integer
|
|
mPitchHeight += amount;
|
|
if (mPitchHeight <= 0) mPitchHeight = 1;
|
|
PrepareIPitchToY(trackRect); // update because mPitchHeight changed
|
|
int newCenterPitch = YToIPitch(GetY() + GetHeight() / 2);
|
|
// center the pitch that the user clicked on
|
|
SetBottomNote(mBottomNote + (centerPitch - newCenterPitch));
|
|
}
|
|
|
|
|
|
void NoteTrack::ZoomTo(int start, int end)
|
|
{
|
|
wxRect trackRect(0, GetY(), 1, GetHeight());
|
|
PrepareIPitchToY(trackRect);
|
|
int topPitch = YToIPitch(start);
|
|
int botPitch = YToIPitch(end);
|
|
if (topPitch < botPitch) { // swap
|
|
int temp = topPitch; topPitch = botPitch; botPitch = temp;
|
|
}
|
|
if (topPitch == botPitch) { // can't divide by zero, do something else
|
|
Zoom(start, 1);
|
|
return;
|
|
}
|
|
int trialPitchHeight = trackRect.height / (topPitch - botPitch);
|
|
if (trialPitchHeight > 25) { // keep mPitchHeight in bounds [1...25]
|
|
trialPitchHeight = 25;
|
|
} else if (trialPitchHeight == 0) {
|
|
trialPitchHeight = 1;
|
|
}
|
|
Zoom((start + end) / 2, trialPitchHeight - mPitchHeight);
|
|
}
|
|
|
|
int NoteTrack::YToIPitch(int y)
|
|
{
|
|
y = mBottom - y; // pixels above pitch 0
|
|
int octave = (y / GetOctaveHeight());
|
|
y -= octave * GetOctaveHeight();
|
|
// result is approximate because C and G are one pixel taller than
|
|
// mPitchHeight.
|
|
return (y / mPitchHeight) + octave * 12;
|
|
}
|
|
|
|
#endif // USE_MIDI
|