//________________________________________________________________________________________ // // FOF synthesis demo // 2025, Stanko Juzbasic // Inspired by IRCAM historical program CHANT, by Jean-Baptiste Barrière, // Xavier Rodet, and Pierre-François Baisnée // synthesizes 24 distinct vowels, oral and nasal // // cc -o fof fof.c -framework AudioToolbox -framework AudioUnit -framework CoreAudio\ // -framework CoreMidi -framework CoreFoundation -lm -lpthread // // // Start the program in demo mode by typing ./fof // or hook a MIDI keyboard and start the program by typing ./fof (without arguments) //________________________________________________________________________________________ #include #include #include #include #include #include #include #include #include #define MAX_HELD_NOTES 16 #ifndef M_PI #define M_PI 3.14159265358979323846 #endif #define TWO_PI 6.28318530717958647692 #define PI_OVER_20 0.15707963267948966192 #define MAX_GRAINS 1024 #define NUM_FORMANTS 6 #define NUM_VOWELS 12 // basic vowels, with nasal versions = 24 #define SMOOTHING_FACTOR 0.08 typedef struct { double phase; double phase_inc; double age; double duration; double attack; double decay_rate; double amplitude; int active; } FofGrain; typedef struct { double freq; double bandwidth; double db_amplitude; } FormantTarget; typedef struct { double tongue_front_back; double tongue_high_low; double mouth_stretch; FormantTarget formants[NUM_FORMANTS]; } VowelVertex; typedef struct { double sample_rate; double inv_sample_rate; double attack_time; double decay_time; volatile double control_tongue_x; volatile double control_tongue_y; volatile double control_stretch_z; volatile double control_f0_hz; double smoothed_x; double smoothed_y; double smoothed_z; double current_f0_inc; double target_phase_incs[NUM_FORMANTS]; double target_decay_rates[NUM_FORMANTS]; double target_gains[NUM_FORMANTS]; volatile int control_gate; // 1 = MIDI key held, 0 = released volatile double control_velocity; // 0.0 to 1.0 double current_amp_env; // Smooth anti-click gate envelope double f0_phase; FofGrain grains[MAX_GRAINS]; //volatile double control_f0_hz; double smoothed_f0_hz; // Current gliding pitch double glide_coeff; // Portamento speed (e.g., 0.0005 to 0.01) volatile double control_mod_wheel; // 0.0 to 1.0 (from MIDI CC 1) double smoothed_mod_wheel; double vibrato_phase; // Vowel coarticulation state int prev_vowel_idx; int target_vowel_idx; double vowel_morph; double morph_progress; //(linear 0.0 → 1.0 timer) volatile double control_nasal; // 0.0 = oral, 1.0 = full nasalization double smoothed_nasal; } FofSynthesizer; typedef struct { uint8_t notes[MAX_HELD_NOTES]; int count; FofSynthesizer *fof; } MidiMonoController; int g_demoMode = 0; int debug = 0; //_______________________________________________________________________________________ FofSynthesizer *g_fofEngine = NULL; static MidiMonoController g_midiCtrl = { .count = 0, .fof = NULL }; //_______________________________________________________________________________________ // Complete 3D Spatial Vowel library mapped according to tongue coordinates and lip configuration const VowelVertex ANATOMICAL_SPACE[NUM_VOWELS] = { // 0. Vowel "A": Back, Low, Stretched Wide {-1.0, -1.0, 1.0, {{600, 60, 0.0}, {1040, 70, -7.0}, {2250, 110, -16.0}, {2450, 120, -22.0}, {3100, 50, -28.0}, {265, 50, -90.0}}}, // 1. Vowel "E": Front, Mid, Stretched Wide { 0.5, 0.0, 1.0, {{400, 40, 0.0}, {1600, 80, -12.0}, {2200, 100, -16.0}, {2400, 120, -24.0}, {3120, 50, -30.0}, {265, 50, -90.0}}}, // 2. Vowel "I": Front, High, Stretched Wide { 1.0, 1.0, 1.0, {{250, 60, 0.0}, {1750, 90, -12.0}, {2600, 100, -18.0}, {3050, 120, -30.0}, {3150, 50, -35.0}, {265, 50, -90.0}}}, // 3. Vowel "O": Back, Mid, Rounded/Compressed {-0.5, 0.0, -1.0, {{450, 40, 0.0}, {800, 80, -11.0}, {2300, 100, -21.0}, {2675, 120, -20.0}, {3100, 50, -30.0}, {265, 50, -90.0}}}, // 4. Vowel "U": Back, High, Rounded/Compressed {-1.0, 1.0, -1.0, {{325, 50, 0.0}, {700, 60, -16.0}, {2100, 110, -28.0}, {2400, 120, -32.0}, {3100, 50, -40.0}, {265, 50, -90.0}}}, // 5. Vowel "Ö": Front, Mid (Same Tongue space as E), but Rounded/Compressed Lips (-1.0) { 0.5, 0.0, -1.0, {{370, 40, 0.0}, {1500, 85, -10.0}, {2150, 100, -15.0}, {2350, 120, -22.0}, {3100, 50, -30.0}, {265, 50, -90.0}}}, // 6. Vowel "Ü": Front, High (Same Tongue space as I), but Rounded/Compressed Lips (-1.0) { 1.0, 1.0, -1.0, {{230, 50, 0.0}, {1650, 90, -11.0}, {2500, 100, -16.0}, {2950, 120, -28.0}, {3100, 50, -32.0}, {265, 50, -90.0}}}, // 7. Dark Closed-Lip Hum (Back, Tongue Low, Lips Tightly Closed) {-1.0, -1.0, -1.0, {{220, 40, 0.0}, {520, 50, -18.0}, {1800, 90, -36.0}, {2200, 110, -45.0}, {2750, 50, -54.0}, {265, 50, -90.0}}}, // 8. Neutral "a" ("C" / ə): Center, Mid, Neutral Lips { 0.0, 0.0, 0.0, {{500, 60, 0.0}, {1500, 80, -10.0}, {2500, 100, -18.0}, {3300, 120, -24.0}, {3750, 60, -32.0}, {265, 50, -90.0}}}, // 9. Open Front "Æ" ("X"): Front, Low, Stretched Wide { 0.8, -1.0, 1.0, {{750, 75, 0.0}, {1700, 90, -8.0}, {2450, 110, -15.0}, {3200, 120, -22.0}, {3600, 60, -30.0}, {265, 50, -90.0}}}, // 10. Nasal "M" (/m/): Bilabial Closure (Lips Sealed, Oral Anti-Resonance ~1000 Hz) {-0.8, -0.5, -1.0, {{260, 45, 0.0}, {1000, 200, -30.0}, {2250, 65, -20.0}, {2900, 120, -34.0}, {3400, 80, -44.0}, {265, 50, -90.0}}}, // 11. Nasal "N" (/n/): Alveolar Closure (Tongue Tip Up, Oral Anti-Resonance ~1550 Hz) { 0.6, 0.8, -0.5, {{280, 45, 0.0}, {1550, 220, -28.0}, {2450, 70, -18.0}, {3100, 120, -32.0}, {3600, 80, -42.0}, {265, 50, -90.0}}} }; //_______________________________________________________________________________________ void fof_set_pitch(FofSynthesizer* fof, double f0_hz) { // if (f0_hz < 16.0) f0_hz = 16.0; if (f0_hz > 1500.0) f0_hz = 1500.0; fof->control_f0_hz = f0_hz; } //_______________________________________________________________________________________ void fof_set_portamento(FofSynthesizer* fof, double time_ms) { if (time_ms <= 0.1) { fof->glide_coeff = 1.0; // Instant snap } else { // Convert time in ms to a sample-rate-independent 1-pole coefficient fof->glide_coeff = 1.0 - exp(-1.0 / (time_ms * 0.001 * fof->sample_rate)); } } //_______________________________________________________________________________________ static inline double midi_note_to_hz(uint8_t note) { return 440.0 * pow(2.0, ((double)note - 69.0) / 12.0); } //_______________________________________________________________________________________ void midi_handle_note_off(MidiMonoController *ctrl, uint8_t note) { int found = -1; for (int i = 0; i < ctrl->count; i++) { if (ctrl->notes[i] == note) { found = i; break; } } if (found == -1) return; // Remove note from stack for (int i = found; i < ctrl->count - 1; i++) { ctrl->notes[i] = ctrl->notes[i + 1]; } ctrl->count--; if (ctrl->count > 0) { // Last-note priority: revert pitch to previous held note uint8_t prev_note = ctrl->notes[ctrl->count - 1]; fof_set_pitch(ctrl->fof, midi_note_to_hz(prev_note)); } else { // No keys held: close gate ctrl->fof->control_gate = 0; } } //_______________________________________________________________________________________ void midi_handle_note_on(MidiMonoController *ctrl, uint8_t note, uint8_t velocity) { // check against "hacked" note-off: if (velocity == 0) { midi_handle_note_off(ctrl, note); return; } //// bool is_legato = (ctrl->count > 0); ctrl->notes[ctrl->count++] = note; double new_hz = midi_note_to_hz(note); fof_set_pitch(ctrl->fof, new_hz); if (!is_legato) { // First note from silence: snap pitch immediately so it doesn't slide from 0 Hz ctrl->fof->smoothed_f0_hz = new_hz; ctrl->fof->control_velocity = (double)velocity / 127.0; ctrl->fof->control_gate = 1; } //// // Remove duplicate if already present for (int i = 0; i < ctrl->count; i++) { if (ctrl->notes[i] == note) { for (int j = i; j < ctrl->count - 1; j++) { ctrl->notes[j] = ctrl->notes[j + 1]; } ctrl->count--; break; } } // Push new note onto top of stack if (ctrl->count >= MAX_HELD_NOTES) { for (int i = 0; i < MAX_HELD_NOTES - 1; i++) { ctrl->notes[i] = ctrl->notes[i + 1]; } ctrl->count = MAX_HELD_NOTES - 1; } ctrl->notes[ctrl->count++] = note; // Update FOF synth pitch, velocity, and open gate fof_set_pitch(ctrl->fof, midi_note_to_hz(note)); ctrl->fof->control_velocity = (double)velocity / 127.0; ctrl->fof->control_gate = 1; } //_______________________________________________________________________________________ void midi_read_callback1(const MIDIPacketList *pktlist, void *readProcRefCon, void *srcConnRefCon) { MidiMonoController *ctrl = (MidiMonoController *)readProcRefCon; if (!ctrl || !ctrl->fof) return; const MIDIPacket *packet = pktlist->packet; for (UInt32 i = 0; i < pktlist->numPackets; ++i) { UInt16 j = 0; uint8_t running_status = 0; while (j < packet->length) { uint8_t byte = packet->data[j]; if (byte >= 0x80) { // Real-time system messages (0xF8–0xFF) do NOT cancel Running Status per MIDI spec if (byte >= 0xF8) { j++; continue; } // Channel voice message (0x80–0xEF): update running_status and advance to data byte running_status = byte; j++; } else if (running_status == 0) { // Orphaned data byte with no prior status: skip safely j++; continue; } uint8_t msg_type = running_status & 0xF0; if (msg_type == 0x90 && j + 1 < packet->length) { // Note On (also handles vel == 0) uint8_t note = packet->data[j] & 0x7F; uint8_t vel = packet->data[j + 1] & 0x7F; midi_handle_note_on(ctrl, note, vel); j += 2; } else if (msg_type == 0x80 && j + 1 < packet->length) { // Real Note Off uint8_t note = packet->data[j] & 0x7F; midi_handle_note_off(ctrl, note); j += 2; } else if (msg_type == 0xB0 && j + 1 < packet->length) { // Control Change uint8_t cc_num = packet->data[j] & 0x7F; if (cc_num == 123) { // All Notes Off ctrl->count = 0; ctrl->fof->control_gate = 0; } j += 2; } else if ((msg_type == 0xC0 || msg_type == 0xD0) && j < packet->length) { // Program Change / Channel Aftertouch are 1-byte data messages j += 1; } else if (msg_type == 0xE0 && j + 1 < packet->length) { // Pitch Bend (2-byte data message) j += 2; } else { // Unknown or incomplete SysEx/System Common message: reset running_status and advance running_status = 0; j++; } } packet = MIDIPacketNext(packet); } } //_______________________________________________________________________________________ void midi_read_callback(const MIDIPacketList *pktlist, void *readProcRefCon, void *srcConnRefCon) { MidiMonoController *ctrl = (MidiMonoController *)readProcRefCon; if (!ctrl || !ctrl->fof) return; const MIDIPacket *packet = pktlist->packet; for (UInt32 i = 0; i < pktlist->numPackets; ++i) { UInt16 j = 0; uint8_t running_status = 0; while (j < packet->length) { uint8_t byte = packet->data[j]; if (byte >= 0x80) { // System Real-Time messages (0xF8–0xFF) do not affect Running Status if (byte >= 0xF8) { j++; continue; } // System Common / SysEx messages (0xF0–0xF7) cancel Running Status if (byte >= 0xF0) { running_status = 0; j++; continue; } // Channel Voice status byte (0x80–0xEF) running_status = byte; j++; } else if (running_status == 0) { // Orphaned data byte with no valid status: skip safely j++; continue; } uint8_t msg_type = running_status & 0xF0; if (msg_type == 0x90 && j + 1 < packet->length) { // Note On (velocity == 0 is handled inside midi_handle_note_on as Note Off) uint8_t note = packet->data[j] & 0x7F; uint8_t vel = packet->data[j + 1] & 0x7F; midi_handle_note_on(ctrl, note, vel); j += 2; } else if (msg_type == 0x80 && j + 1 < packet->length) { // Explicit Note Off uint8_t note = packet->data[j] & 0x7F; midi_handle_note_off(ctrl, note); j += 2; } else if (msg_type == 0xB0 && j + 1 < packet->length) { // Control Change (CC) uint8_t cc_num = packet->data[j] & 0x7F; uint8_t cc_val = packet->data[j + 1] & 0x7F; if (cc_num == 1) { // Mod Wheel -> Vibrato Depth (0.0 to 1.0) ctrl->fof->control_mod_wheel = (double)cc_val / 127.0; } else if (cc_num == 120 || cc_num == 123) { // All Sound Off (120) / All Notes Off (123) ctrl->count = 0; ctrl->fof->control_gate = 0; } j += 2; } else if ((msg_type == 0xC0 || msg_type == 0xD0) && j < packet->length) { // 1-byte data messages: Program Change (0xC0) / Channel Pressure (0xD0) j += 1; } else if (msg_type == 0xE0 && j + 1 < packet->length) { // 2-byte data message: Pitch Bend (0xE0) j += 2; } else { // Malformed or truncated packet: reset status and advance running_status = 0; j++; } } packet = MIDIPacketNext(packet); } } //_______________________________________________________________________________________ void midi_notify_callback(const MIDINotification *message, void *refCon) { if (message->messageID == kMIDIMsgObjectAdded) { const MIDIObjectAddRemoveNotification *addMsg = (const MIDIObjectAddRemoveNotification *)message; if (addMsg->childType == kMIDIObjectType_Source) { MIDIPortRef inPort = *(MIDIPortRef *)refCon; MIDIPortConnectSource(inPort, (MIDIEndpointRef)addMsg->child, NULL); } } } //_______________________________________________________________________________________ void fof_init_midi(FofSynthesizer *fof) { g_midiCtrl.fof = fof; g_midiCtrl.count = 0; static MIDIClientRef client; static MIDIPortRef inPort; MIDIClientCreate(CFSTR("FOFSynthMIDI"), midi_notify_callback, &inPort, &client); MIDIInputPortCreate(client, CFSTR("FOFSynthInput"), midi_read_callback, &g_midiCtrl, &inPort); ItemCount sources = MIDIGetNumberOfSources(); printf("Found %lu MIDI sources\n", sources); for (ItemCount i = 0; i < sources; ++i) { MIDIPortConnectSource(inPort, MIDIGetSource(i), NULL); } } //_______________________________________________________________________________________ void fof_set_nasal(FofSynthesizer* fof, double nasal_amount) { if (nasal_amount < 0.0) nasal_amount = 0.0; if (nasal_amount > 1.0) nasal_amount = 1.0; fof->control_nasal = nasal_amount; } //_______________________________________________________________________________________ void fof_set_vocal_tract_3d(FofSynthesizer* fof, double front_back, double high_low, double mouth_stretch) { if (front_back < -1.0) front_back = -1.0; if (front_back > 1.0) front_back = 1.0; if (high_low < -1.0) high_low = -1.0; if (high_low > 1.0) high_low = 1.0; if (mouth_stretch < -1.0) mouth_stretch = -1.0; if (mouth_stretch > 1.0) mouth_stretch = 1.0; fof->control_tongue_x = front_back; fof->control_tongue_y = high_low; fof->control_stretch_z = mouth_stretch; // Find the vowel vertex matching these 3D coordinates int best_v = 0; double min_d2 = 1e9; for (int v = 0; v < NUM_VOWELS; v++) { double dx = front_back - ANATOMICAL_SPACE[v].tongue_front_back; double dy = high_low - ANATOMICAL_SPACE[v].tongue_high_low; double dz = mouth_stretch - ANATOMICAL_SPACE[v].mouth_stretch; double d2 = (dx * dx) + (dy * dy) + (dz * dz); if (d2 < min_d2) { min_d2 = d2; best_v = v; } } // Trigger natural transition to the new vowel if (best_v != fof->target_vowel_idx) { if (fof->vowel_morph > 0.5) { fof->prev_vowel_idx = fof->target_vowel_idx; } fof->target_vowel_idx = best_v; fof->vowel_morph = 0.0; } } //_______________________________________________________________________________________ void fof_set_vocal_tract_3d0(FofSynthesizer* fof, double front_back, double high_low, double mouth_stretch) { if (front_back < -1.0) front_back = -1.0; if (front_back > 1.0) front_back = 1.0; if (high_low < -1.0) high_low = -1.0; if (high_low > 1.0) high_low = 1.0; if (mouth_stretch < -1.0) mouth_stretch = -1.0; if (mouth_stretch > 1.0) mouth_stretch = 1.0; fof->control_tongue_x = front_back; fof->control_tongue_y = high_low; fof->control_stretch_z = mouth_stretch; } //_______________________________________________________________________________________ static inline void fof_audio_thread_sync_controls(FofSynthesizer* fof) { if (fof->smoothed_f0_hz <= 0.0) { fof->smoothed_f0_hz = fof->control_f0_hz; } fof->smoothed_f0_hz += 0.05 * (fof->control_f0_hz - fof->smoothed_f0_hz); fof->smoothed_mod_wheel += 0.08 * (fof->control_mod_wheel - fof->smoothed_mod_wheel); fof->vibrato_phase += (5.2 * 512.0) * fof->inv_sample_rate; if (fof->vibrato_phase >= 1.0) { fof->vibrato_phase -= 1.0; } double vibrato_semitones = sin(fof->vibrato_phase * TWO_PI) * (fof->smoothed_mod_wheel * 0.6); double modulated_f0_hz = fof->smoothed_f0_hz * pow(2.0, vibrato_semitones / 12.0); fof->current_f0_inc = modulated_f0_hz * fof->inv_sample_rate; fof->vowel_morph += 0.12; if (fof->vowel_morph > 1.0) fof->vowel_morph = 1.0; fof->smoothed_nasal += 0.12 * (fof->control_nasal - fof->smoothed_nasal); double nas = fof->smoothed_nasal; double t_raw = fof->vowel_morph; int vA = fof->prev_vowel_idx; int vB = fof->target_vowel_idx; double mid_motion = sin(t_raw * M_PI); double bw_damping_mult = 1.0 + (0.14 * mid_motion); double phase_coef = TWO_PI * fof->inv_sample_rate; for (int f = 0; f < NUM_FORMANTS; f++) { double stagger = (f == 0) ? 1.15 : (f == 1 ? 1.0 : 0.92); double t_staggered = t_raw * stagger; if (t_staggered > 1.0) t_staggered = 1.0; double t_curve = 0.5 * (1.0 - cos(t_staggered * M_PI)); double inv_t = 1.0 - t_curve; double fA = ANATOMICAL_SPACE[vA].formants[f].freq; double fB = ANATOMICAL_SPACE[vB].formants[f].freq; double interp_freq = fA * pow(fB / fA, t_curve); double interp_bw = ((ANATOMICAL_SPACE[vA].formants[f].bandwidth * inv_t) + (ANATOMICAL_SPACE[vB].formants[f].bandwidth * t_curve)) * bw_damping_mult; double interp_db = (ANATOMICAL_SPACE[vA].formants[f].db_amplitude * inv_t) + (ANATOMICAL_SPACE[vB].formants[f].db_amplitude * t_curve); /* if (nas > 0.001) { if (f == 0) { interp_freq = interp_freq * pow(270.0 / interp_freq, nas * 0.65); interp_bw += 75.0 * nas; } else if (f == 1) { interp_bw += 65.0 * nas; interp_db -= 7.0 * nas; } else if (f == 2) { interp_freq = interp_freq * pow(2350.0 / interp_freq, nas * 0.35); interp_bw *= (1.0 - 0.25 * nas); interp_db += 3.5 * nas; } else { interp_db -= 4.0 * nas; } } */ if (nas > 0.001) { if (f == 0) { interp_bw += 35.0 * nas; interp_db -= 3.5 * nas; } else if (f == 1) { interp_bw += 85.0 * nas; interp_db -= 9.5 * nas; } else if (f == 2) { interp_bw *= (1.0 - 0.30 * nas); interp_db += 4.5 * nas; } else if (f == 5) { interp_db = -90.0; } } fof->target_phase_incs[f] = interp_freq * phase_coef; fof->target_decay_rates[f] = M_PI * interp_bw; fof->target_gains[f] = exp(interp_db * PI_OVER_20) * 0.05; } } //_______________________________________________________________________________________ static inline void fof_spawn_formant_grains_rt(FofSynthesizer* fof) { for (int f = 0; f < NUM_FORMANTS; f++) { for (int i = 0; i < MAX_GRAINS; i++) { if (!fof->grains[i].active) { FofGrain* g = &fof->grains[i]; g->phase = 0.0; g->phase_inc = fof->target_phase_incs[f]; g->age = 0.0; g->attack = fof->attack_time; g->duration = fof->attack_time + fof->decay_time; g->amplitude = fof->target_gains[f]; g->decay_rate = fof->target_decay_rates[f]; g->active = 1; fof->glide_coeff = 0.0008; // Default legato portamento speed break; } } } } //_______________________________________________________________________________________ OSStatus FofAudioUnitRenderCallback(void *inRefCon, AudioUnitRenderActionFlags *ioActionFlags, const AudioTimeStamp *inTimeStamp, UInt32 inBusNumber, UInt32 inNumberFrames, AudioBufferList *ioData) { FofSynthesizer *fof = (FofSynthesizer *)inRefCon; if (!fof) return noErr; const int left_channel_idx = 0; const int right_channel_idx = 1; float *leftChannel = (float *)ioData->mBuffers[left_channel_idx].mData; float *rightChannel = (ioData->mNumberBuffers > 1) ? (float *)ioData->mBuffers[right_channel_idx].mData : NULL; double dt = fof->inv_sample_rate; fof_audio_thread_sync_controls(fof); for (UInt32 frame = 0; frame < inNumberFrames; frame++) { double sample_out = 0.0; /// // Smooth gate envelope (prevents clicks on key press/release) double target_env = fof->control_gate ? fof->control_velocity : 0.0; double env_rate = fof->control_gate ? 0.005 : 0.0005; // Fast attack, smooth release fof->current_amp_env += env_rate * (target_env - fof->current_amp_env); fof->f0_phase += fof->current_f0_inc; if (fof->f0_phase >= 1.0) { fof->f0_phase -= 1.0; // Only trigger new vocal glottal pulses while a MIDI key is held if (fof->control_gate) { fof_spawn_formant_grains_rt(fof); } } /// /* fof->f0_phase += fof->current_f0_inc; if (fof->f0_phase >= 1.0) { fof->f0_phase -= 1.0; fof_spawn_formant_grains_rt(fof); } */ for (int i = 0; i < MAX_GRAINS; i++) { if (!fof->grains[i].active) continue; FofGrain* g = &fof->grains[i]; double carrier = sin(g->phase); double env = (g->age < g->attack) ? (g->age / g->attack) : exp(-g->decay_rate * (g->age - g->attack)); sample_out += g->amplitude * env * carrier; g->phase += g->phase_inc; if (g->phase >= TWO_PI) g->phase -= TWO_PI; g->age += dt; if (g->age >= g->duration) g->active = 0; } sample_out *= fof->current_amp_env; leftChannel[frame] = (float)sample_out; if (rightChannel) rightChannel[frame] = (float)sample_out; } return noErr; } //_______________________________________________________________________________________ void* keyboard_input_thread(void* arg) { struct termios orig_termios; struct termios raw_termios; tcgetattr(STDIN_FILENO, &orig_termios); raw_termios = orig_termios; raw_termios.c_lflag &= ~(ICANON | ECHO); tcsetattr(STDIN_FILENO, TCSANOW, &raw_termios); printf("\n--- Extended 3D Anatomical Vocal Space Interface ---\n"); printf("[a] = A (Back, Low, Stretched) | [e] = E (Front, Mid, Stretched)\n"); printf("[i] = I (Front, High, Stretched) | [o] = O (Back, Mid, Rounded)\n"); printf("[u] = U (Back, High, Rounded) | [w] = Ö (Front, Mid, Rounded)\n"); printf("[x] = Æ (Front, Low, Stretched) | [c] = å (Mid, Mid, Mid)\n"); printf("[y] = Ü (Front, High, Rounded) | [m] = M \n"); printf("[n] = N | [v] = Dark, closed Uum \n"); printf("Same keys with play nasal versions of phonemes\n"); printf("MIDI keyboard -> pitch; mod_wheel -> vibrato\n"); printf("In Demo Mode + and - keys step the pitch half-step up/down\n"); printf("No MIDI in Demo Mode \n"); printf("[q] quits the program\n\n"); while (1) { int ch = getchar(); if (debug) { printf("KEY: %d\n", ch); printf("%d\n", g_demoMode); } if (ch == EOF || ch == 'q') break; switch (ch) { // Arguments: (Tongue Front/Back, Tongue High/Low, Mouth Round/Stretch) case 'a': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, -1.0, -1.0, 1.0); printf("Vowel: A (Oral)\n"); break; case 'e': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, 0.5, 0.0, 1.0); printf("Vowel: E (Oral)\n"); break; case 'i': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, 1.0, 1.0, 1.0); printf("Vowel: I (Oral)\n"); break; case 'o': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, -0.5, 0.0, -1.0); printf("Vowel: O (Oral)\n"); break; case 'u': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, -1.0, 1.0, -1.0); printf("Vowel: U (Oral)\n"); break; case 'w': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, 0.5, 0.0, -1.0); printf("Vowel: Ö (Oral)\n"); break; case 'y': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, 1.0, 1.0, -1.0); printf("Vowel: Ü (Oral)\n"); break; case 'v': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, -1.0, -1.0, -1.0); printf("Vowel: Dark Closed Hum (Oral)\n"); break; case 'c': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, 0.0, 0.0, 0.0); printf("Vowel: Schwa (Oral)\n"); break; case 'x': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, 0.8, -1.0, 1.0); printf("Vowel: Æ (Oral)\n"); break; case 'm': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, -0.8, -0.5, -1.0); printf("Consonant: M\n"); break; case 'n': fof_set_nasal(g_fofEngine, 0.0); fof_set_vocal_tract_3d(g_fofEngine, 0.6, 0.8, -0.5); printf("Consonant: N\n"); break; case 'A': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, -1.0, -1.0, 1.0); printf("Vowel: Ã (Nasal)\n"); break; case 'E': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, 0.5, 0.0, 1.0); printf("Vowel: Ẽ (Nasal)\n"); break; case 'I': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, 1.0, 1.0, 1.0); printf("Vowel: Ĩ (Nasal)\n"); break; case 'O': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, -0.5, 0.0, -1.0); printf("Vowel: Õ (Nasal)\n"); break; case 'U': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, -1.0, 1.0, -1.0); printf("Vowel: Ũ (Nasal)\n"); break; case 'W': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, 0.5, 0.0, -1.0); printf("Vowel: Ö (Nasal)\n"); break; case 'Y': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, 1.0, 1.0, -1.0); printf("Vowel: Ü (Nasal)\n"); break; case 'V': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, -1.0, -1.0, -1.0); printf("Vowel: Dark Closed Hum (Nasal)\n"); break; case 'C': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, 0.0, 0.0, 0.0); printf("Vowel: Schwa (Nasal)\n"); break; case 'X': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, 0.8, -1.0, 1.0); printf("Vowel: Æ (Nasal)\n"); break; case 'M': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, -0.8, -0.5, -1.0); printf("Consonant: M (Deep Nasal)\n"); break; case 'N': fof_set_nasal(g_fofEngine, 1.0); fof_set_vocal_tract_3d(g_fofEngine, 0.6, 0.8, -0.5); printf("Consonant: N (Deep Nasal)\n"); break; case 61: case 43: if (g_demoMode) { fof_set_pitch(g_fofEngine, g_fofEngine->control_f0_hz * 1.0594630943592953); printf("Pitch (+1 semitone): %.2f Hz\n", g_fofEngine->control_f0_hz); } break; case 45: case 95: if (g_demoMode) { fof_set_pitch(g_fofEngine, g_fofEngine->control_f0_hz / 1.0594630943592953); printf("Pitch (-1 semitone): %.2f Hz\n", g_fofEngine->control_f0_hz); } break; default: break; } } tcsetattr(STDIN_FILENO, TCSANOW, &orig_termios); return NULL; } //_______________________________________________________________________________________ int main1(int argc, const char *argv[]) { double sample_rate = 44100.0; double initial_pitch = 130.65; if (argc > 1) { char *endptr; double parsed_pitch = strtod(argv[1], &endptr); if (endptr != argv[1]) { initial_pitch = parsed_pitch; } else { fprintf(stderr, "Invalid pitch value provided. Using default %.2f Hz.\n", initial_pitch); } } g_fofEngine = calloc(1, sizeof(FofSynthesizer)); if (!g_fofEngine) return 1; g_fofEngine->sample_rate = sample_rate; g_fofEngine->inv_sample_rate = 1.0 / sample_rate; g_fofEngine->attack_time = 0.002; g_fofEngine->decay_time = 0.040; g_fofEngine->smoothed_x = -1.0; g_fofEngine->smoothed_y = -1.0; g_fofEngine->smoothed_z = 1.0; g_fofEngine->vowel_morph = 1.0; fof_set_pitch(g_fofEngine, initial_pitch); g_fofEngine->smoothed_f0_hz = g_fofEngine->control_f0_hz; g_fofEngine->current_f0_inc = g_fofEngine->control_f0_hz * g_fofEngine->inv_sample_rate; fof_set_vocal_tract_3d(g_fofEngine, -1.0, -1.0, 1.0); if (argc > 1) { g_fofEngine->control_velocity = 0.8; g_fofEngine->control_gate = 1; printf("Playing at %.2f Hz\n", g_fofEngine->control_f0_hz); } else { g_fofEngine->control_gate = 0; printf("Waiting for MIDI Note-On...\n"); } fof_init_midi(g_fofEngine); printf("FOF Synth listening for MIDI. Press Ctrl+C to quit.\n"); AudioComponentDescription desc = { .componentType = kAudioUnitType_Output, .componentSubType = kAudioUnitSubType_DefaultOutput, .componentManufacturer = kAudioUnitManufacturer_Apple }; AudioComponent comp = AudioComponentFindNext(NULL, &desc); AudioUnit outputAudioUnit; AudioComponentInstanceNew(comp, &outputAudioUnit); AudioStreamBasicDescription streamFormat = { .mSampleRate = sample_rate, .mFormatID = kAudioFormatLinearPCM, .mFormatFlags = kAudioFormatFlagIsFloat | kAudioFormatFlagIsNonInterleaved, .mBytesPerPacket = sizeof(float), .mFramesPerPacket = 1, .mBytesPerFrame = sizeof(float), .mChannelsPerFrame = 2, .mBitsPerChannel = 32 }; AudioUnitSetProperty(outputAudioUnit, kAudioUnitProperty_StreamFormat, kAudioUnitScope_Input, 0, &streamFormat, sizeof(streamFormat)); AURenderCallbackStruct callbackStruct = { .inputProc = FofAudioUnitRenderCallback, .inputProcRefCon = g_fofEngine }; AudioUnitSetProperty(outputAudioUnit, kAudioUnitProperty_SetRenderCallback, k AudioUnitScope_Input, 0, &callbackStruct, sizeof(callbackStruct)); AudioUnitInitialize(outputAudioUnit); AudioOutputUnitStart(outputAudioUnit); pthread_t input_thread_id; pthread_create(&input_thread_id, NULL, keyboard_input_thread, NULL); pthread_join(input_thread_id, NULL); AudioOutputUnitStop(outputAudioUnit); AudioUnitUninitialize(outputAudioUnit); AudioComponentInstanceDispose(outputAudioUnit); free(g_fofEngine); return 0; } //_______________________________________________________________________________________ int main(int argc, const char *argv[]) { double sample_rate = 44100.0; double initial_pitch = 130.65; int has_valid_pitch_arg = 0; double parsed_pitch; char *endptr; // 0. Parse arguments if (argc > 1) { parsed_pitch = strtod(argv[1], &endptr); if (endptr != argv[1]) { initial_pitch = parsed_pitch; has_valid_pitch_arg = 1; } else { fprintf(stderr, "Invalid pitch value provided.\n"); } } // 1. Setup CoreAudio g_fofEngine = calloc(1, sizeof(FofSynthesizer)); if (!g_fofEngine) return 1; g_fofEngine->sample_rate = sample_rate; g_fofEngine->inv_sample_rate = 1.0 / sample_rate; g_fofEngine->attack_time = 0.002; g_fofEngine->decay_time = 0.040; g_fofEngine->smoothed_x = -1.0; g_fofEngine->smoothed_y = -1.0; g_fofEngine->smoothed_z = 1.0; g_fofEngine->vowel_morph = 1.0; fof_set_pitch(g_fofEngine, initial_pitch); g_fofEngine->smoothed_f0_hz = g_fofEngine->control_f0_hz; g_fofEngine->current_f0_inc = g_fofEngine->control_f0_hz * g_fofEngine->inv_sample_rate; fof_set_vocal_tract_3d(g_fofEngine, -1.0, -1.0, 1.0); fof_init_midi(g_fofEngine); if (has_valid_pitch_arg) { g_demoMode = 1; g_fofEngine->control_velocity = 0.8; g_fofEngine->control_gate = 1; printf("Demo mode: Playing at %.2f Hz (Press 'q' to quit)\n", g_fofEngine->control_f0_hz); } else { g_demoMode = 0; g_fofEngine->control_gate = 0; printf("FOF Synth listening for MIDI. Press Ctrl+C to quit.\n"); } // 2. Initialize CoreMIDI and connect to MIDI sources fof_init_midi (g_fofEngine); for (int i = 0; i < MAX_GRAINS; i++) g_fofEngine->grains[i].active = 0; fof_set_pitch(g_fofEngine, initial_pitch); fof_set_vocal_tract_3d(g_fofEngine, -1.0, -1.0, 1.0); AudioComponentDescription desc = { .componentType = kAudioUnitType_Output, .componentSubType = kAudioUnitSubType_DefaultOutput, .componentManufacturer = kAudioUnitManufacturer_Apple }; AudioComponent comp = AudioComponentFindNext(NULL, &desc); AudioUnit outputAudioUnit; AudioComponentInstanceNew(comp, &outputAudioUnit); AudioStreamBasicDescription streamFormat = { .mSampleRate = sample_rate, .mFormatID = kAudioFormatLinearPCM, .mFormatFlags = kAudioFormatFlagIsFloat | kAudioFormatFlagIsNonInterleaved, .mBytesPerPacket = sizeof(float), .mFramesPerPacket = 1, .mBytesPerFrame = sizeof(float), .mChannelsPerFrame = 2, .mBitsPerChannel = 32 }; AudioUnitSetProperty(outputAudioUnit, kAudioUnitProperty_StreamFormat, kAudioUnitScope_Input, 0, &streamFormat, sizeof(streamFormat)); AURenderCallbackStruct callbackStruct = { .inputProc = FofAudioUnitRenderCallback, .inputProcRefCon = g_fofEngine }; AudioUnitSetProperty(outputAudioUnit, kAudioUnitProperty_SetRenderCallback, kAudioUnitScope_Input, 0, &callbackStruct, sizeof(callbackStruct)); AudioUnitInitialize(outputAudioUnit); AudioOutputUnitStart(outputAudioUnit); pthread_t input_thread_id; pthread_create(&input_thread_id, NULL, keyboard_input_thread, NULL); pthread_join(input_thread_id, NULL); AudioOutputUnitStop(outputAudioUnit); AudioComponentInstanceDispose(outputAudioUnit); free(g_fofEngine); return 0; }