optimize and fix some stuff

README.md

@@ -5,8 +5,7 @@ FM95 is a audio processor for FM, it does:
 - Pre-Emphasis
 - Low Pass Filtering
 - AGC
-- Stereo (Polar too)
+- Stereo
-- SCA
 - BS412 (mpx power limiter, simplest implementation ever)
 
 Supports these inputs:
@@ -36,11 +35,11 @@ Done!
 
 ## CPU Usage?
 
-Should run completly fine on a pi 5, fine on a pi 3b (30% cpu, 45% with lpf)
+Should run completly fine on a pi 5, fine on a pi 3b (~30% cpu)
 
 ## Other Apps
 
-FM95 also includes some other apps, such as chimer95 which generates GTS tones each half hour, and dcf95 which creates a DCF77 compatible signal, and vban95 now which is a buffered VBAN receiver. And now also SCA generation was moved to sca95 from fm95!
+FM95 also includes some other apps, such as chimer95 which generates GTS tones each half hour, and vban95 now which is a buffered VBAN receiver. And now also SCA generation was moved to sca95 from fm95!
 
 ## Usage of other projects
 

fm95.md

@@ -35,10 +35,6 @@ Simple integer, set to 1 to output a sine wave at 400 Hz, which should correspon
 
 Float value to set the master volume, 1 by default, set to 0 to mute MPX
 
-### audio_volume
-
-Float value to set the audio volume, also 1 by default, set to 0 to mute audio itself, but stereo pilot and RDS should remain
-
 ### audio_preamp
 
 Pre-amplification of the audio, before any filters, set to 1 by default, set to 0 to mute audio

modulation/stereo_encoder.c

@@ -5,7 +5,7 @@ void init_stereo_encoder(StereoEncoder* st, uint8_t stereo_ssb, uint8_t multipli
     st->multiplier = multiplier;
     st->osc = osc;
     st->pilot_volume = pilot_volume;
-    st->audio_volume = audio_volume;
+    st->audio_volume = audio_volume * 0.5f;
     if(stereo_ssb) {
         init_delay_line(&st->delay_pilot, stereo_ssb*2);
         init_delay_line(&st->delay, stereo_ssb*2);
@@ -15,7 +15,10 @@ void init_stereo_encoder(StereoEncoder* st, uint8_t stereo_ssb, uint8_t multipli
 
 float stereo_encode(StereoEncoder* st, uint8_t enabled, float left, float right, float *audio) {
     float mid = (left+right) * 0.5f;
-    if(!enabled) return mid * st->audio_volume;
+    if(!enabled) {
+        *audio = mid * 2.0f * st->audio_volume;
+        return 0.0f;
+    }
 
     if(st->stereo_hilbert) mid = delay_line(&st->delay, mid);
 
@@ -33,12 +36,12 @@ float stereo_encode(StereoEncoder* st, uint8_t enabled, float left, float right,
     }
     float signalx2 = get_oscillator_sin_multiplier_ni(st->osc, st->multiplier * 2.0f);
 
-    *audio = (mid*st->audio_volume);
+    *audio = mid * st->audio_volume;
     if(st->stereo_hilbert) {
         float stereo = (crealf(stereo_hilbert) * signalx2) - (cimagf(stereo_hilbert) * signalx2cos);
-        *audio += (stereo * st->audio_volume);
+        *audio += stereo * st->audio_volume;
-    } else *audio += ((side*signalx2) * st->audio_volume);
+    } else *audio += (side*signalx2) * st->audio_volume;
-    return (signalx1*st->pilot_volume);
+    return signalx1 * st->pilot_volume;
 }
 
 void exit_stereo_encoder(StereoEncoder* st) {

src/fm95.c

@@ -18,19 +18,9 @@
 
 #include "../io/audio.h"
 
-#define DEFAULT_PILOT_VOLUME 0.09f // 9%
-#define DEFAULT_RDS_VOLUME 0.0475f // 4.75%
-#define DEFAULT_RDS_VOLUME_STEP 0.9f // 90%, so RDS2 stream 4 is 90% of stream 3 which is 90% of stream 2, which again is 90% of stream 1...
-
 static volatile sig_atomic_t to_run = 1;
 static volatile sig_atomic_t to_reload = 0;
 
-static inline float soft_clip(float x) { return tanhf(x); }
-
-static inline float soft_clip_drive(float x, float drive) {
-    return tanhf(x * drive) / tanhf(drive);
-}
-
 typedef struct {
 	bool rds_on;
 	bool mpx_on;
@@ -59,7 +49,6 @@ typedef struct {
 	float mpx_deviation;
 	float audio_deviation;
 	float master_volume;
-	float audio_volume;
 	float audio_preamp;
 
 	uint32_t sample_rate;
@@ -104,10 +93,7 @@ typedef struct {
 } FM95_SetupContext;
 
 bool compare_dvs(const FM95_DeviceNames *a, const FM95_DeviceNames *b) {
-    return strcmp(a->input,  b->input)  == 0 &&
+    return strcmp(a->input, b->input) == 0 && strcmp(a->output, b->output) == 0 && strcmp(a->mpx, b->mpx) == 0 && strcmp(a->rds, b->rds) == 0;
-           strcmp(a->output, b->output) == 0 &&
-           strcmp(a->mpx,    b->mpx)    == 0 &&
-           strcmp(a->rds,    b->rds)    == 0;
 }
 
 float calculate_sharedaudio_volume(const FM95_Volumes volumes, const int rds_streams) {
@@ -169,7 +155,7 @@ int run_fm95(const FM95_Config config, FM95_Runtime* runtime) {
 				else if(config.calibration == 3) sample *= (19000/config.mpx_deviation);
 				output[i] = sample*config.master_volume;
 			}
-			if((pulse_error = write_PulseOutputDevice(&runtime->output_device, output, sizeof(output)))) { // get output from the function and assign it into pulse_error, this comment to avoid confusion
+			if((pulse_error = write_PulseOutputDevice(&runtime->output_device, output, sizeof(output)))) {
 				fprintf(stderr, "Error writing to output device: %s\n", pa_strerror(pulse_error));
 				to_run = 0;
 				break;
@@ -185,6 +171,8 @@ int run_fm95(const FM95_Config config, FM95_Runtime* runtime) {
 	bool mpx_on = config.options.mpx_on;
 	bool rds_on = config.options.rds_on;
 
+	float softclip_norm = config.volumes.makeup / tanhf(config.volumes.drive);
+
 	while (to_run) {
 		if((pulse_error = read_PulseInputDevice(&runtime->input_device, audio_stereo_input, sizeof(audio_stereo_input)))) { // get output from the function and assign it into pulse_error, this comment to avoid confusion
 			fprintf(stderr, "Error reading from input device: %s\n", pa_strerror(pulse_error));
@@ -229,8 +217,8 @@ int run_fm95(const FM95_Config config, FM95_Runtime* runtime) {
 				mod_r = apply_preemphasis(&runtime->preemp_r, mod_r);
 			}
 
-			mod_l = soft_clip_drive(mod_l, config.volumes.drive) * config.volumes.makeup;
+			mod_l = tanhf(mod_l * config.volumes.drive) * softclip_norm;
-			mod_r = soft_clip_drive(mod_r, config.volumes.drive) * config.volumes.makeup;
+			mod_r = tanhf(mod_r * config.volumes.drive) * softclip_norm;
 
 			mpx = stereo_encode(&runtime->stencode, config.stereo, mod_l, mod_r, &audio);
 
@@ -250,7 +238,7 @@ int run_fm95(const FM95_Config config, FM95_Runtime* runtime) {
 
 			mpx = bs412_compress(&runtime->bs412, audio, mpx+mpx_in[i]);
 
-			output[i] = soft_clip(mpx)*config.master_volume; // Ensure peak deviation of 75 khz (or the set deviation), assuming we're calibrated correctly
+			output[i] = tanhf(mpx)*config.master_volume; // Ensure peak deviation of 75 khz (or the set deviation), assuming we're calibrated correctly
 			advance_oscillator(&runtime->osc);
 		}
 
@@ -296,9 +284,8 @@ static int config_handler(void* user, const char* section, const char* name, con
 
     #define MATCH(s, n) strcmp(section, s) == 0 && strcmp(name, n) == 0
 
-    if (MATCH("fm95", "stereo")) {
+    if (MATCH("fm95", "stereo")) pconfig->stereo = atoi(value);
-        pconfig->stereo = atoi(value);
+	else if (MATCH("devices", "input")) {
-	} else if (MATCH("devices", "input")) {
         strncpy(dv->input, value, 63);
         dv->input[63] = '\0';
     } else if (MATCH("devices", "output")) {
@@ -316,47 +303,26 @@ static int config_handler(void* user, const char* section, const char* name, con
             printf("RDS Streams more than 4? Nuh uh\n");
             return 0;
         }
-    } else if (MATCH("fm95", "preemphasis")) {
+	} else if (MATCH("fm95", "preemphasis")) pconfig->preemphasis = atoi(value);
-        pconfig->preemphasis = atoi(value);
+    else if (MATCH("fm95", "calibration")) pconfig->calibration = atoi(value);
-    } else if (MATCH("fm95", "calibration")) {
+    else if (MATCH("fm95", "mpx_power")) pconfig->mpx_power = strtof(value, NULL);
-        pconfig->calibration = atoi(value);
+    else if (MATCH("fm95", "mpx_deviation")) pconfig->mpx_deviation = strtof(value, NULL);
-    } else if (MATCH("fm95", "mpx_power")) {
+    else if (MATCH("fm95", "master_volume")) pconfig->master_volume = strtof(value, NULL);
-        pconfig->mpx_power = strtof(value, NULL);
+    else if (MATCH("fm95", "audio_preamp")) pconfig->audio_preamp = strtof(value, NULL);
-    } else if (MATCH("fm95", "mpx_deviation")) {
+    else if (MATCH("fm95", "deviation")) pconfig->audio_deviation = strtof(value, NULL);
-        pconfig->mpx_deviation = strtof(value, NULL);
+	else if(MATCH("fm95", "agc_target")) pconfig->agc_target = strtof(value, NULL);
-    } else if (MATCH("fm95", "master_volume")) {
+	else if(MATCH("fm95", "agc_attack")) pconfig->agc_attack = strtof(value, NULL);
-        pconfig->master_volume = strtof(value, NULL);
+	else if(MATCH("fm95", "agc_release")) pconfig->agc_release = strtof(value, NULL);
-    } else if (MATCH("fm95", "audio_volume")) {
+	else if(MATCH("fm95", "agc_min")) pconfig->agc_min = strtof(value, NULL);
-        pconfig->audio_volume = strtof(value, NULL);
+	else if(MATCH("fm95", "agc_max")) pconfig->agc_max = strtof(value, NULL);
-    } else if (MATCH("fm95", "audio_preamp")) {
+	else if(MATCH("fm95", "bs412_attack")) pconfig->bs412_attack = strtof(value, NULL);
-        pconfig->audio_preamp = strtof(value, NULL);
+	else if(MATCH("fm95", "bs412_release")) pconfig->bs412_release = strtof(value, NULL);
-    } else if (MATCH("fm95", "deviation")) {
+	else if(MATCH("fm95", "bs412_max")) pconfig->bs412_max = strtof(value, NULL);
-        pconfig->audio_deviation = strtof(value, NULL);
+	else if(MATCH("advanced", "lpf_order")) pconfig->lpf_order = atoi(value);
-	} else if(MATCH("fm95", "agc_target")) {
+	else if(MATCH("advanced", "stereo_ssb")) pconfig->stereo_ssb = atoi(value);
-		pconfig->agc_target = strtof(value, NULL);
+	else if(MATCH("advanced", "preemp_unity")) pconfig->preemp_unity_freq = strtof(value, NULL);
-	} else if(MATCH("fm95", "agc_attack")) {
+	else if(MATCH("advanced", "sample_rate")) pconfig->sample_rate = atoi(value);
-		pconfig->agc_attack = strtof(value, NULL);
+	else if(MATCH("advanced", "lpf_cutoff")) {
-	} else if(MATCH("fm95", "agc_release")) {
-		pconfig->agc_release = strtof(value, NULL);
-	} else if(MATCH("fm95", "agc_min")) {
-		pconfig->agc_min = strtof(value, NULL);
-	} else if(MATCH("fm95", "agc_max")) {
-		pconfig->agc_max = strtof(value, NULL);
-	} else if(MATCH("fm95", "bs412_attack")) {
-		pconfig->bs412_attack = strtof(value, NULL);
-	} else if(MATCH("fm95", "bs412_release")) {
-		pconfig->bs412_release = strtof(value, NULL);
-	} else if(MATCH("fm95", "bs412_max")) {
-		pconfig->bs412_max = strtof(value, NULL);
-	} else if(MATCH("advanced", "lpf_order")) {
-		pconfig->lpf_order = atoi(value);
-	} else if(MATCH("advanced", "stereo_ssb")) {
-		pconfig->stereo_ssb = atoi(value);
-	} else if(MATCH("advanced", "preemp_unity")) {
-		pconfig->preemp_unity_freq = strtof(value, NULL);
-	} else if(MATCH("advanced", "sample_rate")) {
-		pconfig->sample_rate = atoi(value);
-	} else if(MATCH("advanced", "lpf_cutoff")) {
 		pconfig->lpf_cutoff = strtof(value, NULL);
 		if(pconfig->lpf_cutoff > (pconfig->sample_rate * 0.5)) {
 			pconfig->lpf_cutoff = (pconfig->sample_rate * 0.5);
@@ -491,13 +457,13 @@ void init_runtime(FM95_Runtime* runtime, const FM95_Config config) {
 }
 
 int main(int argc, char **argv) {
-	printf("fm95 (an FM Processor by radio95) version 2.4\n");
+	printf("fm95 (an FM Processor by radio95) version 2.5\n");
 
 	FM95_Config config = {
 		.volumes = {
-			.pilot = DEFAULT_PILOT_VOLUME,
+			.pilot = 0.09f,
-			.rds = DEFAULT_RDS_VOLUME,
+			.rds = 0.045f,
-			.rds_step = DEFAULT_RDS_VOLUME_STEP,
+			.rds_step = 0.9f,
 			.headroom = 0.05f,
 			.drive = 1.0f,
 			.makeup = 1.0f
@@ -513,7 +479,6 @@ int main(int argc, char **argv) {
 		.mpx_deviation = 75000.0f, // for BS412, this is what deviation does the compressor see as peak, so if i set here 150 khz, then the compressor will act as if it was two times louder
 		.audio_deviation = 75000.0f, // another way to set the volume
 		.master_volume = 1.0f, // Volume of everything combined, for calibration
-		.audio_volume = 1.0f, // Volume of the audio, before stereo encoding, before clipper
 		.audio_preamp = 1.0f, // Volume of the audio before the filters
 
 		.sample_rate = 192000, // Sample rate for this whole gizmo to run on
@@ -560,6 +525,8 @@ int main(int argc, char **argv) {
 		return 1;
 	}
 
+	if (config.volumes.drive < 0.01f) config.volumes.drive = 0.01f;
+
 	config.master_volume *= config.audio_deviation/75000.0f;
 
 	config.volumes.audio = calculate_sharedaudio_volume(config.volumes, config.rds_streams);