Small API change of Synth_Dexed.

dexed_sd.cpp

@@ -74,7 +74,7 @@ bool load_sd_voice(uint8_t b, uint8_t v, uint8_t instance_id)
       Serial.println(F("]"));
 #endif
       uint8_t tmp_data[156];
-      bool ret = MicroDexed[instance_id]->decodeVoice(tmp_data, tmp);
+      bool ret = MicroDexed[instance_id]->decodeVoice(tmp_data, data);
       MicroDexed[instance_id]->loadVoiceParameters(tmp_data);
 #ifdef DEBUG
       show_patch(instance_id);

third-party/Synth_Dexed/README.md

@@ -36,7 +36,7 @@ See example sketch.
     void setMaxNotes(uint8_t n);
     uint8_t getMaxNotes(void);
     void doRefreshVoice(void);
-    bool decodeVoice(uint8_t* encoded_data, uint8_t* data);
+    bool decodeVoice(uint8_t* data, uint8_t* encoded_data);
     bool encodeVoice(uint8_t* encoded_data);
     bool getVoiceData(uint8_t* data_copy);
     void setVoiceDataElement(uint8_t address, uint8_t value);
@@ -169,4 +169,4 @@ See example sketch.
     void setTranspose(uint8_t transpose);
     uint8_t getTranspose(void);
     void setName(char* name);
-    void getName(char* buffer);
+    void getName(char* buffer);

third-party/Synth_Dexed/examples/Banks/Banks.ino

@@ -0,0 +1,65 @@
+#include <Audio.h>
+#include "synth_dexed.h"
+#include "banks.h"
+
+AudioSynthDexed          dexed(4, SAMPLE_RATE); // 4 voices max
+AudioOutputI2S           i2s1;
+AudioControlSGTL5000     sgtl5000_1;
+AudioConnection          patchCord1(dexed, 0, i2s1, 0);
+AudioConnection          patchCord2(dexed, 0, i2s1, 1);
+
+void setup()
+{
+  Serial.begin(230400);
+
+  AudioMemory(32);
+
+  sgtl5000_1.enable();
+  sgtl5000_1.lineOutLevel(29);
+  sgtl5000_1.dacVolumeRamp();
+  sgtl5000_1.dacVolume(1.0);
+  sgtl5000_1.unmuteHeadphone();
+  sgtl5000_1.unmuteLineout();
+  sgtl5000_1.volume(0.8, 0.8); // Headphone volume
+
+  randomSeed(analogRead(0));
+}
+
+void loop()
+{
+  uint8_t b = random(0, 10);
+  uint8_t v = random(0, 32);
+  uint8_t t = random(0, 36);
+  char voice_name[11];
+  uint8_t decoded_voice[156];
+
+  memset(voice_name, 0, 11);
+  memcpy(voice_name, &progmem_bank[b][v][117], 10);
+
+  Serial.print("Voice: ");
+  Serial.print(voice_name);
+  Serial.print(" ");
+  Serial.print("Transpose: ");
+  Serial.println(t);
+
+  dexed.decodeVoice(progmem_bank[b][v],decoded_voice); 
+  dexed.loadVoiceParameters(decoded_voice);
+  dexed.setTranspose(t);
+
+  Serial.println("Key-Down");
+  dexed.keydown(48, 100);
+  delay(100);
+  dexed.keydown(52, 100);
+  delay(100);
+  dexed.keydown(55, 100);
+  delay(100);
+  dexed.keydown(60, 100);
+  delay(2000);
+
+  Serial.println("Key-Up");
+  dexed.keyup(48);
+  dexed.keyup(52);
+  dexed.keyup(55);
+  dexed.keyup(60);
+  delay(1000);
+}

third-party/Synth_Dexed/src/synth_dexed.cpp

@@ -751,7 +751,7 @@ uint8_t Dexed::getNumNotesPlaying(void)
   return (count_playing_voices);
 }
 
-bool Dexed::decodeVoice(uint8_t* encoded_data, uint8_t* new_data)
+bool Dexed::decodeVoice(uint8_t* new_data, uint8_t* encoded_data)
 {
   uint8_t* p_data = new_data;
   uint8_t op;

third-party/Synth_Dexed/src/synth_dexed.h

@@ -1126,7 +1126,7 @@ class Dexed
     uint8_t getMaxNotes(void);
     void doRefreshVoice(void);
     void setOPAll(uint8_t ops);
-    bool decodeVoice(uint8_t* encoded_data, uint8_t* data);
+    bool decodeVoice(uint8_t* data, uint8_t* encoded_data);
     bool encodeVoice(uint8_t* encoded_data);
     bool getVoiceData(uint8_t* data_copy);
     void setVoiceDataElement(uint8_t address, uint8_t value);

third-party/Synth_Dexed/tools/sysex2c.py

@@ -0,0 +1,148 @@
+#!/usr/bin/python3
+
+import sys
+import os.path
+
+# From: https://github.com/bwhitman/learnfm/blob/master/dx7db.py
+def unpack_packed_patch(p):
+    # Input is a 128 byte thing from compact.bin
+    # Output is a 156 byte thing that the synth knows about
+    o = [0]*156
+    for op in range(6):
+        o[op*21:op*21 + 11] = p[op*17:op*17+11]
+        leftrightcurves = p[op*17+11]
+        o[op * 21 + 11] = leftrightcurves & 3
+        o[op * 21 + 12] = (leftrightcurves >> 2) & 3
+        detune_rs = p[op * 17 + 12]
+        o[op * 21 + 13] = detune_rs & 7
+        o[op * 21 + 20] = detune_rs >> 3
+        kvs_ams = p[op * 17 + 13]
+        o[op * 21 + 14] = kvs_ams & 3
+        o[op * 21 + 15] = kvs_ams >> 2
+        o[op * 21 + 16] = p[op * 17 + 14]
+        fcoarse_mode = p[op * 17 + 15]
+        o[op * 21 + 17] = fcoarse_mode & 1
+        o[op * 21 + 18] = fcoarse_mode >> 1
+        o[op * 21 + 19] = p[op * 17 + 16]
+    
+    o[126:126+9] = p[102:102+9]
+    oks_fb = p[111]
+    o[135] = oks_fb & 7
+    o[136] = oks_fb >> 3
+    o[137:137+4] = p[112:112+4]
+    lpms_lfw_lks = p[116]
+    o[141] = lpms_lfw_lks & 1
+    o[142] = (lpms_lfw_lks >> 1) & 7
+    o[143] = lpms_lfw_lks >> 4
+    o[144:144+11] = p[117:117+11]
+    o[155] = 0x3f
+
+    # Clamp the unpacked patches to a known max. 
+    maxes =  [
+        99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, # osc6
+        3, 3, 7, 3, 7, 99, 1, 31, 99, 14,
+        99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, # osc5
+        3, 3, 7, 3, 7, 99, 1, 31, 99, 14,
+        99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, # osc4
+        3, 3, 7, 3, 7, 99, 1, 31, 99, 14,
+        99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, # osc3
+        3, 3, 7, 3, 7, 99, 1, 31, 99, 14,
+        99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, # osc2
+        3, 3, 7, 3, 7, 99, 1, 31, 99, 14,
+        99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, # osc1
+        3, 3, 7, 3, 7, 99, 1, 31, 99, 14,
+        99, 99, 99, 99, 99, 99, 99, 99, # pitch eg rate & level 
+        31, 7, 1, 99, 99, 99, 99, 1, 5, 7, 48, # algorithm etc
+        126, 126, 126, 126, 126, 126, 126, 126, 126, 126, # name
+        127 # operator on/off
+    ]
+    for i in range(156):
+        if(o[i] > maxes[i]): o[i] = maxes[i]
+        if(o[i] < 0): o[i] = 0
+    return o
+
+def print_header_data(voice_data):
+	print("\t\t\t",end="")
+	for y in range(0,len(voice_data)):
+		if(y!=len(voice_data)-1):
+			print("%3d, " % voice_data[y],end="")
+		else:
+			print("%3d\n" % voice_data[y],end="")
+		if((y+1)%8==0 and y!=len(voice_data)-1):
+			print("\n\t\t\t",end="")
+
+def help_message():
+	print(progname+" [--decode] <sysex1> [<sysex2> ... <sysexn>]")
+
+#---------------------------------------------------------------------------
+
+progname=sys.argv.pop(0)
+if(len(sys.argv)==0):
+	help_message()
+	exit(1)
+if(sys.argv[0]=="-h" or sys.argv[0]=="--help"):
+	help_message()
+	exit(1)
+if(sys.argv[0]=="--decode"):
+	decode=True
+	sys.argv.pop(0)
+else:
+	decode=False
+
+print("""
+//
+// File generated with sysex2c.py
+//
+""")
+if(decode==True):
+	print("uint8_t progmem_bank[%d][32][156] PROGMEM =\n{" % int(len(sys.argv)))
+else:
+	print("uint8_t progmem_bank[%d][32][128] PROGMEM =\n{" % int(len(sys.argv)))
+for sysex in sys.argv:
+	if(not os.path.isfile(sysex)):
+		print("* File "+sysex+" does not exists.")
+		exit(10)
+	if(not os.access(sysex,os.R_OK)):
+		print("* File "+sysex+" does not readable.")
+		exit(11)
+	print("\t{\t// %s" % os.path.basename(sysex))
+	with open(sysex, "rb") as f:
+		header = f.read(6)
+		if(header[0]!=240):
+			print("* %s: Start of sysex not found." % sysex)
+			exit(200)
+		if(header[1]!=67):
+                	print("* %s: Manufactorer-ID not Yamaha." % sysex)
+                	exit(201)
+		if(header[3]!=9):
+                	print("* %s: Not a 32 voice sysex file." % sysex)
+                	exit(202)
+		byte_count = header[4]*128+header[5]
+		if(byte_count!=4096):
+                	print("* %s: Byte count mismatch." % sysex)
+                	exit(203)
+		patch_data=f.read(4096)
+		check = ~sum(patch_data) + 1 & 0x7F
+		f.seek(4102) # Bulk checksum
+		checksum=int.from_bytes(f.read(1),"little")
+		if(check!=checksum):
+			print("* %s: Checksum mismatch!" % sysex)
+			exit(204)
+		f.seek(6)
+
+		for v in range(1,33):
+			data=f.read(128)
+			print("\t\t{\t// %d: %s" % (v, data[118:128].decode('ascii')))
+			if(decode==True):
+				print_header_data(unpack_packed_patch(data))
+			else:
+				print_header_data(data)
+			if(v!=32):
+				print("\t\t},")
+			else:
+				print("\t\t}")
+		if(sys.argv[len(sys.argv)-1]==sysex):
+			print("\t}")
+		else:
+			print("\t},")
+print("};")