Add support for no AI and preprocessed stems

.gitignore

@@ -3,3 +3,4 @@ test.m4a
 separated
 srdr_work
 __pycache__
+stems

base.sds

@@ -1,4 +1,4 @@
-%Srdr-1.0
+%Srdr-1.1
 
 '
 ' Surrounder base scene file
@@ -13,19 +13,9 @@ Chs: CHANNELS_5_1; 16; 44100
 Lay: LAYOUT_5_1
 
 ' Splitting into stems
-Sep: STEMS_ALL
+Sep: STEMS_SIMPLE
 
 ' Original version of the song in the front channels
-Map: STEMS_VOCALS_L; 0; 1
-Map: STEMS_VOCALS_R; 1; 1
-Map: STEMS_PIANO_L; 0; 1
-Map: STEMS_PIANO_R; 1; 1
-Map: STEMS_GUITAR_L; 0; 1
-Map: STEMS_GUITAR_R; 1; 1
-Map: STEMS_BASS_L; 0; 1
-Map: STEMS_BASS_R; 1; 1
-Map: STEMS_DRUMS_L; 0; 1
-Map: STEMS_DRUMS_R; 1; 1
 Map: STEMS_OTHER_L; 0; 1
 Map: STEMS_OTHER_R; 1; 1
 
@@ -38,18 +28,11 @@ Map: STEMS_LFE_L; 3; 1
 Map: STEMS_LFE_R; 3; 1
 
 ' Quieter version of the song in the back channels
-Map: STEMS_VOCALS_L; 4; 0.75
-Map: STEMS_VOCALS_R; 5; 0.75
-Map: STEMS_PIANO_L; 4; 0.75
-Map: STEMS_PIANO_R; 5; 0.75
-Map: STEMS_GUITAR_L; 4; 0.75
-Map: STEMS_GUITAR_R; 5; 0.75
-Map: STEMS_BASS_L; 4; 0.75
-Map: STEMS_BASS_R; 5; 0.75
-Map: STEMS_DRUMS_L; 4; 0.75
-Map: STEMS_DRUMS_R; 5; 0.75
 Map: STEMS_OTHER_L; 4; 0.75
 Map: STEMS_OTHER_R; 5; 0.75
 
+' New in 0.2.0: Adding distance between the back channels
+Dis: 4; 5
+
 ' FLAC output
 Out: OUTPUT_FLAC

base_ai.sds

@@ -0,0 +1,58 @@
+%Srdr-1.1
+
+'
+' Surrounder base scene file
+'
+' This is the primary scene file tested during development
+' of Surrounder. It produces 5.1 surround output from stereo
+' input.
+'
+
+' Base configuration
+Chs: CHANNELS_5_1; 16; 44100
+Lay: LAYOUT_5_1
+
+' Splitting into stems
+Sep: STEMS_ALL
+
+' Original version of the song in the front channels
+Map: STEMS_VOCALS_L; 0; 1
+Map: STEMS_VOCALS_R; 1; 1
+Map: STEMS_PIANO_L; 0; 1
+Map: STEMS_PIANO_R; 1; 1
+Map: STEMS_GUITAR_L; 0; 1
+Map: STEMS_GUITAR_R; 1; 1
+Map: STEMS_BASS_L; 0; 1
+Map: STEMS_BASS_R; 1; 1
+Map: STEMS_DRUMS_L; 0; 1
+Map: STEMS_DRUMS_R; 1; 1
+Map: STEMS_OTHER_L; 0; 1
+Map: STEMS_OTHER_R; 1; 1
+
+' Vocals on the center channel
+Map: STEMS_VOCALS_L; 2; 0.5
+Map: STEMS_VOCALS_R; 2; 0.5
+
+' Low frequency elements
+Map: STEMS_LFE_L; 3; 1
+Map: STEMS_LFE_R; 3; 1
+
+' Quieter version of the song in the back channels
+Map: STEMS_VOCALS_L; 4; 0.75
+Map: STEMS_VOCALS_R; 5; 0.75
+Map: STEMS_PIANO_L; 4; 0.75
+Map: STEMS_PIANO_R; 5; 0.75
+Map: STEMS_GUITAR_L; 4; 0.75
+Map: STEMS_GUITAR_R; 5; 0.75
+Map: STEMS_BASS_L; 4; 0.75
+Map: STEMS_BASS_R; 5; 0.75
+Map: STEMS_DRUMS_L; 4; 0.75
+Map: STEMS_DRUMS_R; 5; 0.75
+Map: STEMS_OTHER_L; 4; 0.75
+Map: STEMS_OTHER_R; 5; 0.75
+
+' New in 0.2.0: Adding distance to the back channels
+Dis: 4; 5
+
+' FLAC output
+Out: OUTPUT_FLAC

base_manual.sds

@@ -0,0 +1,58 @@
+%Srdr-1.1
+
+'
+' Surrounder base scene file
+'
+' This is the primary scene file tested during development
+' of Surrounder. It produces 5.1 surround output from stereo
+' input.
+'
+
+' Base configuration
+Chs: CHANNELS_5_1; 16; 44100
+Lay: LAYOUT_5_1
+
+' Splitting into stems
+Sep: STEMS_MANUAL
+
+' Original version of the song in the front channels
+Map: STEMS_VOCALS_L; 0; 1
+Map: STEMS_VOCALS_R; 1; 1
+Map: STEMS_PIANO_L; 0; 1
+Map: STEMS_PIANO_R; 1; 1
+Map: STEMS_GUITAR_L; 0; 1
+Map: STEMS_GUITAR_R; 1; 1
+Map: STEMS_BASS_L; 0; 1
+Map: STEMS_BASS_R; 1; 1
+Map: STEMS_DRUMS_L; 0; 1
+Map: STEMS_DRUMS_R; 1; 1
+Map: STEMS_OTHER_L; 0; 1
+Map: STEMS_OTHER_R; 1; 1
+
+' Vocals on the center channel
+Map: STEMS_VOCALS_L; 2; 0.5
+Map: STEMS_VOCALS_R; 2; 0.5
+
+' Low frequency elements
+Map: STEMS_LFE_L; 3; 1
+Map: STEMS_LFE_R; 3; 1
+
+' Quieter version of the song in the back channels
+Map: STEMS_VOCALS_L; 4; 0.75
+Map: STEMS_VOCALS_R; 5; 0.75
+Map: STEMS_PIANO_L; 4; 0.75
+Map: STEMS_PIANO_R; 5; 0.75
+Map: STEMS_GUITAR_L; 4; 0.75
+Map: STEMS_GUITAR_R; 5; 0.75
+Map: STEMS_BASS_L; 4; 0.75
+Map: STEMS_BASS_R; 5; 0.75
+Map: STEMS_DRUMS_L; 4; 0.75
+Map: STEMS_DRUMS_R; 5; 0.75
+Map: STEMS_OTHER_L; 4; 0.75
+Map: STEMS_OTHER_R; 5; 0.75
+
+' New in 0.2.0: Adding distance to the back channels
+Dis: 4; 5
+
+' FLAC output
+Out: OUTPUT_FLAC

constants.py

@@ -12,7 +12,8 @@ def value_to_constant(value, filter_str=None):
     return keys[0]
 
 Versions = [
-    "1.0"
+    "1.0",
+    "1.1"
 ]
 
 Channels = [
@@ -122,6 +123,8 @@ def value_to_constant(value, filter_str=None):
     'LAYOUT_ATMOS_11_1_4': 'FL; FR; FC; LFE; BL; BR; FLC; FRC; BC; SL; SR; TBL; TBR',
 
     # Stem configurations
+    'STEMS_MANUAL': -2,  # Supported for Sep
+    'STEMS_SIMPLE': -1,  # Supported for Sep
     'STEMS_ALL': 0,  # Supported for Sep
     'STEMS_VOCALS': 1,  # Supported for Sep
     'STEMS_BASS': 2,  # Supported for Sep

legacy.sds

@@ -1,4 +1,4 @@
-%Srdr-1.0
+%Srdr-1.1
 
 Chs: CHANNELS_5_1; 16; 44100
 Sep: STEMS_ALL

main.py

@@ -4,7 +4,6 @@
 import shutil
 import subprocess
 import json
-import demucs.separate
 
 parser = argparse.ArgumentParser(
     prog='surrounder',
@@ -21,6 +20,10 @@
     print(f'Input file {args.input} not found')
     exit(1)
 
+if os.path.isdir(args.input):
+    print(f"Notice: {args.input} is a directory, using the first file in this directory as the reference input.")
+    args.input = list(os.scandir(args.input))[0].path
+
 if not os.path.exists(args.scene):
     print(f'Scene file {args.scene} not found')
     exit(1)
@@ -102,6 +105,10 @@
 for (name, parameters) in operations:
     match name:
         case "Chs":
+            if len(parameters) != 3:
+                print(f"Expected 3 parameters but got {len(parameters)}.")
+                exit(2)
+
             try:
                 channels = int(parameters[0])
 
@@ -168,6 +175,10 @@
                     f"./srdr_work/channels/{i}.wav"
                 ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
         case "Sep":
+            if len(parameters) != 1:
+                print(f"Expected 1 parameter but got {len(parameters)}.")
+                exit(2)
+
             if not os.path.exists(f"./srdr_work/input.wav"):
                 print(f"Separation is not ready: not configured.")
                 exit(2)
@@ -180,7 +191,15 @@
                 print(f"Invalid stem format value: {parameters[0]}")
                 exit(2)
 
-            if stem_format != 0:
+            if stem_format == -2:
+                if not os.path.exists(f"./stems"):
+                    print("Could not find stems in a 'stems' folder.")
+                    exit(2)
+
+                shutil.copytree(f"./stems", f"./srdr_work/stems")
+            elif stem_format == -1:
+                pass
+            elif stem_format != 0:
                 match stem_format:
                     case 1:
                         print("Isolating vocals")
@@ -206,11 +225,26 @@
 
             args += ["./srdr_work/input.wav"]
 
-            print("Separating, this might take a while.")
-            demucs.separate.main(args)
-
-            os.rename("./srdr_work/stems_tmp/htdemucs_6s/input", "./srdr_work/stems")
-            shutil.rmtree('./srdr_work/stems_tmp')
+            if stem_format > -1:
+                import demucs.separate
+                print("Separating using machine learning, this might take a while.")
+                demucs.separate.main(args)
+                os.rename("./srdr_work/stems_tmp/htdemucs_6s/input", "./srdr_work/stems")
+                shutil.rmtree('./srdr_work/stems_tmp')
+            elif stem_format == -1:
+                os.mkdir("./srdr_work/stems")
+                subprocess.run([
+                    "sox",
+                    "./srdr_work/input.wav",
+                    "./srdr_work/stems/other.wav"
+                ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+                subprocess.run([
+                    "sox",
+                    "./srdr_work/input.wav",
+                    "./srdr_work/stems/vocals.wav",
+                    "highpass", "300",
+                    "lowpass", "3500"
+                ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
 
             if os.path.exists("./srdr_work/stems/vocals.wav"):
                 os.rename("./srdr_work/stems/vocals.wav", "./srdr_work/stems/1.wav")
@@ -302,6 +336,10 @@
                 "./srdr_work/stems/21.wav"
             ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
         case "Map":
+            if len(parameters) != 3:
+                print(f"Expected 3 parameters but got {len(parameters)}.")
+                exit(2)
+
             try:
                 stem = int(parameters[0])
             except ValueError:
@@ -357,6 +395,69 @@
             os.unlink(f"./srdr_work/channels/{channel}.wav")
             os.unlink(f"./srdr_work/stems/{stem}_2.wav")
             os.rename(f"./srdr_work/channels/{channel}_2.wav", f"./srdr_work/channels/{channel}.wav")
+        case "Dis":
+            if len(parameters) != 2:
+                print(f"Expected 2 parameters but got {len(parameters)}.")
+                exit(2)
+
+            try:
+                channel1 = int(parameters[0])
+            except ValueError:
+                print(f"Invalid channel value: {parameters[0]}")
+                exit(2)
+
+            try:
+                channel2 = int(parameters[1])
+            except ValueError:
+                print(f"Invalid channel value: {parameters[1]}")
+                exit(2)
+
+            if not os.path.exists(f"./srdr_work/channels/{channel1}.wav"):
+                print(f"Invalid or nonexistent channel: {channel1}")
+                exit(2)
+
+            if not os.path.exists(f"./srdr_work/channels/{channel2}.wav"):
+                print(f"Invalid or nonexistent channel: {channel2}")
+                exit(2)
+
+            if layout is not None:
+                print(f"Distance: Channel {channel1} ({layout[channel1]}) <-> Channel {channel2} ({layout[channel2]})")
+            else:
+                print(f"Distance: Channel {channel1} <-> Channel {channel2}")
+
+            subprocess.run([
+                "sox",
+                "-M",
+                f"./srdr_work/channels/{channel1}.wav",
+                f"./srdr_work/channels/{channel2}.wav",
+                f"./srdr_work/channels/_2.wav",
+            ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+
+            subprocess.run([
+                "sox",
+                f"./srdr_work/channels/_2.wav",
+                f"./srdr_work/channels/_3.wav",
+                "remix", "1v-0.8718,2v0.4898", "1v0.4898,2v-0.8718",
+                "delay", "0.15"
+            ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+
+            subprocess.run([
+                "ffmpeg",
+                "-y",
+                "-i", "./srdr_work/channels/_3.wav",
+                "-filter_complex", "[0:0]pan=1|c0=c0[left];[0:0]pan=1|c0=c1[right]",
+                "-map", "[left]",
+                f"./srdr_work/channels/{channel1}_2.wav",
+                "-map", "[right]",
+                f"./srdr_work/channels/{channel2}_2.wav",
+            ], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+
+            os.unlink(f"./srdr_work/channels/{channel1}.wav")
+            os.unlink(f"./srdr_work/channels/{channel2}.wav")
+            os.unlink(f"./srdr_work/channels/_2.wav")
+            os.unlink(f"./srdr_work/channels/_3.wav")
+            os.rename(f"./srdr_work/channels/{channel1}_2.wav", f"./srdr_work/channels/{channel1}.wav")
+            os.rename(f"./srdr_work/channels/{channel2}_2.wav", f"./srdr_work/channels/{channel2}.wav")
         case "Lay":
             layout = parameters
 
@@ -367,6 +468,10 @@
 
             print(f"Channel layout: {', '.join(layout)}")
         case "Out":
+            if len(parameters) != 1:
+                print(f"Expected 1 parameter but got {len(parameters)}.")
+                exit(2)
+
             if not os.path.exists(f"./srdr_work/input.wav"):
                 print(f"Output is not ready: not configured.")
                 exit(2)