feat: audio.cpp backend support

1 files changed, 65 insertions, 25 deletions

diff --git a/audiobook.py b/audiobook.py
index b8c445e..c2639ce 100755
--- a/audiobook.py
+++ b/audiobook.py
@@ -27,7 +27,14 @@ from converter.converter import (
     setup_directories,
     setup_logging,
 )
-from converter.tts import VOICE_MODE_CLONE, VOICE_MODE_CUSTOM, normalize_language
+from converter.tts import (
+    BACKEND_AUDIOCPP,
+    BACKEND_FASTER,
+    BACKEND_GRADIO,
+    VOICE_MODE_CLONE,
+    VOICE_MODE_CUSTOM,
+    normalize_language,
+)
 
 
 def main() -> None:
@@ -37,14 +44,17 @@ def main() -> None:
         formatter_class=argparse.RawDescriptionHelpFormatter,
         epilog="""
 Examples:
-  # Use custom voice (default - Vivian speaker)
+  # Use the Qwen Gradio demo with a custom voice (default - Vivian speaker)
   python audiobook.py
 
-  # Use voice cloning with reference audio
+  # Use the Qwen Gradio demo with voice cloning from reference audio
   python audiobook.py --clone path/to/reference.wav
 
   # Use the faster-qwen3-tts server (voice cloning, configured server-side)
-  python audiobook.py --faster
+  python audiobook.py --backend faster [--voice NAME]
+
+  # Use an audio.cpp audiocpp_server (speaker mode, or a server-side voice)
+  python audiobook.py --backend audiocpp [--voice NAME]
         """
     )
 
@@ -105,23 +115,26 @@ Examples:
     )
 
     parser.add_argument(
-        "--faster",
-        action="store_true",
-        help=("Use a faster-qwen3-tts OpenAI-compatible server instead of the Qwen "
-              "Gradio demos (5-10x faster inference via CUDA graphs). Always voice "
-              "cloning: the reference audio and transcript are configured on the "
-              "server itself (see the 'Faster backend' section of the README).")
+        "--backend",
+        choices=[BACKEND_GRADIO, BACKEND_FASTER, BACKEND_AUDIOCPP],
+        default=config.BACKEND,
+        help=("TTS server to talk to: the Qwen3-TTS Gradio demos (gradio), the "
+              "faster-qwen3-tts OpenAI-compatible server (faster), or an "
+              "audio.cpp audiocpp_server (audiocpp). Defaults to the BACKEND "
+              "setting in converter/config.py (gradio).")
     )
 
     parser.add_argument(
-        "--faster-voice",
+        "--voice",
         type=str,
         default=None,
         metavar="NAME",
-        help=("Voice entry to request from the faster server's voice config "
-              "(default: the FASTER_VOICE setting in converter/config.py, "
-              "typically 'default'). Must match a key in the server's voices.json, "
-              "or 'default' when the server was started with --ref-audio.")
+        help=("Voice to request from a server-side voice configuration. faster: "
+              "a key in the server's voices.json ('default' when it was started "
+              "with --ref-audio). audiocpp: a voice_preset or voice_dir entry "
+              "(cloning); without this flag the audiocpp backend uses a built-in "
+              "CustomVoice speaker instead. Not used by the gradio backend "
+              "(use converter/config.py SPEAKER or --clone there).")
     )
 
     parser.add_argument(
@@ -137,23 +150,44 @@ Examples:
     if args.speed <= 0:
         parser.error(f"--speed must be a positive number (got {args.speed:g})")
 
-    if args.faster:
+    if args.backend == BACKEND_FASTER:
         if args.clone:
-            print("[WARNING] --clone is ignored with --faster: the faster backend "
+            print("[WARNING] --clone is ignored with --backend faster: that backend "
                   "always uses voice cloning, and the reference voice is configured "
-                  "on the faster server (see README)")
+                  "on the server (see README)")
             args.clone = None
         if args.transcription or args.no_transcription:
             print("[WARNING] --transcription/--no-transcription are ignored with "
-                  "--faster: the reference transcript is configured on the faster "
+                  "--backend faster: the reference transcript is configured on the "
                   "server (--ref-text or voices.json, see README)")
             args.transcription = None
             args.no_transcription = False
         if args.language is not None:
-            print("[WARNING] --language is ignored with --faster: language is "
-                  "configured on the faster server (see README)")
+            print("[WARNING] --language is ignored with --backend faster: language "
+                  "is configured on the server (see README)")
             args.language = None
+    elif args.backend == BACKEND_AUDIOCPP:
+        if args.clone:
+            print("[WARNING] --clone is ignored with --backend audiocpp: cloning "
+                  "uses a voice configured on the server (voice_presets or "
+                  "voice_dir in its config); select it with --voice (see README)")
+            args.clone = None
+        if args.transcription or args.no_transcription:
+            print("[WARNING] --transcription/--no-transcription are ignored with "
+                  "--backend audiocpp: the reference transcript is configured on "
+                  "the server (see README)")
+            args.transcription = None
+            args.no_transcription = False
+        if args.language is not None:
+            try:
+                args.language = normalize_language(args.language)
+            except ValueError as exc:
+                parser.error(str(exc))
     else:
+        if args.voice is not None:
+            parser.error("--voice requires --backend faster or audiocpp; the "
+                         "gradio backend uses built-in speakers "
+                         "(converter/config.py SPEAKER) or --clone")
         if args.language is not None:
             try:
                 args.language = normalize_language(args.language)
@@ -167,10 +201,16 @@ Examples:
     setup_logging(debug=args.debug)
     setup_directories()
 
+    if args.backend == BACKEND_FASTER:
+        voice_mode = VOICE_MODE_CLONE
+    elif args.backend == BACKEND_AUDIOCPP:
+        voice_mode = VOICE_MODE_CLONE if args.voice else VOICE_MODE_CUSTOM
+    else:
+        voice_mode = VOICE_MODE_CLONE if args.clone else VOICE_MODE_CUSTOM
+
     try:
         converter = AudiobookConverter(
-            voice_mode=VOICE_MODE_CLONE if (args.clone or args.faster)
-            else VOICE_MODE_CUSTOM,
+            voice_mode=voice_mode,
             voice_clone_ref_audio=args.clone,
             voice_clone_ref_text=args.transcription,
             skip_transcription=args.no_transcription,
@@ -178,8 +218,8 @@ Examples:
             single_file=args.single_file,
             output_format=args.format,
             language=args.language,
-            faster=args.faster,
-            faster_voice=args.faster_voice,
+            backend=args.backend,
+            voice=args.voice,
             debug=args.debug,
         )
         ok = converter.run()