Fix tts_with_vc (#3275)

* Revert "fix for issue 3067"

This reverts commit 041b4b6.

Fixes #3143. The original issue (#3067) was people trying to use
tts.tts_with_vc_to_file() with XTTS and was "fixed" in #3109. But XTTS has
integrated VC and you can just do tts.tts_to_file(..., speaker_wav="..."), there
is no point in passing it through FreeVC afterwards. So, reverting this commit
because it breaks tts.tts_with_vc_to_file() for any model that doesn't have
integrated VC, i.e. all models this method is meant for.

* fix: support multi-speaker models in tts_with_vc/tts_with_vc_to_file

* fix: only compute spk embeddings for models that support it

Fixes #1440. Passing a `speaker_wav` argument to regular Vits models failed
because they don't support voice cloning. Now that argument is simply ignored.

TTS/api.py

@@ -440,7 +440,7 @@ def voice_conversion_to_file(
         save_wav(wav=wav, path=file_path, sample_rate=self.voice_converter.vc_config.audio.output_sample_rate)
         return file_path
 
-    def tts_with_vc(self, text: str, language: str = None, speaker_wav: str = None):
+    def tts_with_vc(self, text: str, language: str = None, speaker_wav: str = None, speaker: str = None):
         """Convert text to speech with voice conversion.
 
         It combines tts with voice conversion to fake voice cloning.
@@ -457,17 +457,25 @@ def tts_with_vc(self, text: str, language: str = None, speaker_wav: str = None):
             speaker_wav (str, optional):
                 Path to a reference wav file to use for voice cloning with supporting models like YourTTS.
                 Defaults to None.
+            speaker (str, optional):
+                Speaker name for multi-speaker. You can check whether loaded model is multi-speaker by
+                `tts.is_multi_speaker` and list speakers by `tts.speakers`. Defaults to None.
         """
         with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
             # Lazy code... save it to a temp file to resample it while reading it for VC
-            self.tts_to_file(text=text, speaker=None, language=language, file_path=fp.name, speaker_wav=speaker_wav)
+            self.tts_to_file(text=text, speaker=speaker, language=language, file_path=fp.name)
         if self.voice_converter is None:
             self.load_vc_model_by_name("voice_conversion_models/multilingual/vctk/freevc24")
         wav = self.voice_converter.voice_conversion(source_wav=fp.name, target_wav=speaker_wav)
         return wav
 
     def tts_with_vc_to_file(
-        self, text: str, language: str = None, speaker_wav: str = None, file_path: str = "output.wav"
+        self,
+        text: str,
+        language: str = None,
+        speaker_wav: str = None,
+        file_path: str = "output.wav",
+        speaker: str = None,
     ):
         """Convert text to speech with voice conversion and save to file.
 
@@ -484,6 +492,9 @@ def tts_with_vc_to_file(
                 Defaults to None.
             file_path (str, optional):
                 Output file path. Defaults to "output.wav".
+            speaker (str, optional):
+                Speaker name for multi-speaker. You can check whether loaded model is multi-speaker by
+                `tts.is_multi_speaker` and list speakers by `tts.speakers`. Defaults to None.
         """
-        wav = self.tts_with_vc(text=text, language=language, speaker_wav=speaker_wav)
+        wav = self.tts_with_vc(text=text, language=language, speaker_wav=speaker_wav, speaker=speaker)
         save_wav(wav=wav, path=file_path, sample_rate=self.voice_converter.vc_config.audio.output_sample_rate)

TTS/utils/synthesizer.py

@@ -358,7 +358,11 @@ def tts(
                 )
 
         # compute a new d_vector from the given clip.
-        if speaker_wav is not None and self.tts_model.speaker_manager is not None:
+        if (
+            speaker_wav is not None
+            and self.tts_model.speaker_manager is not None
+            and self.tts_model.speaker_manager.encoder_ap is not None
+        ):
             speaker_embedding = self.tts_model.speaker_manager.compute_embedding_from_clip(speaker_wav)
 
         vocoder_device = "cpu"