Add preferred_bitrate option to TTS audio conversion (#179760)

homeassistant/components/tts/__init__.py

@@ -71,6 +71,7 @@ from .models import Voice
 
 __all__ = [
     "ATTR_AUDIO_OUTPUT",
+    "ATTR_PREFERRED_BITRATE",
     "ATTR_PREFERRED_FORMAT",
     "ATTR_PREFERRED_SAMPLE_BYTES",
     "ATTR_PREFERRED_SAMPLE_CHANNELS",
@@ -99,6 +100,7 @@ ATTR_PREFERRED_FORMAT = "preferred_format"
 ATTR_PREFERRED_SAMPLE_RATE = "preferred_sample_rate"
 ATTR_PREFERRED_SAMPLE_CHANNELS = "preferred_sample_channels"
 ATTR_PREFERRED_SAMPLE_BYTES = "preferred_sample_bytes"
+ATTR_PREFERRED_BITRATE = "preferred_bitrate"
 ATTR_MEDIA_PLAYER_ENTITY_ID = "media_player_entity_id"
 ATTR_VOICE = "voice"
 
@@ -108,6 +110,7 @@ _PREFFERED_FORMAT_OPTIONS: Final[set[str]] = {
     ATTR_PREFERRED_SAMPLE_RATE,
     ATTR_PREFERRED_SAMPLE_CHANNELS,
     ATTR_PREFERRED_SAMPLE_BYTES,
+    ATTR_PREFERRED_BITRATE,
 }
 
 CONF_LANG = "language"
@@ -317,6 +320,7 @@ async def _async_convert_audio(
     to_sample_rate: int | None = None,
     to_sample_channels: int | None = None,
     to_sample_bytes: int | None = None,
+    to_bitrate: int | None = None,
 ) -> AsyncGenerator[bytes]:
     """Convert audio to a preferred format using ffmpeg."""
     ffmpeg_manager = ffmpeg.get_ffmpeg_manager(hass)
@@ -345,8 +349,13 @@ async def _async_convert_audio(
     if to_sample_channels is not None:
         command.extend(["-ac", str(to_sample_channels)])
     if to_extension == "mp3":
-        # Max quality for MP3.
-        command.extend(["-q:a", "0"])
+        if to_bitrate is not None:
+            # Constant bitrate. Some hardware decoders cannot handle the
+            # variable bitrate that -q:a produces.
+            command.extend(["-b:a", f"{to_bitrate}k"])
+        else:
+            # Max quality for MP3.
+            command.extend(["-q:a", "0"])
     if to_sample_bytes == 2:
         # 16-bit samples.
         command.extend(["-sample_fmt", "s16"])
@@ -588,6 +597,7 @@ class ResultStream:
                 ATTR_PREFERRED_SAMPLE_RATE,
                 ATTR_PREFERRED_SAMPLE_CHANNELS,
                 ATTR_PREFERRED_SAMPLE_BYTES,
+                ATTR_PREFERRED_BITRATE,
             )
         )
 
@@ -633,6 +643,7 @@ class ResultStream:
             to_sample_rate=self.options.get(ATTR_PREFERRED_SAMPLE_RATE),
             to_sample_channels=self.options.get(ATTR_PREFERRED_SAMPLE_CHANNELS),
             to_sample_bytes=self.options.get(ATTR_PREFERRED_SAMPLE_BYTES),
+            to_bitrate=self.options.get(ATTR_PREFERRED_BITRATE),
         )
         async for chunk in converted_audio:
             yield chunk
@@ -1082,6 +1093,14 @@ class SpeechManager:
         if sample_bytes is not None:
             sample_bytes = int(sample_bytes)
 
+        if ATTR_PREFERRED_BITRATE in supported_options:
+            bitrate = options.get(ATTR_PREFERRED_BITRATE)
+        else:
+            bitrate = options.pop(ATTR_PREFERRED_BITRATE, None)
+
+        if bitrate is not None:
+            bitrate = int(bitrate)
+
         if engine_instance.name is None or engine_instance.name is UNDEFINED:
             raise HomeAssistantError("TTS engine name is not set.")
 
@@ -1134,6 +1153,7 @@ class SpeechManager:
             or (sample_rate is not None)
             or (sample_channels is not None)
             or (sample_bytes is not None)
+            or (bitrate is not None)
         )
 
         if needs_conversion:
@@ -1145,6 +1165,7 @@ class SpeechManager:
                 to_sample_rate=sample_rate,
                 to_sample_channels=sample_channels,
                 to_sample_bytes=sample_bytes,
+                to_bitrate=bitrate,
             )
 
         async for chunk in data_gen:

tests/components/tts/test_init.py

@@ -1930,6 +1930,54 @@ async def test_async_convert_audio_probe_size(
     ]
 
 
+@pytest.mark.parametrize(
+    ("to_bitrate", "expected_encoder_args"),
+    [
+        pytest.param(None, ["-q:a", "0"], id="default_vbr"),
+        pytest.param(48, ["-b:a", "48k"], id="cbr_48k"),
+    ],
+)
+async def test_async_convert_audio_mp3_bitrate(
+    hass: HomeAssistant,
+    to_bitrate: int | None,
+    expected_encoder_args: list[str],
+) -> None:
+    """Test that a preferred bitrate produces a constant bitrate MP3."""
+    assert await async_setup_component(hass, ffmpeg.DOMAIN, {})
+
+    mock_process = MagicMock()
+    mock_process.stdin.drain = AsyncMock()
+    mock_process.stdout.read = AsyncMock(return_value=b"")
+    mock_process.wait = AsyncMock(return_value=0)
+
+    with patch(
+        "asyncio.create_subprocess_exec", return_value=mock_process
+    ) as mock_create_subprocess_exec:
+        async for _chunk in tts._async_convert_audio(
+            hass,
+            "wav",
+            _audio_data_gen(),
+            "mp3",
+            to_sample_rate=24000,
+            to_sample_channels=1,
+            to_bitrate=to_bitrate,
+        ):
+            pass
+
+    command = list(mock_create_subprocess_exec.call_args.args)
+    input_index = command.index("-i")
+    assert command[input_index + 2 :] == [
+        "-f",
+        "mp3",
+        "-ar",
+        "24000",
+        "-ac",
+        "1",
+        *expected_encoder_args,
+        "pipe:1",
+    ]
+
+
 async def test_default_engine_prefer_entity(
     hass: HomeAssistant,
     mock_tts_entity: MockTTSEntity,