From d2c2ba6fe557a23a8b3652b3e70e0f331cfa58a4 Mon Sep 17 00:00:00 2001
From: zzz27578 <2950506809@qq.com>
Date: Mon, 20 Jul 2026 16:55:49 +0800
Subject: [PATCH 1/2] fix: convert mislabeled audio before STT

---
 astrbot/core/utils/media_utils.py |  6 +++++-
 tests/test_media_utils.py         | 28 ++++++++++++++++++++++++++++
 2 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/astrbot/core/utils/media_utils.py b/astrbot/core/utils/media_utils.py
index a3990775ac..a15d481799 100644
--- a/astrbot/core/utils/media_utils.py
+++ b/astrbot/core/utils/media_utils.py
@@ -1127,7 +1127,11 @@ async def convert_audio_format(
     Raises:
         Exception: Raised when ffmpeg is unavailable or conversion fails.
     """
-    if audio_path.lower().endswith(f".{output_format}"):
+    detected_format = _get_audio_magic_type(audio_path)
+    if audio_path.lower().endswith(f".{output_format}") and (
+        detected_format == output_format
+        or (output_format == "ogg" and detected_format == "opus")
+    ):
         return audio_path
 
     if output_path is None:
diff --git a/tests/test_media_utils.py b/tests/test_media_utils.py
index 0cc3c85a12..78cdc01b54 100644
--- a/tests/test_media_utils.py
+++ b/tests/test_media_utils.py
@@ -399,6 +399,34 @@ async def test_media_resolver_accepts_unpadded_base64_payloads(tmp_path, monkeyp
     assert not list(tmp_path.iterdir())
 
 
+@pytest.mark.asyncio
+async def test_ensure_wav_converts_non_wav_bytes_with_wav_suffix(tmp_path, monkeypatch):
+    monkeypatch.setattr(media_utils, "get_astrbot_temp_path", lambda: str(tmp_path))
+    source_path = tmp_path / "voice.wav"
+    source_path.write_bytes(b"#!AMR\n" + b"\x00" * 64)
+
+    class _Process:
+        returncode = 0
+
+        async def communicate(self):
+            return b"", b""
+
+    async def fake_create_subprocess_exec(*args, **_kwargs):
+        Path(args[-1]).write_bytes(b"RIFF\x24\x00\x00\x00WAVEfmt " + b"\x00" * 16)
+        return _Process()
+
+    monkeypatch.setattr(
+        media_utils.asyncio,
+        "create_subprocess_exec",
+        fake_create_subprocess_exec,
+    )
+
+    resolved_path = await media_utils.ensure_wav(str(source_path))
+
+    assert resolved_path != str(source_path)
+    assert Path(resolved_path).read_bytes().startswith(b"RIFF")
+
+
 @pytest.mark.asyncio
 async def test_media_resolver_cleans_materialized_file_when_audio_conversion_fails(
     tmp_path, monkeypatch

From 7dd356b4114e1534a60d78a14a5e4036b798b910 Mon Sep 17 00:00:00 2001
From: zzz27578 <2950506809@qq.com>
Date: Mon, 20 Jul 2026 17:13:37 +0800
Subject: [PATCH 2/2] fix: offload audio format probe

---
 astrbot/core/utils/media_utils.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/astrbot/core/utils/media_utils.py b/astrbot/core/utils/media_utils.py
index a15d481799..0d5775acc7 100644
--- a/astrbot/core/utils/media_utils.py
+++ b/astrbot/core/utils/media_utils.py
@@ -1127,7 +1127,7 @@ async def convert_audio_format(
     Raises:
         Exception: Raised when ffmpeg is unavailable or conversion fails.
     """
-    detected_format = _get_audio_magic_type(audio_path)
+    detected_format = await asyncio.to_thread(_get_audio_magic_type, audio_path)
     if audio_path.lower().endswith(f".{output_format}") and (
         detected_format == output_format
         or (output_format == "ogg" and detected_format == "opus")