compare only on rank, as voices may be subclassed

qt/aqt/sound.py

@@ -14,6 +14,7 @@ from concurrent.futures import Future
 from typing import Any, Callable, Dict, List, Optional, Tuple
 
 import pyaudio
+from operator import itemgetter
 
 import anki
 import aqt
@@ -169,7 +170,7 @@ class AVPlayer:
             if rank is not None:
                 ranked.append((rank, p))
 
-        ranked.sort()
+        ranked.sort(key=itemgetter(0))
 
         if ranked:
             return ranked[-1][1]

qt/aqt/tts.py

@@ -9,7 +9,7 @@ or
 
 {{tts ja_JP voices=Kyoko,Otoya,Another_name:Field}}
 
-The first argument must be a language code.
+The first argument must be an underscored language code, eg en_US.
 
 If provided, voices is a comma-separated list of one or more voices that
 the user would prefer. Spaces must not be included. Underscores will be