diff --git a/frigate/data_processing/real_time/audio_transcription.py b/frigate/data_processing/real_time/audio_transcription.py
index 9dc2783c3..c0c8b86ae 100644
--- a/frigate/data_processing/real_time/audio_transcription.py
+++ b/frigate/data_processing/real_time/audio_transcription.py
@@ -44,7 +44,7 @@ class AudioTranscriptionRealTimeProcessor(RealTimeProcessorApi):
 
         if self.config.audio_transcription.model_size == "large":
             self.asr = FasterWhisperASR(
-                modelsize="tiny",  # could use 'base' for CPU, switch to 'small' or 'large-v2' for GPU
+                modelsize="tiny",
                 device="cuda"
                 if self.config.audio_transcription.device == "GPU"
                 else "cpu",
diff --git a/frigate/embeddings/maintainer.py b/frigate/embeddings/maintainer.py
index 44ad41eef..a07797781 100644
--- a/frigate/embeddings/maintainer.py
+++ b/frigate/embeddings/maintainer.py
@@ -179,12 +179,10 @@ class EmbeddingMaintainer(threading.Thread):
                 )
             )
 
-        audio_transcription_cameras = [
-            c
+        if any(
+            c.enabled_in_config and c.audio_transcription.enabled
             for c in self.config.cameras.values()
-            if c.enabled_in_config and c.audio_transcription.enabled
-        ]
-        if audio_transcription_cameras:
+        ):
             self.post_processors.append(
                 AudioTranscriptionPostProcessor(self.config, self.requestor, metrics)
             )
diff --git a/frigate/events/audio.py b/frigate/events/audio.py
index a19746f3d..11e8a2ae8 100644
--- a/frigate/events/audio.py
+++ b/frigate/events/audio.py
@@ -226,7 +226,6 @@ class AudioEventMaintainer(threading.Thread):
 
         # run audio transcription
         if self.transcription_processor is not None and (
-            # rms >= self.camera_config.audio.min_volume or self.is_endpoint is False
             self.camera_config.audio_transcription.live_enabled
         ):
             self.transcribing = True