diff --git a/src/lerobot/datasets/audio_utils.py b/src/lerobot/datasets/audio_utils.py
index 95bee3f02..1156fa33f 100644
--- a/src/lerobot/datasets/audio_utils.py
+++ b/src/lerobot/datasets/audio_utils.py
@@ -41,7 +41,7 @@ def decode_audio(
     audio_path: Path | str,
     timestamps: list[float],
     duration: float,
-    backend: str | None = "torchaudio",
+    backend: str | None = "torchcodec",
 ) -> torch.Tensor:
     """
     Decodes audio using the specified backend.
@@ -49,7 +49,7 @@ def decode_audio(
         audio_path (Path): Path to the audio file.
         timestamps (list[float]): List of (starting) timestamps to extract audio chunks.
         duration (float): Duration of the audio chunks in seconds.
-        backend (str, optional): Backend to use for decoding. Defaults to "torchaudio".
+        backend (str, optional): Backend to use for decoding. Defaults to "torchcodec".
 
     Returns:
         torch.Tensor: Decoded audio chunks.
@@ -57,8 +57,7 @@ def decode_audio(
     Currently supports torchaudio.
     """
     if backend == "torchcodec":
-        #   return decode_audio_torchcodec(audio_path, timestamps, duration)    #TODO(CarolinePascal): uncomment this line at next torchcodec release
-        raise ValueError("torchcodec backend is not available yet.")
+        return decode_audio_torchcodec(audio_path, timestamps, duration)
     elif backend == "torchaudio":
         return decode_audio_torchaudio(audio_path, timestamps, duration)
     else:
diff --git a/src/lerobot/datasets/lerobot_dataset.py b/src/lerobot/datasets/lerobot_dataset.py
index 05e6ec5cf..c26bb3c4c 100644
--- a/src/lerobot/datasets/lerobot_dataset.py
+++ b/src/lerobot/datasets/lerobot_dataset.py
@@ -757,7 +757,7 @@ class LeRobotDataset(torch.utils.data.Dataset):
             download_audio (bool, optional): Flag to download the audio. Defaults to True.
             video_backend (str | None, optional): Video backend to use for decoding videos. Defaults to torchcodec when available int the platform; otherwise, defaults to 'pyav'.
                 You can also use the 'pyav' decoder used by Torchvision, which used to be the default option, or 'video_reader' which is another decoder of Torchvision.
-            audio_backend (str | None, optional): Audio backend to use for decoding audio. Defaults to 'torchaudio'.
+            audio_backend (str | None, optional): Audio backend to use for decoding audio. Defaults to 'torchcodec'.
             batch_encoding_size (int, optional): Number of episodes to accumulate before batch encoding videos.
                 Set to 1 for immediate encoding (default), or higher for batched encoding. Defaults to 1.
             vcodec (str, optional): Video codec for encoding videos during recording. Options: 'h264', 'hevc',
@@ -776,7 +776,7 @@ class LeRobotDataset(torch.utils.data.Dataset):
         self.revision = revision if revision else CODEBASE_VERSION
         self.video_backend = video_backend if video_backend else get_safe_default_codec()
         self.audio_backend = (
-            audio_backend if audio_backend else "torchaudio"
+            audio_backend if audio_backend else "torchcodec"
         )  # Waiting for torchcodec release #TODO(CarolinePascal)
         self.delta_indices = None
         self.batch_encoding_size = batch_encoding_size
@@ -1950,7 +1950,7 @@ class LeRobotDataset(torch.utils.data.Dataset):
         obj._recorded_frames = 0
         obj._writer_closed_for_reading = False
         obj.audio_backend = (
-            audio_backend if audio_backend is not None else "torchaudio"
+            audio_backend if audio_backend is not None else "torchcodec"
         )  # Waiting for torchcodec release #TODO(CarolinePascal)
         return obj