feat(datasets): expose video codec option for dataset recording (#2771)

* expose codec options + add tests

* pre-commit run -a

src/lerobot/scripts/lerobot_record.py

@@ -27,6 +27,8 @@ lerobot-record \
     --dataset.num_episodes=2 \
     --dataset.single_task="Grab the cube" \
     --display_data=true
+    # <- Optional: specify video codec (h264, hevc, libsvtav1). Default is libsvtav1. \
+    # --dataset.vcodec=h264 \
     # <- Teleop optional if you want to teleoperate to record or in between episodes with a policy \
     # --teleop.type=so100_leader \
     # --teleop.port=/dev/tty.usbmodem58760431551 \
@@ -165,6 +167,9 @@ class DatasetRecordConfig:
     # Number of episodes to record before batch encoding videos
     # Set to 1 for immediate encoding (default behavior), or higher for batched encoding
     video_encoding_batch_size: int = 1
+    # Video codec for encoding videos. Options: 'h264', 'hevc', 'libsvtav1'.
+    # Use 'h264' for faster encoding on systems where AV1 encoding is CPU-heavy.
+    vcodec: str = "libsvtav1"
     # Rename map for the observation to override the image and state keys
     rename_map: dict[str, str] = field(default_factory=dict)
 
@@ -427,6 +432,7 @@ def record(cfg: RecordConfig) -> LeRobotDataset:
                 cfg.dataset.repo_id,
                 root=cfg.dataset.root,
                 batch_encoding_size=cfg.dataset.video_encoding_batch_size,
+                vcodec=cfg.dataset.vcodec,
             )
 
             if hasattr(robot, "cameras") and len(robot.cameras) > 0:
@@ -448,6 +454,7 @@ def record(cfg: RecordConfig) -> LeRobotDataset:
                 image_writer_processes=cfg.dataset.num_image_writer_processes,
                 image_writer_threads=cfg.dataset.num_image_writer_threads_per_camera * len(robot.cameras),
                 batch_encoding_size=cfg.dataset.video_encoding_batch_size,
+                vcodec=cfg.dataset.vcodec,
             )
 
         # Load pretrained policy