[FEA]: Add utilities to the API that allow for stripping audio tracks from common video formats

Is this a new feature, an improvement, or a change to existing functionality?

New Feature

How would you describe the priority of this feature request

Significant improvement

Please provide a clear description of problem this feature solves

As part of planning for video ingestion, there are a number of use cases we want to support, while minimizing the movement of data unnecessarily. In some cases we may only be interested in the audio component of a video file and would like to strip that data out as early as possible.

To support this, it would be good to have a set of utility functions in the API, and some initial prototyping to allow for converting files from their movie format into audio.

Describe the feature, and optionally a solution or implementation and any alternatives

These utilities should:

• Extract audio from a video file to a standard format (e.g., .wav, .mp3)
• Support mono, stereo, or multi-channel configurations
• Optionally downmix to mono
• Allow specifying output sampling rate and bit depth
• Avoid full decode if passthrough is possible (e.g., extracting AAC directly)

Sample function signature.

def extract_audio_from_video(
    video_path: str,
    output_path: str,
    audio_format: str = "wav",
    stream_index: int | None = None,
    downmix: bool = False,
    sample_rate: int | None = None,
    bit_depth: int | None = None,
) -> None:
    """
    Extracts audio from a video file.

    Parameters
    ----------
    video_path : str
        Path to the input video file.
    output_path : str
        Destination path for the extracted audio.
    audio_format : str, optional
        Desired output format (e.g., 'wav', 'mp3'), by default 'wav'.
    stream_index : int, optional
        Index of the audio stream to extract, if multiple streams are present.
    downmix : bool, optional
        If True, convert audio to mono.
    sample_rate : int, optional
        Resample output to this frequency (e.g., 16000, 44100).
    bit_depth : int, optional
        Set bit depth (e.g., 16, 24) if supported by format.
    """
    ...


def list_audio_streams(video_path: str) -> list[dict]:
    """
    Lists audio streams present in a video file.

    Parameters
    ----------
    video_path : str
        Path to the video file.

    Returns
    -------
    list of dict
        Metadata about each audio stream, e.g., codec, channels, sample rate.
    """
    ...

Example usage

from nv_ingest_api.util.media.audio_utils import extract_audio_from_video, list_audio_streams

# Inspect available audio streams
streams = list_audio_streams("my_video.mp4")
for i, stream in enumerate(streams):
    print(f"Stream {i}: {stream['codec_name']}, {stream['channels']}ch")

# Extract stereo WAV
extract_audio_from_video(
    video_path="my_video.mp4",
    output_path="out_stereo.wav"
)

# Extract and downmix to mono MP3
extract_audio_from_video(
    video_path="my_video.mp4",
    output_path="out_mono.mp3",
    audio_format="mp3",
    downmix=True
)

# Extract second stream with resampling and custom bit depth
extract_audio_from_video(
    video_path="my_multilang_video.mkv",
    output_path="out_highres.wav",
    stream_index=1,
    sample_rate=48000,
    bit_depth=24
)

Additional context

No response