Operator ID: daft.las.functions.audio.audio_concat_fast.AudioConcatFast

Description #

Fast audio concatenation processor (same-source audio)

• Uses concat demuxer to quickly concatenate same-source audio (no re-encoding required)
• Applicable to audio files with identical format, codec, sample rate, and channel count
• Concatenate segments from the same recording
• Concatenate audio clips with the same format
• For scenarios requiring fast concatenation without format conversion

Key features #

• High speed, no quality loss
• Supports multiple input formats: local file paths, TOS/S3 storage paths, HTTP/HTTPS paths
• Supports output to a specified path

Format support and requirements #

• Supports common audio formats: mp3, wav, flac, aac, m4a, and more
• Requirement: All input audio must have the same format, codec, sample rate, and channel count

Difference from AudioConcat #

• AudioConcat: Uses concat filter, supports audio with different formats, requires re-encoding
• AudioConcatFast: Uses concat demuxer, supports only same-source audio, no re-encoding required, faster speed

Operator parameters #

Input #

Output #

An array containing the paths of the concatenated audio files; returns the output path on success, returns None on failure

Parameters #

If a parameter does not have a default value, it is required

Examples #

The following code demonstrates how to use Daft (for distributed environments) to run the operator for fast concatenation of same-source audio files (no re-encoding required).

from __future__ import annotations

import os

import daft
from daft import col
from daft.las.functions.audio import AudioConcatFast
from daft.las.functions.udf import las_udf

if __name__ == "__main__":
    # The converted audio will be saved to the specified TOS path. Therefore, environment variables must be set to ensure write permissions to TOS, including: ACCESS_KEY, SECRET_KEY, TOS_ENDPOINT, TOS_REGION, TOS_TEST_DIR
    TOS_TEST_DIR_URL = os.getenv("TOS_TEST_DIR_URL", "las-cn-beijing-public-online.tos-cn-beijing.volces.com")
    TOS_TEST_DIR = os.getenv("TOS_TEST_DIR", "tos_bucket")

    if os.getenv("DAFT_RUNNER", "native") == "ray":
        import logging

        import ray

        def configure_logging():
            logging.basicConfig(
                level=logging.INFO,
                format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
                datefmt="%Y-%m-%d %H:%M:%S",
            )
            logging.getLogger("tracing.span").setLevel(logging.WARNING)
            logging.getLogger("daft_io.stats").setLevel(logging.WARNING)
            logging.getLogger("DaftStatisticsManager").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaScheduler").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaDispatcher").setLevel(logging.WARNING)

        ray.init(dashboard_host="0.0.0.0", runtime_env={"worker_process_setup_hook": configure_logging})
        daft.set_runner_ray()

    daft.set_execution_config(actor_udf_ready_timeout=600)
    daft.set_execution_config(min_cpu_per_task=0)

    samples = {
        "audio_paths": [
            [
                f"https://{TOS_TEST_DIR_URL}/public/archive/audio_concat_fast/sample_a.wav",
                f"https://{TOS_TEST_DIR_URL}/public/archive/audio_concat_fast/sample_b.wav",
            ],
        ],
        "output_path": [f"tos://{TOS_TEST_DIR}/audio_concat_fast/output/concatenated_audio_fast.wav"],
    }
    ds = daft.from_pydict(samples)

    # Using Daft to concatenate audio files (fast, no re-encoding)
    constructor_kwargs = {
        "output_format": "wav",  # Same as input format
    }

    ds = ds.with_column(
        "output_path",
        las_udf(
            AudioConcatFast,
            construct_args=constructor_kwargs,
            num_cpus=1,
            concurrency=1,
            batch_size=1,
        )(col("audio_paths"), col("output_path")),
    )

    ds.show()
    # ╭───────────────────┬───────────────────╮
    # │ audio_paths                    ┆ output_path                    │
    # │ ---                            ┆ ---                            │
    # │ List[String]                   ┆ String                         │
    # ╞═══════════════════╪═══════════════════╡
    # │ [https://las-public-data-qa.t… ┆ tos://tos_bucket/audio_concat… │
    # ╰───────────────────┴───────────────────╯