Description #

Audio quality scoring module – evaluates audio quality using the DNSMOS model

Key features #

• Evaluates audio quality using the DNSMOS (Deep Noise Suppression Mean Opinion Score) model
• Provides scores for three dimensions: overall quality (OVRL), signal quality (SIG), and background noise (BAK)
• The scoring range is from 1 to 5, with higher scores indicating better audio quality
• Supports automatic sample rate conversion and audio preprocessing
• Suitable for scenarios such as speech quality evaluation, audio filtering, and quality control

Explanation of scoring dimensions #

• OVRL (Overall): Overall audio quality score, comprehensively evaluates audibility
• SIG (Signal): Signal quality score, evaluates the clarity and naturalness of the speech signal
• BAK (Background): Background noise score, evaluates the degree of interference from background noise

Operator parameters #

Input #

Output #

Structured result array, where each element contains the following fields:

• ovrl (float): Overall audio quality score, comprehensively evaluates audibility, range from 1.0 to 5.0
• sig (float): Signal quality score, evaluates the clarity and naturalness of the speech signal, range from 1.0 to 5.0
• bak (float): Background noise score, evaluates the degree of interference from background noise, range from 1.0 to 5.0

Audio that fails processing returns a structure containing null values

Parameters #

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

Examples #

The following code demonstrates how to use Daft to run the operator for audio quality scoring.

from __future__ import annotations

import os

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

if __name__ == "__main__":
    TOS_TEST_DIR_URL = os.getenv("TOS_TEST_DIR_URL", "las-cn-beijing-public-online.tos-cn-beijing.volces.com")
    model_path = os.getenv("MODEL_PATH", "/opt/las/models")

    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_path": [f"https://{TOS_TEST_DIR_URL}/public/archive/audio_quality_score/test_audio.wav"]}
    df = daft.from_pydict(samples)
    df = df.with_column(
        "audio_quality_score",
        las_udf(
            AudioQualityScore,
            construct_args={"model_path": model_path, "device": "cuda"},
            num_gpus=1,
            batch_size=8,
            concurrency=1,
        )(col("audio_path")),
    )

    df.show()
    # ╭────────────────────────────────┬───────────────────────────────────────────────────╮
    # │ audio_path                     ┆ audio_quality_score                               │
    # │ ---                            ┆ ---                                               │
    # │ String                         ┆ Struct[ovrl: Float64, sig: Float64, bak: Float64] │
    # ╞════════════════════════════════╪═══════════════════════════════════════════════════╡
    # │ https://las-public-data-qa.to… ┆ {ovrl: 1.7469981067293439,                        │
    # │                                ┆ si…                                               │
    # ╰────────────────────────────────┴───────────────────────────────────────────────────╯