Implementation:Datajuicer Data juicer AudioDurationFilter

Knowledge Sources	Datajuicer_Data_juicer
Domains	Data_Quality, Filtering
Last Updated	2026-02-14 16:00 GMT

Overview

Concrete tool for filtering data samples based on audio file duration provided by Data-Juicer.

Description

AudioDurationFilter is a filter operator that keeps data samples whose audio durations are within a specified range in seconds. It loads audio files using librosa and computes their durations via librosa.get_duration(), caching results under the audio_duration stats key. The operator supports 'any' (keep if any audio qualifies) or 'all' (keep only if every audio qualifies) strategies. It extends the Filter base class and implements the two-phase compute_stats/process pattern.

Usage

Import this operator when you need to filter dataset samples based on the duration of audio files. Configure it in your Data-Juicer YAML config or instantiate directly.

Code Reference

Source Location

• Repository: Datajuicer_Data_juicer
• File: data_juicer/ops/filter/audio_duration_filter.py
• Lines: 1-90

Signature

@OPERATORS.register_module("audio_duration_filter")
@LOADED_AUDIOS.register_module("audio_duration_filter")
class AudioDurationFilter(Filter):
    def __init__(
        self, min_duration: int = 0, max_duration: int = sys.maxsize, any_or_all: str = "any", *args, **kwargs
    ):
        ...

Import

from data_juicer.ops.filter.audio_duration_filter import AudioDurationFilter

I/O Contract

Inputs

Name	Type	Required	Description
min_duration	int	No	The minimum audio duration to keep samples in seconds. Default: 0
max_duration	int	No	The maximum audio duration to keep samples in seconds. Default: sys.maxsize
any_or_all	str	No	Keep strategy: 'any' keeps if any audio meets condition, 'all' keeps only if all audios meet condition. Default: "any"

Outputs

Name	Type	Description
samples	Dict	Filtered samples with stats field updated (audio_duration)

Usage Examples

YAML Configuration

process:
  - audio_duration_filter:
      min_duration: 0
      max_duration: 300
      any_or_all: "any"

Python API

from data_juicer.ops.filter.audio_duration_filter import AudioDurationFilter

op = AudioDurationFilter(min_duration=0, max_duration=300)
# Apply to dataset
result = dataset.process(op)

Related Pages

• Environment:Datajuicer_Data_juicer_Python_Runtime_Environment