Audio Format Support

1
from nemo_curator.stages.audio.common import GetAudioDurationStage
2
3
# Calculate duration with built-in error handling
4
duration_stage = GetAudioDurationStage(
5
    audio_filepath_key="audio_filepath",
6
    duration_key="duration"
7
)

1
from nemo_curator.stages.audio.common import PreserveByValueStage
2
3
# Filter out corrupted files (duration = -1.0)
4
valid_files_filter = PreserveByValueStage(
5
    input_value_key="duration",
6
    target_value=0.0,
7
    operator="gt"  # greater than 0
8
)

1
from nemo_curator.pipeline import Pipeline
2
from nemo_curator.stages.audio.common import GetAudioDurationStage, PreserveByValueStage
3
from nemo_curator.stages.audio.inference.asr_nemo import InferenceAsrNemoStage
4
5
# Create pipeline with built-in error handling
6
pipeline = Pipeline(name="audio_validation")
7
8
# 1. Calculate duration (automatically handles format validation)
9
pipeline.add_stage(GetAudioDurationStage(
10
    audio_filepath_key="audio_filepath",
11
    duration_key="duration"
12
))
13
14
# 2. Filter out corrupted files (duration = -1.0 indicates issues)
15
pipeline.add_stage(PreserveByValueStage(
16
    input_value_key="duration",
17
    target_value=0.0,
18
    operator="gt"
19
))
20
21
# 3. Proceed with ASR inference on valid files only
22
pipeline.add_stage(InferenceAsrNemoStage(
23
    model_name="nvidia/stt_en_fastconformer_hybrid_large_pc"
24
))

1
import soundfile as sf
2
3
# Check available formats
4
print("Supported formats:")
5
for format_name, format_info in sf.available_formats().items():
6
    print(f"  {format_name}: {format_info}")
7
8
# Check specific file
9
try:
10
    info = sf.info("your_audio_file.wav")
11
    print(f"File info: {info}")
12
except Exception as e:
13
    print(f"File validation failed: {e}")