Datasets:
Tasks:
Automatic Speech Recognition
Formats:
webdataset
Languages:
Icelandic
Size:
100K - 1M
License:
# Directory where the converted .wav files will be stored. | |
# The prompt mentioned storing them "directly under the upper test folder". | |
# Assuming the script is run from /home/dem/projects/k2/isl_zipformer/asr_data/samromur_asr/ | |
# and the 'upper test folder' is 'test/', this script will place files in 'test/wav_files/'. | |
# If the script is run from /home/dem/projects/k2, it will be in isl_zipformer/asr_data/samromur_asr/test/wav_files | |
# Let's assume the script is run from where the `test` directory is. | |
the_set=train | |
OUTPUT_DIR="${the_set}/wav_files" | |
mkdir -p "$OUTPUT_DIR" | |
# Number of parallel jobs. `nproc` gets the number of CPU cores. | |
# You can change this to a fixed number, e.g., PARALLEL_JOBS=4 | |
PARALLEL_JOBS=8 | |
echo "Converting files in parallel using up to $PARALLEL_JOBS jobs..." | |
# Find all .flac files and pipe them to xargs for parallel processing. | |
# The `-P` option for xargs sets the maximum number of parallel processes. | |
# The `-I {}` option tells xargs to replace `{}` with the input line (the file path). | |
find ${the_set}/train_part_01 -type f -name "*.flac" | xargs -P "$PARALLEL_JOBS" -I {} bash -c ' | |
flac_file="{}" | |
# Get the base filename without the .flac extension | |
base_filename=$(basename "$flac_file" .flac) | |
output_wav_file="'"$OUTPUT_DIR"'/${base_filename}.wav" | |
# Run the ffmpeg conversion. | |
# -y overwrites the output file if it exists. | |
ffmpeg -y -hide_banner -loglevel quiet -i "$flac_file" -ac 1 -ar 16000 "$output_wav_file" | |
' | |
echo "All conversions are complete. Check the '$OUTPUT_DIR' directory." | |