Datasets:
| one liners to create | |
| metadata: | |
| awk -F'\t' 'BEGIN{OFS=","} NR==1{print "id","gender","age","ethnicity","first_language","socioeconomic_bkgd","transcription"} NR>1{print $1,$4,$3,$7,$5,$6,$2}' ./data/voices-asrfairness/metadata_all.tsv > data/voices-asrfairness/metadata_all.csv | |
| head -n 10001 data/voices-asrfairness/metadata_all.csv > data/voices-asrfairness/metadata.csv | |
| to copy audio files: | |
| tail -n +2 /data/jsalt/data/voices-asrfairness/metadata.csv | cut -d',' -f1 | while read id; do cp /data/jsalt/data/voices-asrfairness/audio_all/"${id}.wav" /data/jsalt/data/voices-asrfairness/audio/ 2>/dev/null; done && echo "Done: $(ls /data/jsalt/data/voices-asrfairness/audio/ | wc -l) files copied" | |