README.md

metadata

license: cc-by-nc-4.0
task_categories:
  - automatic-speech-recognition
  - text-to-speech
  - audio-classification
language:
  - af
  - am
  - ar
  - az
  - be
  - bn
  - ca
  - ckb
  - cnr
  - crs
  - cs
  - de
  - dv
  - el
  - en
  - eo
  - es
  - fa
  - fr
  - ga
  - grc
  - ha
  - he
  - hi
  - hu
  - hy
  - id
  - ig
  - iu
  - ja
  - ka
  - kk
  - km
  - ko
  - la
  - lb
  - lo
  - mfe
  - mi
  - ml
  - mn
  - mr
  - ms
  - my
  - ne
  - nl
  - nr
  - nso
  - om
  - pa
  - pl
  - pt
  - rm
  - ro
  - ru
  - rw
  - si
  - sm
  - sn
  - sq
  - ss
  - st
  - sv
  - sw
  - ta
  - th
  - ti
  - tl
  - tn
  - tr
  - ts
  - ug
  - uz
  - ve
  - vi
  - xh
  - yue
  - zh
  - zu
size_categories:
  - 10M<n<100M
pretty_name: WorldSpeech
tags:
  - speech
  - multilingual
  - low-resource
  - parliamentary
  - asr
  - tts
  - audio
configs:
  - config_name: af_za
    data_files:
      - split: train
        path: data/af_za/*train-*.parquet
      - split: test
        path: data/af_za/*test-*.parquet
  - config_name: am_et
    data_files:
      - split: train
        path: data/am_et/*train-*.parquet
      - split: test
        path: data/am_et/*test-*.parquet
  - config_name: ar_bh
    data_files:
      - split: train
        path: data/ar_bh/*train-*.parquet
      - split: test
        path: data/ar_bh/*test-*.parquet
  - config_name: ar_dz
    data_files:
      - split: train
        path: data/ar_dz/*train-*.parquet
      - split: test
        path: data/ar_dz/*test-*.parquet
  - config_name: ar_eg
    data_files:
      - split: train
        path: data/ar_eg/*train-*.parquet
      - split: test
        path: data/ar_eg/*test-*.parquet
  - config_name: ar_iq
    data_files:
      - split: train
        path: data/ar_iq/*train-*.parquet
      - split: test
        path: data/ar_iq/*test-*.parquet
  - config_name: ar_kw
    data_files:
      - split: train
        path: data/ar_kw/*train-*.parquet
      - split: test
        path: data/ar_kw/*test-*.parquet
  - config_name: ar_ma
    data_files:
      - split: train
        path: data/ar_ma/*train-*.parquet
      - split: test
        path: data/ar_ma/*test-*.parquet
  - config_name: ar_sa
    data_files:
      - split: train
        path: data/ar_sa/*train-*.parquet
      - split: test
        path: data/ar_sa/*test-*.parquet
  - config_name: ar_tn
    data_files:
      - split: train
        path: data/ar_tn/*train-*.parquet
      - split: test
        path: data/ar_tn/*test-*.parquet
  - config_name: ar_un
    data_files:
      - split: train
        path: data/ar_un/*train-*.parquet
      - split: test
        path: data/ar_un/*test-*.parquet
  - config_name: as_in
    data_files:
      - split: train
        path: data/as_in/in_as_air-train-*.parquet
      - split: test
        path: data/as_in/in_as_air-test-*.parquet
  - config_name: az_az
    data_files:
      - split: train
        path: data/az_az/*train-*.parquet
      - split: test
        path: data/az_az/*test-*.parquet
  - config_name: be_by
    data_files:
      - split: train
        path: data/be_by/*train-*.parquet
      - split: test
        path: data/be_by/*test-*.parquet
  - config_name: bn_bd
    data_files:
      - split: train
        path: data/bn_bd/*train-*.parquet
      - split: test
        path: data/bn_bd/*test-*.parquet
  - config_name: bn_in
    data_files:
      - split: train
        path: data/bn_in/in_bn_air-train-*.parquet
      - split: test
        path: data/bn_in/in_bn_air-test-*.parquet
  - config_name: ca_es
    data_files:
      - split: train
        path: data/ca_es/*train-*.parquet
      - split: test
        path: data/ca_es/*test-*.parquet
  - config_name: ca_fr
    default: true
    data_files:
      - split: train
        path: data/ca_fr/train-*.parquet
      - split: test
        path: data/ca_fr/test-*.parquet
  - config_name: ckb_iq
    data_files:
      - split: train
        path: data/ckb_iq/*train-*.parquet
      - split: test
        path: data/ckb_iq/*test-*.parquet
  - config_name: cn_ug
    data_files:
      - split: train
        path: data/cn_ug/*train-*.parquet
      - split: test
        path: data/cn_ug/*test-*.parquet
  - config_name: cnr_me
    data_files:
      - split: train
        path: data/cnr_me/*train-*.parquet
      - split: test
        path: data/cnr_me/*test-*.parquet
  - config_name: crs_sc
    data_files:
      - split: train
        path: data/crs_sc/*train-*.parquet
      - split: test
        path: data/crs_sc/*test-*.parquet
  - config_name: cs_cz
    data_files:
      - split: train
        path:
          - data/cs_cz/*train-*.parquet
          - data/cz_cs/*train-*.parquet
      - split: test
        path:
          - data/cs_cz/*test-*.parquet
          - data/cz_cs/*test-*.parquet
  - config_name: de_at
    data_files:
      - split: train
        path: data/de_at/*train-*.parquet
      - split: test
        path: data/de_at/*test-*.parquet
  - config_name: de_li
    data_files:
      - split: train
        path: data/de_li/train-*.parquet
      - split: test
        path: data/de_li/test-*.parquet
  - config_name: dgo_in
    data_files:
      - split: train
        path: data/dgo_in/in_dgo_air-train-*.parquet
      - split: test
        path: data/dgo_in/in_dgo_air-test-*.parquet
  - config_name: dv_mv
    data_files:
      - split: train
        path: data/dv_mv/*train-*.parquet
      - split: test
        path: data/dv_mv/*test-*.parquet
  - config_name: el_cy
    data_files:
      - split: train
        path: data/el_cy/*train-*.parquet
      - split: test
        path: data/el_cy/*test-*.parquet
  - config_name: el_gr
    data_files:
      - split: train
        path: data/el_gr/*train-*.parquet
      - split: test
        path: data/el_gr/*test-*.parquet
  - config_name: en_au
    data_files:
      - split: train
        path: data/en_au/*train-*.parquet
      - split: test
        path: data/en_au/*test-*.parquet
  - config_name: en_jm
    data_files:
      - split: train
        path: data/en_jm/*train-*.parquet
      - split: test
        path: data/en_jm/*test-*.parquet
  - config_name: en_ke
    data_files:
      - split: train
        path: data/en_ke/*train-*.parquet
      - split: test
        path: data/en_ke/*test-*.parquet
  - config_name: en_nz
    data_files:
      - split: train
        path: data/en_nz/*train-*.parquet
      - split: test
        path: data/en_nz/*test-*.parquet
  - config_name: en_pk
    data_files:
      - split: train
        path: data/en_pk/train-*.parquet
      - split: test
        path: data/en_pk/test-*.parquet
  - config_name: en_sl
    data_files:
      - split: train
        path: data/en_sl/*train-*.parquet
      - split: test
        path: data/en_sl/*test-*.parquet
  - config_name: en_us
    data_files:
      - split: train
        path: data/en_us/*train-*.parquet
      - split: test
        path: data/en_us/*test-*.parquet
  - config_name: en_zm
    data_files:
      - split: train
        path: data/en_zm/*train-*.parquet
      - split: test
        path: data/en_zm/*test-*.parquet
  - config_name: eo
    data_files:
      - split: train
        path: data/eo/*train-*.parquet
      - split: test
        path: data/eo/*test-*.parquet
  - config_name: es_ar
    data_files:
      - split: train
        path: data/es_ar/*train-*.parquet
      - split: test
        path: data/es_ar/*test-*.parquet
  - config_name: es_cl
    data_files:
      - split: train
        path: data/es_cl/*train-*.parquet
      - split: test
        path: data/es_cl/*test-*.parquet
  - config_name: es_co
    data_files:
      - split: train
        path: data/es_co/*train-*.parquet
      - split: test
        path: data/es_co/*test-*.parquet
  - config_name: es_es
    data_files:
      - split: train
        path: data/es_es/*train-*.parquet
      - split: test
        path: data/es_es/*test-*.parquet
  - config_name: es_mx
    data_files:
      - split: train
        path: data/es_mx/*train-*.parquet
      - split: test
        path: data/es_mx/*test-*.parquet
  - config_name: es_pe
    data_files:
      - split: train
        path: data/es_pe/*train-*.parquet
      - split: test
        path: data/es_pe/*test-*.parquet
  - config_name: es_pr
    data_files:
      - split: train
        path: data/es_pr/*train-*.parquet
      - split: test
        path: data/es_pr/*test-*.parquet
  - config_name: es_py
    data_files:
      - split: train
        path: data/es_py/*train-*.parquet
      - split: test
        path: data/es_py/*test-*.parquet
  - config_name: es_uy
    data_files:
      - split: train
        path: data/es_uy/*train-*.parquet
      - split: test
        path: data/es_uy/*test-*.parquet
  - config_name: fa_ir
    data_files:
      - split: train
        path: data/fa_ir/*train-*.parquet
      - split: test
        path: data/fa_ir/*test-*.parquet
  - config_name: fr_ca
    data_files:
      - split: train
        path: data/fr_ca/*train-*.parquet
      - split: test
        path: data/fr_ca/*test-*.parquet
  - config_name: fr_cd
    data_files:
      - split: train
        path: data/fr_cd/*train-*.parquet
      - split: test
        path: data/fr_cd/*test-*.parquet
  - config_name: fr_ci
    data_files:
      - split: train
        path: data/fr_ci/*train-*.parquet
      - split: test
        path: data/fr_ci/*test-*.parquet
  - config_name: grc_gr
    data_files:
      - split: train
        path: data/grc_gr/*train-*.parquet
      - split: test
        path: data/grc_gr/*test-*.parquet
  - config_name: gu_in
    data_files:
      - split: train
        path: data/gu_in/in_gu_air-train-*.parquet
      - split: test
        path: data/gu_in/in_gu_air-test-*.parquet
  - config_name: ha_ng
    data_files:
      - split: train
        path: data/ha_ng/*train-*.parquet
      - split: test
        path: data/ha_ng/*test-*.parquet
  - config_name: ha_td
    data_files:
      - split: train
        path: data/ha_td/*train-*.parquet
      - split: test
        path: data/ha_td/*test-*.parquet
  - config_name: he_il
    data_files:
      - split: train
        path: data/he_il/*train-*.parquet
      - split: test
        path: data/he_il/*test-*.parquet
  - config_name: hi_in
    data_files:
      - split: train
        path:
          - data/hi_in/train-*.parquet
          - data/hi_in/in_hi_bh-train-*.parquet
          - data/hi_in/in_hi_mkb-train-*.parquet
          - data/hi_in/in_hi_rj-train-*.parquet
          - data/hi_in/in_hi_air-train-*.parquet
          - data/hi_in/in_hi_rs-train-*.parquet
      - split: test
        path:
          - data/hi_in/test-*.parquet
          - data/hi_in/in_hi_bh-test-*.parquet
          - data/hi_in/in_hi_mkb-test-*.parquet
          - data/hi_in/in_hi_rj-test-*.parquet
          - data/hi_in/in_hi_air-test-*.parquet
          - data/hi_in/in_hi_rs-test-*.parquet
  - config_name: hu_hu
    data_files:
      - split: train
        path: data/hu_hu/*train-*.parquet
      - split: test
        path: data/hu_hu/*test-*.parquet
  - config_name: hy_am
    data_files:
      - split: train
        path: data/hy_am/*train-*.parquet
      - split: test
        path: data/hy_am/*test-*.parquet
  - config_name: id_id
    data_files:
      - split: train
        path: data/id_id/*train-*.parquet
      - split: test
        path: data/id_id/*test-*.parquet
  - config_name: ig_ng
    data_files:
      - split: train
        path: data/ig_ng/*train-*.parquet
      - split: test
        path: data/ig_ng/*test-*.parquet
  - config_name: ja_jp
    data_files:
      - split: train
        path: data/ja_jp/*train-*.parquet
      - split: test
        path: data/ja_jp/*test-*.parquet
  - config_name: ka_ge
    data_files:
      - split: train
        path: data/ka_ge/*train-*.parquet
      - split: test
        path: data/ka_ge/*test-*.parquet
  - config_name: kh_km
    data_files:
      - split: train
        path: data/kh_km/*train-*.parquet
      - split: test
        path: data/kh_km/*test-*.parquet
  - config_name: kk_kz
    data_files:
      - split: train
        path: data/kk_kz/*train-*.parquet
      - split: test
        path: data/kk_kz/*test-*.parquet
  - config_name: kn_in
    data_files:
      - split: train
        path: data/kn_in/in_kn_air-train-*.parquet
      - split: test
        path: data/kn_in/in_kn_air-test-*.parquet
  - config_name: ko_kr
    data_files:
      - split: train
        path: data/ko_kr/*train-*.parquet
      - split: test
        path: data/ko_kr/*test-*.parquet
  - config_name: kok_in
    data_files:
      - split: train
        path: data/kok_in/in_kok_air-train-*.parquet
      - split: test
        path: data/kok_in/in_kok_air-test-*.parquet
  - config_name: la_lo
    data_files:
      - split: train
        path: data/la_lo/*train-*.parquet
      - split: test
        path: data/la_lo/*test-*.parquet
  - config_name: la_va
    data_files:
      - split: train
        path: data/la_va/*train-*.parquet
      - split: test
        path: data/la_va/*test-*.parquet
  - config_name: lb_lu
    data_files:
      - split: train
        path: data/lb_lu/*train-*.parquet
      - split: test
        path: data/lb_lu/*test-*.parquet
  - config_name: mai_in
    data_files:
      - split: train
        path: data/mai_in/in_mai_air-train-*.parquet
      - split: test
        path: data/mai_in/in_mai_air-test-*.parquet
  - config_name: mfe_mu
    data_files:
      - split: train
        path: data/mfe_mu/*train-*.parquet
      - split: test
        path: data/mfe_mu/*test-*.parquet
  - config_name: mi_nz
    data_files:
      - split: train
        path: data/mi_nz/*train-*.parquet
      - split: test
        path: data/mi_nz/*test-*.parquet
  - config_name: ml_in
    data_files:
      - split: train
        path:
          - data/ml_in/*train-*.parquet
          - data/ml_in/in_ml_air-train-*.parquet
      - split: test
        path:
          - data/ml_in/*test-*.parquet
          - data/ml_in/in_ml_air-test-*.parquet
  - config_name: mm_my
    data_files:
      - split: train
        path: data/mm_my/*train-*.parquet
      - split: test
        path: data/mm_my/*test-*.parquet
  - config_name: mn_mn
    data_files:
      - split: train
        path: data/mn_mn/*train-*.parquet
      - split: test
        path: data/mn_mn/*test-*.parquet
  - config_name: mr_in
    data_files:
      - split: train
        path:
          - data/mr_in/*train-*.parquet
          - data/mr_in/in_mr_air-train-*.parquet
      - split: test
        path:
          - data/mr_in/*test-*.parquet
          - data/mr_in/in_mr_air-test-*.parquet
  - config_name: ms_my
    data_files:
      - split: train
        path: data/ms_my/*train-*.parquet
      - split: test
        path: data/ms_my/*test-*.parquet
  - config_name: ne_in
    data_files:
      - split: train
        path: data/ne_in/in_ne_air-train-*.parquet
      - split: test
        path: data/ne_in/in_ne_air-test-*.parquet
  - config_name: ne_np
    data_files:
      - split: train
        path: data/ne_np/*train-*.parquet
      - split: test
        path: data/ne_np/*test-*.parquet
  - config_name: nl_be
    data_files:
      - split: train
        path: data/nl_be/*train-*.parquet
      - split: test
        path: data/nl_be/*test-*.parquet
  - config_name: nl_nl
    data_files:
      - split: train
        path: data/nl_nl/*train-*.parquet
      - split: test
        path: data/nl_nl/*test-*.parquet
  - config_name: nr_za
    data_files:
      - split: train
        path: data/nr_za/*train-*.parquet
      - split: test
        path: data/nr_za/*test-*.parquet
  - config_name: nso_za
    data_files:
      - split: train
        path: data/nso_za/*train-*.parquet
      - split: test
        path: data/nso_za/*test-*.parquet
  - config_name: om_et
    data_files:
      - split: train
        path: data/om_et/*train-*.parquet
      - split: test
        path: data/om_et/*test-*.parquet
  - config_name: or_in
    data_files:
      - split: train
        path: data/or_in/in_or_air-train-*.parquet
      - split: test
        path: data/or_in/in_or_air-test-*.parquet
  - config_name: pa_in
    data_files:
      - split: train
        path:
          - data/pa_in/*train-*.parquet
          - data/pa_in/in_pa_air-train-*.parquet
      - split: test
        path:
          - data/pa_in/*test-*.parquet
          - data/pa_in/in_pa_air-test-*.parquet
  - config_name: pl_pl
    data_files:
      - split: train
        path: data/pl_pl/*train-*.parquet
      - split: test
        path: data/pl_pl/*test-*.parquet
  - config_name: pt_br
    data_files:
      - split: train
        path: data/pt_br/*train-*.parquet
      - split: test
        path: data/pt_br/*test-*.parquet
  - config_name: rm_ch
    data_files:
      - split: train
        path: data/rm_ch/*train-*.parquet
      - split: test
        path: data/rm_ch/*test-*.parquet
  - config_name: ro_md
    data_files:
      - split: train
        path: data/ro_md/*train-*.parquet
      - split: test
        path: data/ro_md/*test-*.parquet
  - config_name: ro_ro
    data_files:
      - split: train
        path: data/ro_ro/*train-*.parquet
      - split: test
        path: data/ro_ro/*test-*.parquet
  - config_name: ru_by
    data_files:
      - split: train
        path: data/ru_by/*train-*.parquet
      - split: test
        path: data/ru_by/*test-*.parquet
  - config_name: ru_ru
    data_files:
      - split: train
        path: data/ru_ru/*train-*.parquet
      - split: test
        path: data/ru_ru/*test-*.parquet
  - config_name: rw_rw
    data_files:
      - split: train
        path: data/rw_rw/*train-*.parquet
      - split: test
        path: data/rw_rw/*test-*.parquet
  - config_name: si_lk
    data_files:
      - split: train
        path: data/si_lk/*train-*.parquet
      - split: test
        path: data/si_lk/*test-*.parquet
  - config_name: sm_ws
    data_files:
      - split: train
        path: data/sm_ws/*train-*.parquet
      - split: test
        path: data/sm_ws/*test-*.parquet
  - config_name: sn_zw
    data_files:
      - split: train
        path: data/sn_zw/*train-*.parquet
      - split: test
        path: data/sn_zw/*test-*.parquet
  - config_name: sq_al
    data_files:
      - split: train
        path: data/sq_al/*train-*.parquet
      - split: test
        path: data/sq_al/*test-*.parquet
  - config_name: sq_xk
    data_files:
      - split: train
        path: data/sq_xk/*train-*.parquet
      - split: test
        path: data/sq_xk/*test-*.parquet
  - config_name: ss_za
    data_files:
      - split: train
        path: data/ss_za/*train-*.parquet
      - split: test
        path: data/ss_za/*test-*.parquet
  - config_name: st_za
    data_files:
      - split: train
        path: data/st_za/*train-*.parquet
      - split: test
        path: data/st_za/*test-*.parquet
  - config_name: sv_ax
    data_files:
      - split: train
        path: data/sv_ax/*train-*.parquet
      - split: test
        path: data/sv_ax/*test-*.parquet
  - config_name: sw_ke
    data_files:
      - split: train
        path: data/sw_ke/*train-*.parquet
      - split: test
        path: data/sw_ke/*test-*.parquet
  - config_name: sw_tz
    data_files:
      - split: train
        path:
          - data/sw_tz/tz_sw_znz-train-*.parquet
          - data/sw_tz/tz_sw_bunge-train-*.parquet
      - split: test
        path:
          - data/sw_tz/tz_sw_znz-test-*.parquet
          - data/sw_tz/tz_sw_bunge-test-*.parquet
  - config_name: ta_in
    data_files:
      - split: train
        path: data/ta_in/in_ta_air-train-*.parquet
      - split: test
        path: data/ta_in/in_ta_air-test-*.parquet
  - config_name: ta_lk
    data_files:
      - split: train
        path: data/ta_lk/*train-*.parquet
      - split: test
        path: data/ta_lk/*test-*.parquet
  - config_name: te_in
    data_files:
      - split: train
        path: data/te_in/in_te_air-train-*.parquet
      - split: test
        path: data/te_in/in_te_air-test-*.parquet
  - config_name: th_th
    data_files:
      - split: train
        path: data/th_th/*train-*.parquet
      - split: test
        path: data/th_th/*test-*.parquet
  - config_name: ti_et
    data_files:
      - split: train
        path: data/ti_et/*train-*.parquet
      - split: test
        path: data/ti_et/*test-*.parquet
  - config_name: tl_ph
    data_files:
      - split: train
        path: data/tl_ph/*train-*.parquet
      - split: test
        path: data/tl_ph/*test-*.parquet
  - config_name: tn_bw
    data_files:
      - split: train
        path: data/tn_bw/*train-*.parquet
      - split: test
        path: data/tn_bw/*test-*.parquet
  - config_name: tn_za
    data_files:
      - split: train
        path: data/tn_za/*train-*.parquet
      - split: test
        path: data/tn_za/*test-*.parquet
  - config_name: tr_tr
    data_files:
      - split: train
        path: data/tr_tr/*train-*.parquet
      - split: test
        path: data/tr_tr/*test-*.parquet
  - config_name: ts_za
    data_files:
      - split: train
        path: data/ts_za/*train-*.parquet
      - split: test
        path: data/ts_za/*test-*.parquet
  - config_name: ur_in
    data_files:
      - split: train
        path: data/ur_in/in_ur_air-train-*.parquet
      - split: test
        path: data/ur_in/in_ur_air-test-*.parquet
  - config_name: ur_pk
    data_files:
      - split: train
        path: data/ur_pk/train-*.parquet
      - split: test
        path: data/ur_pk/test-*.parquet
  - config_name: uz_uz
    data_files:
      - split: train
        path: data/uz_uz/*train-*.parquet
      - split: test
        path: data/uz_uz/*test-*.parquet
  - config_name: ve_za
    data_files:
      - split: train
        path: data/ve_za/*train-*.parquet
  - config_name: vn_vi
    data_files:
      - split: train
        path: data/vn_vi/*train-*.parquet
      - split: test
        path: data/vn_vi/*test-*.parquet
  - config_name: xh_za
    data_files:
      - split: train
        path: data/xh_za/*train-*.parquet
      - split: test
        path: data/xh_za/*test-*.parquet
  - config_name: yue_hk
    data_files:
      - split: train
        path: data/yue_hk/*train-*.parquet
      - split: test
        path: data/yue_hk/*test-*.parquet
  - config_name: zh_tw
    data_files:
      - split: train
        path: data/zh_tw/*train-*.parquet
      - split: test
        path: data/zh_tw/*test-*.parquet
  - config_name: zu_za
    data_files:
      - split: train
        path: data/zu_za/*train-*.parquet
      - split: test
        path: data/zu_za/*test-*.parquet

WorldSpeech

A multilingual ASR dataset containing over 65k hours of human transcribed speech across 127 language-region variants, drawn from national parliaments, public broadcasters, public-domain audiobooks, and international institutions. Rows consist of 24 kHz speech utterances paired with a human-provided transcript, an aligned ASR transcript, character error rate (CER) between the two, a WADA-SNR estimate, and four DNSMOS-P.835 quality scores.

Dataset Overview

Metric	Value
Total hours	65,072
Distinct languages	88 (76 above 10 hours)
Language variations	127
Languages ≥ 1,000 h	24
Languages ≥ 500 h	28
Languages ≥ 200 h	37
Languages ≥ 50 h	53
Avg DNSMOS-P.835 OVR	2.83
Sample rate	24 kHz

How to use

from datasets import load_dataset

ds = load_dataset("disco-eth/WorldSpeech", "nl_nl", split="train")

row = ds[0]
wav    = row["audio"]["array"]         
sr     = row["audio"]["sampling_rate"]
text   = row["human_transcript"]
snr    = row["snr"]
dnsmos = row["dnsmos_ovr"]

Streaming

from datasets import load_dataset

ds = load_dataset("disco-eth/WorldSpeech", "nl_nl", split="train", streaming=True)

for row in ds:
    wav = row["audio"]["array"]
    break

Schema

Field	Description
audio	OGG Opus decoded to {"array": np.float32[N], "sampling_rate": 24000}
human_transcript	Human-provided transcript
asr_transcript	ASR output used for alignment and CER computation
cer	Character error rate between asr_transcript and human_transcript
snr	WADA-SNR estimate, dB
dnsmos_sig	DNSMOS-P.835 signal quality
dnsmos_bak	DNSMOS-P.835 background noise
dnsmos_ovr	DNSMOS-P.835 overall MOS
dnsmos_p808	DNSMOS-P.808 MOS
duration	Segment duration in seconds
source	Source identifier (e.g. parliament_nl, librivox, voa_hausa)
source_url	URL to the original recording
source_start_s	Start offset within the source recording (seconds)
source_end_s	End offset within the source recording (seconds)
session_date	ISO-8601 date of the original recording
segment_id	Unique identifier for this segment
language	BCP-47 language tag
country	ISO 3166-1 alpha-2 country code (un / va for international)

Language Coverage

Config	Country	Language	Hours	DNSMOS OVR
af_za	South Africa	Afrikaans	20.2	2.79
am_et	Ethiopia	Amharic	39.6	3.11
ar_bh	Bahrain	Arabic	272.5	2.72
ar_dz	Algeria	Arabic	92.9	2.66
ar_eg	Egypt	Arabic	22.0	2.86
ar_iq	Iraq	Arabic	291.9	2.94
ar_kw	Kuwait	Arabic	175.5	2.82
ar_ma	Morocco	Arabic	78.3	2.68
ar_sa	Saudi Arabia	Arabic	6.1	2.01
ar_tn	Tunisia	Arabic	50.9	2.74
ar_un	United Nations	Arabic	11.1	3.03
as_in	India	Assamese	54.5	3.06
az_az	Azerbaijan	Azerbaijani	305.4	2.76
be_by	Belarus	Belarusian	24.2	3.10
bn_bd	Bangladesh	Bengali	46.1	3.03
bn_in	India	Bengali	26.7	2.83
ca_es	Spain	Catalan	1,171.0	2.99
ckb_iq	Iraq	Sorani Kurdish	35.3	2.93
cn_ug	China	Uyghur	200.0	2.86
cnr_me	Montenegro	Montenegrin	47.9	2.98
crs_sc	Seychelles	Kreol Seselwa	1,602.3	3.15
cs_cz	Czech Republic	Czech	3,717.4	2.86
de_at	Austria	German	1,077.5	2.93
de_li	Liechtenstein	German	829.7	2.79
dgo_in	India	Dogri	34.5	3.17
dv_mv	Maldives	Dhivehi	20.0	2.92
el_cy	Cyprus	Greek	394.4	2.90
el_gr	Greece	Greek	35.9	3.37
en_au	Australia	English	568.4	3.02
en_jm	Jamaica	English	9.6	2.52
en_ke	Kenya	English	170.0	2.97
en_nz	New Zealand	English	435.7	2.68
en_pk	Pakistan	English	8.5	2.28
en_sl	Sierra Leone	English	102.4	2.70
en_us	USA	English	3,725.9	2.85
en_zm	Zambia	English	291.6	2.70
eo	(international)	Esperanto	15.0	3.27
es_ar	Argentina	Spanish	251.7	2.78
es_cl	Chile	Spanish	1,740.5	2.94
es_co	Colombia	Spanish	156.8	2.75
es_es	Spain	Spanish	2,097.4	2.52
es_mx	Mexico	Spanish	901.6	2.63
es_pe	Peru	Spanish	387.4	2.63
es_pr	Puerto Rico	Spanish	228.6	2.73
es_py	Paraguay	Spanish	133.2	3.14
es_uy	Uruguay	Spanish	894.4	2.94
fa_ir	Iran	Persian	27.6	2.94
ca_fr	Canada	French	5,989.3	2.88
fr_cd	DR Congo	French	28.0	3.00
fr_ci	Côte d'Ivoire	French	11.9	2.97
ga_ie	Ireland	Irish	60.6	3.00
grc_gr	Greece	Ancient Greek	1.3	3.25
gu_in	India	Gujarati	27.2	3.09
ha_ng	Nigeria	Hausa	54.9	2.88
ha_td	Chad	Hausa	71.5	2.87
he_il	Israel	Hebrew	41.8	3.21
hi_in	India	Hindi	1,706.7	2.68
hu_hu	Hungary	Hungarian	1,350.1	2.96
hy_am	Armenia	Armenian	1,138.9	2.59
id_id	Indonesia	Indonesian	340.0	2.95
ig_ng	Nigeria	Igbo	40.7	3.31
iu_ca	Canada	Inuktitut	33.8	2.81
ja_jp	Japan	Japanese	1,387.2	2.86
ka_ge	Georgia	Georgian	206.4	2.37
kh_km	Cambodia	Khmer	1,323.0	2.85
kk_kz	Kazakhstan	Kazakh	179.0	2.84
kn_in	India	Kannada	30.2	3.19
ko_kr	South Korea	Korean	1,454.1	3.07
kok_in	India	Konkani	0.7	3.19
la_lo	Laos	Lao	827.0	2.82
la_va	Vatican	Latin	34.8	3.37
lb_lu	Luxembourg	Luxembourgish	1,805.3	2.87
mai_in	India	Maithili	9.6	3.19
mfe_mu	Mauritius	Mauritian Creole	44.3	3.45
mi_nz	New Zealand	Māori	1.6	2.85
ml_in	India	Malayalam	56.7	3.10
mm_my	Myanmar	Burmese	865.0	2.88
mn_mn	Mongolia	Mongolian	181.0	2.80
mr_in	India	Marathi	114.3	3.17
ms_my	Malaysia	Malay	432.0	2.91
ne_in	India	Nepali	5.7	2.94
ne_np	Nepal	Nepali	58.0	2.79
nl_be	Belgium	Dutch (Flemish)	960.5	2.59
nl_nl	Netherlands	Dutch	4,497.6	3.00
nr_za	South Africa	Southern Ndebele	0.2	2.60
nso_za	South Africa	Northern Sotho	0.7	2.70
om_et	Ethiopia	Oromo	16.3	3.06
or_in	India	Odia	57.8	3.05
pa_in	India	Punjabi	4.0	2.61
pl_pl	Poland	Polish	2,731.9	2.70
pt_br	Brazil	Portuguese	1,763.5	2.68
rm_ch	Switzerland	Romansh	163.1	3.14
ro_md	Moldova	Romanian	1,059.5	2.66
ro_ro	Romania	Romanian	686.3	2.75
ru_by	Belarus	Russian	2.9	2.79
ru_ru	Russia	Russian	1,534.2	3.03
rw_rw	Rwanda	Kinyarwanda	17.6	2.63
rw_voa	Rwanda	Kinyarwanda (VOA)	14.5	2.77
si_lk	Sri Lanka	Sinhala	154.0	2.40
sm_ws	Samoa	Samoan	55.9	2.79
sn_zw	Zimbabwe	Shona	18.2	2.88
sq_al	Albania	Albanian	257.4	2.69
sq_xk	Kosovo	Albanian	176.6	2.69
ss_za	South Africa	Swati	0.1	2.84
st_za	South Africa	Sesotho	0.4	2.67
sv_ax	Åland Islands	Swedish	66.0	2.73
sw_ke	Kenya	Swahili	257.4	2.93
sw_tz	Tanzania	Swahili	748.8	2.33
ta_in	India	Tamil	36.1	3.06
ta_lk	Sri Lanka	Tamil	204.0	2.80
te_in	India	Telugu	77.0	3.13
th_th	Thailand	Thai	1,175.6	3.23
ti_et	Ethiopia	Tigrinya	14.0	3.10
tl_ph	Philippines	Tagalog	219.0	2.88
tn_bw	Botswana	Tswana	49.7	2.88
tn_za	South Africa	Tswana	0.9	2.69
tr_tr	Turkey	Turkish	1,007.6	2.22
ts_za	South Africa	Tsonga	0.2	2.75
ur_in	India	Urdu	12.6	3.01
ur_pk	Pakistan	Urdu	73.1	2.44
uz_uz	Uzbekistan	Uzbek	33.7	2.84
ve_za	South Africa	Venda	0.1	2.72
vn_vi	Vietnam	Vietnamese	726.0	2.90
xh_za	South Africa	Xhosa	10.1	2.74
yue_hk	Hong Kong	Cantonese	1,943.5	3.02
zh_tw	Taiwan	Mandarin	1,482.1	2.47
zu_za	South Africa	Zulu	19.0	2.71
Total	—	—	65,072.4	2.83

⚠️ Five configs (kh_km Khmer, la_lo Lao, mm_my Burmese, vn_vi Vietnamese, cn_ug Uyghur, sourced from Radio Free Asia include only metadata, transcripts, and source URLs. Audio must be downloaded separately from source_url.

Citation

If you use WorldSpeech in your research, please cite:

@misc{asonitis2026worldspeechmultilingualspeechcorpus,
      title={WorldSpeech: A Multilingual Speech Corpus from Around the World}, 
      author={Antonis Asonitis and Luca A. Lanzendörfer and Frédéric Berdoz and Roger Wattenhofer},
      year={2026},
      eprint={2605.09167},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2605.09167}, 
}

Dataset License

WorldSpeech is released under CC BY-NC 4.0. Individual source licenses apply per config and are listed in the Licenses section. Where a source license is more restrictive than CC BY-NC 4.0, the more restrictive terms govern that config.

Licenses

Config	Country / Language	License	Notes
af_za	South Africa / Afrikaans	Parliamentary Proceedings
am_et	Ethiopia / Amharic	Public Domain (17 USC §105)	US federal agency content (VOA).
ar_bh	Bahrain / Arabic	Parliamentary Proceedings
ar_dz	Algeria / Arabic	Parliamentary Proceedings
ar_eg	Egypt / Arabic	Parliamentary Proceedings
ar_iq	Iraq / Arabic	Parliamentary Proceedings
ar_kw	Kuwait / Arabic	Parliamentary Proceedings
ar_ma	Morocco / Arabic	Parliamentary Proceedings
ar_sa	Saudi Arabia / Arabic	Government Archive (Public Record)
ar_tn	Tunisia / Arabic	Parliamentary Proceedings
ar_un	United Nations / Arabic	UN Official Records	UN official records; non-commercial/research use per UN terms. Contact permissions@un.org for commercial use.
as_in	India / Assamese	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
az_az	Azerbaijan / Azerbaijani	Public Domain (17 USC §105)	US federal agency content (VOA).
be_by	Belarus / Belarusian	CC BY-NC-SA 4.0
bn_bd	Bangladesh / Bengali	Public Domain (17 USC §105)	US federal agency content (VOA).
bn_in	India / Bengali	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
ckb_iq	Iraq / Sorani Kurdish	Parliamentary Proceedings
cn_ug	China / Uyghur	RFA Terms of Use (non-commercial)	Metadata only — audio not embedded. Download audio from source_url. Non-commercial research use; see RFA terms.
cnr_me	Montenegro / Montenegrin	Parliamentary Proceedings
crs_sc	Seychelles / Kreol Seselwa	Parliamentary Proceedings
cs_cz	Czech Republic / Czech	Parliamentary Proceedings
cz_cs	Czech Republic / Czech	Parliamentary Proceedings	Czech Senate (Senát Parlamentu ČR).
de_at	Austria / German	Parliamentary Proceedings
de_li	Liechtenstein / German	Parliamentary Proceedings
dgo_in	India / Dogri	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
dv_mv	Maldives / Dhivehi	Parliamentary Proceedings
el_cy	Cyprus / Greek	Parliamentary Proceedings
el_gr	Greece / Greek	Parliamentary Proceedings
en_au	Australia / English	Parliamentary Proceedings
en_jm	Jamaica / English	Parliamentary Proceedings
en_ke	Kenya / English	Parliamentary Proceedings
en_nz	New Zealand / English	Parliamentary Proceedings
en_sl	Sierra Leone / English	Parliamentary Proceedings
en_us	USA / English	Public Domain (17 USC §105)	US federal proceedings (Congress.gov / C-SPAN public domain content).
en_zm	Zambia / English	Parliamentary Proceedings
eo	(international) / Esperanto	CC0 1.0 (Public Domain)
es_ar	Argentina / Spanish	Parliamentary Proceedings
es_cl	Chile / Spanish	Parliamentary Proceedings
es_co	Colombia / Spanish	Parliamentary Proceedings
es_es	Spain / Spanish	Parliamentary Proceedings
es_mx	Mexico / Spanish	Parliamentary Proceedings
es_pe	Peru / Spanish	Parliamentary Proceedings
es_pr	Puerto Rico / Spanish	Parliamentary Proceedings
es_py	Paraguay / Spanish	Parliamentary Proceedings
es_uy	Uruguay / Spanish	Parliamentary Proceedings
fa_ir	Iran / Persian	Public Domain (17 USC §105)	US federal agency content (VOA).
ca_fr	Canada / French	Parliamentary Proceedings
fr_cd	DR Congo / French	ICC Court Records	Public judicial proceedings of an international court; non-commercial/educational use per ICC terms.
fr_ci	Cote d'Ivoire / French	ICC Court Records	Public judicial proceedings of an international court; non-commercial/educational use per ICC terms.
ga_ie	Ireland / Irish	CC BY 4.0 (Oireachtas PSI Licence)
grc_gr	Greece / Ancient Greek	CC0 1.0 (Public Domain)
gu_in	India / Gujarati	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
ha_ng	Nigeria / Hausa	Public Domain (17 USC §105)	US federal agency content (VOA).
ha_td	Chad	Public Domain (17 USC §105)	US federal agency content (VOA).
he_il	Israel / Hebrew	CC0 1.0 (Public Domain)	LibriVox recordings (CC0) + Ben-Yehuda Project texts (public domain per Israeli copyright law).
hi_in	India / Hindi	Parliamentary Proceedings
hu_hu	Hungary / Hungarian	Parliamentary Proceedings
hy_am	Armenia / Armenian	Parliamentary Proceedings
id_id	Indonesia / Indonesian	Public Domain (17 USC §105)	US federal agency content (VOA).
ig_ng	Nigeria / Igbo	Public Domain (17 USC §105)	US federal agency content (VOA).
iu_ca	Canada / Inuktitut	Parliamentary Proceedings
ja_jp	Japan / Japanese	CC0 1.0 (Public Domain)	LibriVox recordings (CC0) + Aozora Bunko texts (public domain, 50yr Japanese law).
ka_ge	Georgia / Georgian	Parliamentary Proceedings
kh_km	Cambodia / Khmer	RFA Terms of Use (non-commercial)	Metadata only — audio not embedded. Download audio from source_url. Non-commercial research use; see RFA terms.
kk_kz	Kazakhstan / Kazakh	Parliamentary Proceedings
kn_in	India / Kannada	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
ko_kr	South Korea / Korean	Parliamentary Proceedings
kok_in	India / Konkani	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
la_lo	Laos / Lao	RFA Terms of Use (non-commercial)	Metadata only — audio not embedded. Download audio from source_url. Non-commercial research use; see RFA terms.
la_va	Vatican / Latin	Vatican Radio Archive	Vatican Radio historical archive; contact Vatican Media for commercial use.
lb_lu	Luxembourg / Luxembourgish	Parliamentary Proceedings
mai_in	India / Maithili	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
mfe_mu	Mauritius / Mauritian Creole	Parliamentary Proceedings
mi_nz	New Zealand / Maori	Parliamentary Proceedings
ml_in	India / Malayalam	Parliamentary Proceedings
mm_my	Myanmar / Burmese	RFA Terms of Use (non-commercial)	Metadata only — audio not embedded. Download audio from source_url. Non-commercial research use; see RFA terms.
mn_mn	Mongolia / Mongolian	Parliamentary Proceedings / CC0	Parliament sessions (public record) + Latter-day Saints addresses (CC0 per lds.org).
mr_in	India / Marathi	Parliamentary Proceedings
ms_my	Malaysia / Malay	Parliamentary Proceedings
ne_in	India / Nepali	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
ne_np	Nepal / Nepali	Parliamentary Proceedings
nl_be	Belgium / Dutch	Parliamentary Proceedings
nl_nl	Netherlands / Dutch	Parliamentary Proceedings
nr_za	South Africa / Southern Ndebele	Parliamentary Proceedings
nso_za	South Africa / Northern Sotho	Parliamentary Proceedings
om_et	Ethiopia / Oromo	Public Domain (17 USC §105)	US federal agency content (VOA).
or_in	India / Odia	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
pa_in	India / Punjabi	Parliamentary Proceedings
pl_pl	Poland / Polish	Parliamentary Proceedings
pt_br	Brazil / Portuguese	Parliamentary Proceedings
rm_ch	Switzerland / Romansh	SRG SSR Terms of Use	Corpus compiled from publicly broadcast RTR programming; contact SRG SSR for commercial use.
ro_md	Moldova / Romanian	Parliamentary Proceedings
ro_ro	Romania / Romanian	Parliamentary Proceedings
ru_by	Belarus / Russian	Government Publication (Public Record)
ru_ru	Russia / Russian	Parliamentary Proceedings
rw_rw	Rwanda / Kinyarwanda	Parliamentary Proceedings
si_lk	Sri Lanka / Sinhala	Parliamentary Proceedings
sm_ws	Samoa / Samoan	Parliamentary Proceedings
sn_zw	Zimbabwe / Shona	Public Domain (17 USC §105)	US federal agency content (VOA).
sq_al	Albania / Albanian	Parliamentary Proceedings
sq_xk	Kosovo / Albanian	Parliamentary Proceedings
ss_za	South Africa / Swati	Parliamentary Proceedings
st_za	South Africa / Sesotho	Parliamentary Proceedings
sv_ax	Aland Islands / Swedish	Parliamentary Proceedings
sw_ke	Kenya / Swahili	Parliamentary Proceedings
sw_tz	Tanzania / Swahili	Parliamentary Proceedings
ta_in	India / Tamil	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
ta_lk	Sri Lanka / Tamil	Parliamentary Proceedings
te_in	India / Telugu	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
th_th	Thailand / Thai	Parliamentary Proceedings
ti_et	Ethiopia / Tigrinya	Public Domain (17 USC §105)	US federal agency content (VOA).
tl_ph	Philippines / Tagalog	Parliamentary Proceedings
tn_bw	Botswana / Tswana	Parliamentary Proceedings
tn_za	South Africa / Tswana	Parliamentary Proceedings
tr_tr	Turkey / Turkish	Parliamentary Proceedings
ts_za	South Africa / Tsonga	Parliamentary Proceedings
ur_in	India / Urdu	All India Radio Archive (Prasar Bharati)	Prasar Bharati public broadcaster archive.
uz_uz	Uzbekistan / Uzbek	RFE/RL Terms of Use (non-commercial)	US government-funded broadcaster; non-commercial research use. Original Ozodlik (RFE/RL Uzbek) content.
ve_za	South Africa / Venda	Parliamentary Proceedings
vn_vi	Vietnam / Vietnamese	RFA Terms of Use (non-commercial)	Metadata only — audio not embedded. Download audio from source_url. Non-commercial research use; see RFA terms.
xh_za	South Africa / Xhosa	Parliamentary Proceedings
yue_hk	Hong Kong / Cantonese	Parliamentary Proceedings
zh_tw	Taiwan / Chinese (Mandarin)	Parliamentary Proceedings
zu_za	South Africa / Zulu	Parliamentary Proceedings