The LibriSpeech-Phonetics repository is a rich collection of multi-modal data subsets, which includes waveform (wav files), acoustic (egemaps), phonetic (phonetic alignments), and alphabetic (textual transcripts) data types. These subsets are organized into train-clean, dev-clean, and test-clean sections.
- waveform: https://cmu.box.com/s/b9ww3xjcfvn4frr6uj8i73rz21ohf4z1
- acoustic: https://cmu.box.com/s/p7geisd7g1c9suj3p1n11zsgxk3o7qmq
- phonetic: https://cmu.box.com/s/j0lt4bnk3oour2vo75hs2wkqfud0qja2
- alphabetic: https://cmu.box.com/s/89ukay99h6e8qm46g5g0z9oquep59f3a
- waveform: https://cmu.box.com/s/gedbtpdp4dw2tpmiqc1pi9xyems147vi
- acoustic: https://cmu.box.com/s/ze26iex6pmfpan2o4t1rfv0pn2t8vxvq
- phonetic: https://cmu.box.com/s/jbp9nqcrzvyjr74x5dg1v2ju9anerxi4
- alphabetic: https://cmu.box.com/s/c0vbo4uggord96ubw2audr5zj30ttzxb
- waveform: https://cmu.box.com/s/u1acpkxfw9rtpw5k091gzdk6dvm0izil
- acoustic: https://cmu.box.com/s/iq94rqdkjgb3qosvv7agwc2iejh2cx1g
- phonetic: https://cmu.box.com/s/xc0m8k9yn5yxs54nmqd10pn7w3t82k89
- alphabetic: https://cmu.box.com/s/ifhgjf0apmzdzm4n5q6zhr7ms2bbb9i3
- waveform: https://cmu.box.com/s/067jr86szovd9v1idkd54dty4nm3z7yr
- acoustic: https://cmu.box.com/s/7179fg6z0x82urd0qzfwob7kmze95a0f
- phonetic: https://cmu.box.com/s/f2ward9mp6wjcl25osghv0aoplowg8qx
- alphabetic: https://cmu.box.com/s/011mmc8fa61g7jicbg606kdepfxleo4y
This resource is publicly available for use in research and development. It provides a versatile tool for benchmarking and developing novel methods for speech processing tasks. Please ensure to cite this repository appropriately in any resulting publications.