Full licence texts: Arphic-1999 · CC-BY-SA-4.0 · CJK-Decomp-Apache-2.0 · Hanzi-Writer-MIT · Lucide-ISC · Taiwan-OGD-1.0 · Unicode-License-v3
Download the stroke dataset and its licence
Independent HanziSnack study pilot Character list, readings and forms: Unicode Unihan 17.0, Unicode License v3. Visual reference: CJK Decomposition Data by Gavin Grover and contributors, Apache 2.0. https://github.com/amake/cjk-decomp Meanings and example words: CC-CEDICT contributors and MDBG, CC BY-SA 4.0. https://cc-cedict.org/wiki/ Stroke data: Hanzi Writer Data 2.0.1, Arphic Public License, subset to pilot glyphs. Hanzi Writer and Lucide: MIT and ISC. See the bundled licence texts. Frequency signal from Tatoeba Mandarin sentence export, Tatoeba contributors, CC BY 2.0 France. Han glyphs counted after deduplication. https://downloads.tatoeba.org/exports/per_language/cmn/cmn_sentences.tsv.bz2 https://creativecommons.org/licenses/by/2.0/fr/ Pinyin and example-word pronunciation 376 syllables and the 768 example words are synthetic speech, not recordings of a person. Generated with ElevenLabs text to speech (model eleven_v3, voice "Anna Su - Trustworthy, Clear and Natural" from the ElevenLabs Voice Library) on a paid subscription, which grants a commercial licence for the generated audio. Each syllable clip is cut from a short Chinese input; each example word is synthesised whole. All clips are levelled, slowed to 80 % tempo without changing pitch, and re-encoded. See pinyin-audio-sources.json in the source repository for the input text and processing of every clip. https://elevenlabs.io/terms https://elevenlabs.io/service-specific-terms Two neutral-tone syllables (ma, ne). © 2006 Wei Gao (speaker) and Nicolas Vion, Base Audio Libre de Mots Chinois (Shtooka Project). Creative Commons Attribution 2.0 France. Transcoded to MP3 and slowed to 80 % tempo. https://creativecommons.org/licenses/by/2.0/fr/ https://commons.wikimedia.org/wiki/File:Zh-ma.ogg https://commons.wikimedia.org/wiki/File:Zh-ne.ogg