quran-align timing file package =============================== https://github.com/cpfair/quran-align License ------- These data files are licensed under a [Creative Commons Attribution 4.0 International License](https://creativecommons.org/licenses/by/4.0/). Please consider emailing me if you use this data, so I can let you know when new & revised timing data is available. Contents -------- * Abdul_Basit_Mujawwad_128kbps.json 24 November 2016 d0c5d63917c0a2d59f0d86ce27936251b6d76f90 * Abdul_Basit_Murattal_64kbps.json 24 November 2016 029e32d2cdc869e2f33ae86fd21214952d18f33d * Abdurrahmaan_As-Sudais_192kbps.json 24 November 2016 4315194bc88db14192201426fb4f6617bffac4ac * Abu_Bakr_Ash-Shaatree_128kbps.json 24 November 2016 8861fadb68bad84ae5189227e0bcdd7a7127e976 * Alafasy_128kbps.json 24 November 2016 167bfc53255152aaf8c9a2e5e79c1160134c77a4 * Hani_Rifai_192kbps.json 24 November 2016 22e0cea40ff434848295f8ce959ab885292ebb7b * Husary_64kbps.json 24 November 2016 8e7a24f66f98f176dfddf65b434643a6655c694d * Husary_Muallim_128kbps.json 24 November 2016 b03e6c9188934504b6921d22b1f2e2b72ee773eb * Minshawy_Mujawwad_192kbps.json 24 November 2016 da6c3850917c6e104e650a07542a6c7341539af5 * Minshawy_Murattal_128kbps.json 24 November 2016 2a24d42e7f3a018c56600fc4ddd580f214b20a39 * Mohammad_al_Tablaway_128kbps.json 24 November 2016 dce92289eebd76380d1dae22f2329493a49f06db * Saood_ash-Shuraym_128kbps.json 24 November 2016 59fd80d144caa6fe79a917a8d1dddf18087b3bcc Format ------ Data files are JSON, of the following format: [ { "surah": 1, "ayah": 1, "segments": { [word_start_index, word_end_index, start_msec, end_msec], ... }, "stats": { "insertions": 123, "deletions": 456, "transpositions": 789 } }, ... ] Where... * `word_start_index` is the 0-base index of the first word contained in the segment. * `word_end_index` is the 0-base index of the word _after_ the last word contained in the segment. * `start_msec`/`end_msec` are timestamps within the input audio file. * `stats` contain statistics from the matching routine that aligns the recognized words with reference text. Here, a "word" is defined by splitting the text of the Qur'an by spaces (specifically, `quran-uthmani.txt` from [Tanzil.net](http://tanzil.net/download) - without me_quran tanween differentiation!). Note that the language model used for recognition treats muqata'at as sets of words (ا ل م instead of الم) - but they will appear as a single word in the alignment output.