Tarteel is an open-source project designed to help build digital tools to analyze the recitation of the Quran. Given the important place of reciting the Quran in the lives of Muslims, it is important to build software tools that can help ordinary Muslims recite the Quran with greater accuracy and appreciation. The name tarteel comes from the Quran itself, where God commands us to "recite the Quran with tarteel (slow, measured rhythmic tones)" (73:4).
The goal of the Tarteel 10,000 Challenge is to build the world's first public dataset of Quranic recitations carried out by ordinary Muslims. Most of the available audio of the Quran being recited is from professional reciters with strong fluency in tajweed (rules of recitation) and is recorded in studios. This is valuable when someone wants to listen to a recitation of the Quran.
However, many software tools that Muslim developers are interested in building require training machine learning models on Quranic recitation (e.g. to convert recitation2text), as it is recited by ordinary Muslims. The recitation of ordinary Muslims differs from professional recordings in many ways: for example, it may include background noise, or maybe recited by people with limited knowledge of tajweed, or the demographics of reciters may be different. By collecting this data, we can train machine learning models, which we will release to software developers who are interested in developing a wide variety of applications that are based on recitation2text, things like:
- Hifz helping tools that automatically correct mistakes
- Tajweed teaching tools in a similar vein
- Masjid kiosks that follow the imam and display the translation of the verse.
Tarteel is an open-source project that is maintained by developers interested in furthering the mission of Tarteel. If you'd like to contribute, please check out our GitHub repo: https://github.com/abidlabs/hifz.io.