Home

Welcome to the Eyra wiki!

Running a new Collection

Some questions to ask yourself: What is the point of the collection? Who will be working on what? How many hours of data do you need? What kind of speakers do you need?

Specifications:

• Server with RAM, CPUs, and 100GB of hard Disk space to be hosted for the duration of the collection (1-6 months is usually enough)
• Phones with the app
• Data collection administrators & users willing to lend their voice
• Legal aspects of collecting people\s voices such as a waiver if the data is open sourced
  • waiver agreement for using their voice in this collection, an example in English
• Prompt list for the chosen language
• Linguist to help make the acoustic model, language model, lexicon, and comprehensive prompt list
• If necessary then someone to translate the UI of Eyra
• Server admin to extract data, clean it, and sanitize it