text-to-speech-training
Various data for text-to-speech (TTS) training.
Harvard sentences
Original IEEE Harvard sentences
The IEEE Harvard sentences is a collection of 720 phonemically balanced short sample phrases in English.
Background info
Harvard-NGSL sentences
Kakeru Yazawa has produced a subset of 50 IEEE Harvard sentences from the full list of 720. These are phonemically balanced sentences for English learner speech corpora named the Harvard-NGSL sentences. These comprise core high-frequency vocabulary words for second language English learners were selected based on the New General Service List (NGSL), and are thus learner-oriented, unlike most other phoemically balanced materials which contain low-frequency and low-familiarity words that are of limited use for learners.
Harvard stories
- Harvard-NGSL story 1 “Life’s Challenges and Choices”
- Harvard-NGSL story 2 “A Day Full of Surprises”
- Harvard story 1 “A Day of Many Activities”
- Harvard story 2 “A Day at the Circus and Beyond”
In 2023, Adam Twardoch used ChatGPT to convert the Harvard-NGSL sentences and two sets of 60 original IEEE Harvard sentences into short stories. These short stories use the original vocabulary, but are formatted for more fluent reading.