7 papers
Sometin Beta Pass Notin: Improving Multilingual ASR for Nigerian Languages via Knowledge Distillation
Sewade Ogun
Although modern multilingual Automatic Speech Recognition (ASR) systems support several Nigerian languages, their performance consistently lags behind resource-rich languages such…
The NaijaVoices Dataset: Cultivating Large-Scale, High-Quality, Culturally-Rich Speech Data for African Languages
Chris Emezue, NaijaVoices Community, Busayo Awobade +8
The development of high-performing, robust, and reliable speech technologies depends on large, high-quality datasets. However, African languages -- including our focus, Igbo, Hausa…
An Exhaustive Evaluation of TTS- and VC-based Data Augmentation for ASR
Sewade Ogun, Vincent Colotte, Emmanuel Vincent
Augmenting the training data of automatic speech recognition (ASR) systems with synthetic data generated by text-to-speech (TTS) or voice conversion (VC) has gained popularity in r…
1000 African Voices: Advancing inclusive multi-speaker multi-accent speech synthesis
Sewade Ogun, Abraham T. Owodunni, Tobi Olatunji +6
Recent advances in speech synthesis have enabled many useful applications like audio directions in Google Maps, screen readers, and automated content generation on platforms like T…
Performant ASR Models for Medical Entities in Accented Speech
Tejumade Afonja, Tobi Olatunji, Sewade Ogun +3
Recent strides in automatic speech recognition (ASR) have accelerated their application in the medical domain where their performance on accented medical named entities (NE) such a…
Stochastic Pitch Prediction Improves the Diversity and Naturalness of Speech in Glow-TTS
Sewade Ogun, Vincent Colotte, Emmanuel Vincent
Flow-based generative models are widely used in text-to-speech (TTS) systems to learn the distribution of audio features (e.g., Mel-spectrograms) given the input tokens and to samp…
Can we use Common Voice to train a Multi-Speaker TTS system?
Sewade Ogun, Vincent Colotte, Emmanuel Vincent
Training of multi-speaker text-to-speech (TTS) systems relies on curated datasets based on high-quality recordings or audiobooks. Such datasets often lack speaker diversity and are…