Learning Nigerian accent embeddings from speech: preliminary results based on SautiDB-Naija corpus
It addresses the need for accent-specific speech resources in linguistics and AI, though it is incremental as it builds on existing methods with new data.
This paper tackles the problem of learning Nigerian accent embeddings from speech by introducing SautiDB-Naija, a novel corpus of over 900 non-native English recordings, and demonstrates that fine-tuning a pre-trained model yields representations suitable for accent classification.
This paper describes foundational efforts with SautiDB-Naija, a novel corpus of non-native (L2) Nigerian English speech. We describe how the corpus was created and curated as well as preliminary experiments with accent classification and learning Nigerian accent embeddings. The initial version of the corpus includes over 900 recordings from L2 English speakers of Nigerian languages, such as Yoruba, Igbo, Edo, Efik-Ibibio, and Igala. We further demonstrate how fine-tuning on a pre-trained model like wav2vec can yield representations suitable for related speech tasks such as accent classification. SautiDB-Naija has been published to Zenodo for general use under a flexible Creative Commons License.