Common Voice: A Massively-Multilingual Speech Corpus

Abstract

The Common Voice corpus is a massively-multilingual collection of transcribed speech intended for speech technology research and development. Common Voice is designed for Automatic Speech Recognition purposes but can be useful in other domains (e.g. language identification). To achieve scale and sustainability, the Common Voice project employs crowdsourcing for both data collection and data validation. The most recent release includes 29 languages, and as of November 2019 there are a total of 38 languages collecting data. Over 50,000 individuals have participated so far, resulting in 2,500 hours of collected audio. To our knowledge this is the largest audio corpus in the public domain for speech recognition, both in terms of number of hours and number of languages. As an example use case for Common Voice, we present speech recognition experiments using Mozilla's DeepSpeech Speech-to-Text toolkit. By applying transfer learning from a source English model, we find an average Character Error Rate improvement of 5.99 +/- 5.48 for twelve target languages (German, French, Italian, Turkish, Catalan, Slovenian, Welsh, Irish, Breton, Tatar, Chuvash, and Kabyle). For most of these languages, these are the first ever published results on end-to-end Automatic Speech Recognition.

arXiv commentAccepted to LREC 2020

Cited by in corpus (40)

• BigSSL: Exploring the Frontier of Large-Scale Semi-Supervised Learning for Automatic Speech Recognition
• CTC-Segmentation of Large Corpora for German End-to-end Speech Recognition
• SLAM: A Unified Encoder for Speech and Language Modeling via Speech-Text Joint Pre-Training
• UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data
• Few-Shot Keyword Spotting in Any Language
• Neural network interpretation using descrambler groups
• Uncertainty Estimation in Autoregressive Structured Prediction
• Adapting Multilingual Speech Representation Model for a New, Underresourced Language through Multilingual Fine-tuning and Continued Pretraining
• Resources for Turkish Natural Language Processing: A critical survey
• Acoustic correlates of the syllabic rhythm of speech: Modulation spectrum or local features of the temporal envelope
• Cross-Language Transfer Learning, Continuous Learning, and Domain Adaptation for End-to-End Automatic Speech Recognition
• Lossless data compression by large models
• Exploiting Pre-Trained ASR Models for Alzheimer's Disease Recognition Through Spontaneous Speech
• Joint gender and age estimation based on speech signals using x-vectors and transfer learning
• Materiality and Risk in the Age of Pervasive AI Sensors
• Simple and Effective Zero-shot Cross-lingual Phoneme Recognition
• Swiss Parliaments Corpus, an Automatically Aligned Swiss German Speech to Standard German Text Corpus
• Federated Learning in ASR: Not as Easy as You Think
• What all do audio transformer models hear? Probing Acoustic Representations for Language Delivery and its Structure
• Hi-Fi Multi-Speaker English TTS Dataset
• Improving Noise Robustness of an End-to-End Neural Model for Automatic Speech Recognition
• Semantic-WER: A Unified Metric for the Evaluation of ASR Transcript for End Usability
• Automatic Speech Recognition Benchmark for Air-Traffic Communications
• Romanian Speech Recognition Experiments from the ROBIN Project
• Self-Supervised Representations Improve End-to-End Speech Translation
• Scribosermo: Fast Speech-to-Text models for German and other Languages
• A Framework for Adapting Human-Robot Interaction to Diverse User Groups
• Replay to Remember: Continual Layer-Specific Fine-tuning for German Speech Recognition
• Transfer Learning of Transformer-based Speech Recognition Models from Czech to Slovak
• Howl: A Deployed, Open-Source Wake Word Detection System
• Cross-Modal Knowledge Distillation Method for Automatic Cued Speech Recognition
• Zero-Shot vs. Few-Shot Multi-Speaker TTS Using Pre-trained Czech SpeechT5 Model
• Evaluation of End-to-End Continuous Spanish Lipreading in Different Data Conditions
• A GAN-based Approach for Mitigating Inference Attacks in Smart Home Environment
• The HUAWEI Speaker Diarisation System for the VoxCeleb Speaker Diarisation Challenge
• Improving Streaming Automatic Speech Recognition With Non-Streaming Model Distillation On Unsupervised Data
• Cross-modal Speaker Verification and Recognition: A Multilingual Perspective
• QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpus
• Minimum Word Error Rate Training with Language Model Fusion for End-to-End Speech Recognition
• Topic Model Robustness to Automatic Speech Recognition Errors in Podcast Transcripts