From Tens of Hours to Tens of Thousands: Scaling Back-Translation for Speech Recognition

• URL: http://arxiv.org/abs/2505.16972v1
• Date: Thu, 22 May 2025 17:51:05 GMT
• Title: From Tens of Hours to Tens of Thousands: Scaling Back-Translation for Speech Recognition
• Authors: Tianduo Wang, Lu Xu, Wei Lu, Shanbo Cheng,
• Abstract summary: Speech Back-Translation is a scalable pipeline that improves multilingual ASR models by converting large-scale text corpora into synthetic speech.<n>We generate more than 500,000 hours of synthetic speech in ten languages and continue pre-training Whisper-large-v3, achieving average transcription error reductions of over 30%.
• Score: 14.155874873165853
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent advances in Automatic Speech Recognition (ASR) have been largely fueled by massive speech corpora. However, extending coverage to diverse languages with limited resources remains a formidable challenge. This paper introduces Speech Back-Translation, a scalable pipeline that improves multilingual ASR models by converting large-scale text corpora into synthetic speech via off-the-shelf text-to-speech (TTS) models. We demonstrate that just tens of hours of real transcribed speech can effectively train TTS models to generate synthetic speech at hundreds of times the original volume while maintaining high quality. To evaluate synthetic speech quality, we develop an intelligibility-based assessment framework and establish clear thresholds for when synthetic data benefits ASR training. Using Speech Back-Translation, we generate more than 500,000 hours of synthetic speech in ten languages and continue pre-training Whisper-large-v3, achieving average transcription error reductions of over 30\%. These results highlight the scalability and effectiveness of Speech Back-Translation for enhancing multilingual ASR systems.

Related papers

• Stuttering-Aware Automatic Speech Recognition for Indonesian Language [0.04666493857924358]
  We propose a data augmentation framework that generates synthetic stuttered audio by injecting repetitions and prolongations into fluent text.<n>We apply this synthetic data to fine-tune a pre-trained Indonesian Whisper model using transfer learning.<n>Our experiments demonstrate that this targeted synthetic exposure consistently reduces recognition errors on stuttered speech while maintaining performance on fluent segments.
  arXiv  Detail & Related papers  (2026-01-07T09:21:12Z)
• Improving Code-Switching Speech Recognition with TTS Data Augmentation [58.34842693152991]
  This paper explores multilingual text-to-speech (TTS) models as an effective data augmentation technique to address this shortage.<n>We fine-tune the multilingual CosyVoice2 TTS model on the SEAME dataset to generate synthetic conversational Chinese-English code-switching speech.
  arXiv  Detail & Related papers  (2026-01-02T10:11:51Z)
• Bi-directional Context-Enhanced Speech Large Language Models for Multilingual Conversational ASR [23.285609467633865]
  This paper introduces the integration of language-specific bi-directional context into a speech large language model (SLLM) to improve multilingual continuous conversational automatic speech recognition (ASR)<n>We propose a character-level contextual masking strategy during training, which randomly removes portions of the context to enhance robustness and better emulate the flawed transcriptions that may occur during inference.
  arXiv  Detail & Related papers  (2025-06-16T12:03:23Z)
• CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training [70.31925012315064]
  We present CosyVoice 3, an improved model designed for zero-shot multilingual speech synthesis in the wild.<n>Key features of CosyVoice 3 include a novel speech tokenizer to improve prosody naturalness.<n>Data is expanded from ten thousand hours to one million hours, encompassing 9 languages and 18 Chinese dialects.
  arXiv  Detail & Related papers  (2025-05-23T07:55:21Z)
• An Exhaustive Evaluation of TTS- and VC-based Data Augmentation for ASR [12.197936305117407]
  Augmenting the training data of automatic speech recognition systems with synthetic data generated by text-to-speech (TTS) or voice conversion (VC) has gained popularity in recent years.<n>We leverage recently proposed flow-based TTS/VC models allowing greater speech diversity, and assess the respective impact of augmenting various speech attributes on the word error rate (WER) achieved by several ASR models.
  arXiv  Detail & Related papers  (2025-03-11T23:09:06Z)
• Scaling Speech-Text Pre-training with Synthetic Interleaved Data [31.77653849518526]
  Speech language models (SpeechLMs) accept speech input and produce speech output, allowing for more natural human-computer interaction.<n>Traditional approaches for developing SpeechLMs are constrained by the limited availability of unsupervised speech data and parallel speech-text data.<n>We propose a novel approach to scaling speech-text pre-training by leveraging large-scale synthetic interleaved data derived from text corpora.
  arXiv  Detail & Related papers  (2024-11-26T17:19:09Z)
• Towards Unsupervised Speech Recognition Without Pronunciation Models [57.222729245842054]
  In this article, we tackle the challenge of developing ASR systems without paired speech and text corpora.<n>We experimentally demonstrate that an unsupervised speech recognizer can emerge from joint speech-to-speech and text-to-text masked token-infilling.<n>This innovative model surpasses the performance of previous unsupervised ASR models under the lexicon-free setting.
  arXiv  Detail & Related papers  (2024-06-12T16:30:58Z)
• HierSpeech++: Bridging the Gap between Semantic and Acoustic Representation of Speech by Hierarchical Variational Inference for Zero-shot Speech Synthesis [39.892633589217326]
  Large language models (LLM)-based speech synthesis has been widely adopted in zero-shot speech synthesis. This paper proposes HierSpeech++, a fast and strong zero-shot speech synthesizer for text-to-speech (TTS) and voice conversion (VC)
  arXiv  Detail & Related papers  (2023-11-21T09:07:11Z)
• NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers [90.83782600932567]
  We develop NaturalSpeech 2, a TTS system that leverages a neural audio predictor with residual vectorizers to get the quantized latent vectors. We scale NaturalSpeech 2 to large-scale datasets with 44K hours of speech and singing data and evaluate its voice quality on unseen speakers. NaturalSpeech 2 outperforms previous TTS systems by a large margin in terms of prosody/timbre similarity, synthesis, and voice quality in a zero-shot setting.
  arXiv  Detail & Related papers  (2023-04-18T16:31:59Z)
• Google USM: Scaling Automatic Speech Recognition Beyond 100 Languages [76.95115818308918]
  We introduce the Universal Speech Model (USM), a single large model that performs automatic speech recognition (ASR) across 100+ languages. This is achieved by pre-training the encoder of the model on a large unlabeled multilingual dataset of 12 million (M) hours spanning over 300 languages. We use multilingual pre-training with random-projection quantization and speech-text modality matching to achieve state-of-the-art performance on downstream multilingual ASR and speech-to-text translation tasks.
  arXiv  Detail & Related papers  (2023-03-02T07:47:18Z)
• ASR data augmentation in low-resource settings using cross-lingual multi-speaker TTS and cross-lingual voice conversion [49.617722668505834]
  We show that our approach permits the application of speech synthesis and voice conversion to improve ASR systems using only one target-language speaker during model training. It is possible to obtain promising ASR training results with our data augmentation method using only a single real speaker in a target language.
  arXiv  Detail & Related papers  (2022-03-29T11:55:30Z)
• Transfer Learning Framework for Low-Resource Text-to-Speech using a Large-Scale Unlabeled Speech Corpus [10.158584616360669]
  Training a text-to-speech (TTS) model requires a large scale text labeled speech corpus. We propose a transfer learning framework for TTS that utilizes a large amount of unlabeled speech dataset for pre-training.
  arXiv  Detail & Related papers  (2022-03-29T11:26:56Z)
• LRSpeech: Extremely Low-Resource Speech Synthesis and Recognition [148.43282526983637]
  We develop LRSpeech, a TTS and ASR system for languages with low data cost. We conduct experiments on an experimental language (English) and a truly low-resource language (Lithuanian) to verify the effectiveness of LRSpeech. We are currently deploying LRSpeech into a commercialized cloud speech service to support TTS on more rare languages.
  arXiv  Detail & Related papers  (2020-08-09T08:16:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.