Improving Robustness of LLM-based Speech Synthesis by Learning Monotonic Alignment
- URL: http://arxiv.org/abs/2406.17957v1
- Date: Tue, 25 Jun 2024 22:18:52 GMT
- Title: Improving Robustness of LLM-based Speech Synthesis by Learning Monotonic Alignment
- Authors: Paarth Neekhara, Shehzeen Hussain, Subhankar Ghosh, Jason Li, Rafael Valle, Rohan Badlani, Boris Ginsburg,
- Abstract summary: Large Language Model (LLM) based text-to-speech (TTS) systems have demonstrated remarkable capabilities in handling large speech datasets and generating natural speech for new speakers.
However, LLM-based TTS models are not robust as the generated output can contain repeating words, missing words and mis-aligned speech.
We examine these challenges in an encoder-decoder transformer model and find that certain cross-attention heads in such models implicitly learn the text and speech alignment when trained for predicting speech tokens for a given text.
- Score: 19.48653924804823
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Large Language Model (LLM) based text-to-speech (TTS) systems have demonstrated remarkable capabilities in handling large speech datasets and generating natural speech for new speakers. However, LLM-based TTS models are not robust as the generated output can contain repeating words, missing words and mis-aligned speech (referred to as hallucinations or attention errors), especially when the text contains multiple occurrences of the same token. We examine these challenges in an encoder-decoder transformer model and find that certain cross-attention heads in such models implicitly learn the text and speech alignment when trained for predicting speech tokens for a given text. To make the alignment more robust, we propose techniques utilizing CTC loss and attention priors that encourage monotonic cross-attention over the text tokens. Our guided attention training technique does not introduce any new learnable parameters and significantly improves robustness of LLM-based TTS models.
Related papers
- DC-Spin: A Speaker-invariant Speech Tokenizer for Spoken Language Models [45.791472119671916]
Spoken language models (SLMs) process text and speech, enabling simultaneous speech understanding and generation.
DC-Spin aims to improve speech tokenization by bridging audio signals and SLM tokens.
We propose a chunk-wise approach to enable streamable DC-Spin without retraining and degradation.
arXiv Detail & Related papers (2024-10-31T17:43:13Z) - Developing Instruction-Following Speech Language Model Without Speech Instruction-Tuning Data [84.01401439030265]
Recent end-to-end speech language models (SLMs) have expanded upon the capabilities of large language models (LLMs)
We present a simple yet effective automatic process for creating speech-text pair data.
Our model demonstrates general capabilities for speech-related tasks without the need for speech instruction-tuning data.
arXiv Detail & Related papers (2024-09-30T07:01:21Z) - CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokens [49.569695524535454]
We propose to represent speech with supervised semantic tokens, which are derived from a multilingual speech recognition model by inserting vector quantization into the encoder.
Based on the tokens, we further propose a scalable zero-shot TTS synthesizer, CosyVoice, which consists of an LLM for text-to-token generation and a conditional flow matching model for token-to-speech synthesis.
arXiv Detail & Related papers (2024-07-07T15:16:19Z) - Enhancing the Stability of LLM-based Speech Generation Systems through
Self-Supervised Representations [14.437646262239612]
Self-supervised Voice Conversion (VC) architecture can be used to learn to encode transitory features, such as content, separately from stationary ones, such as speaker ID or recording conditions, creating speaker-disentangled representations.
Using speaker-disentangled codes to train LLMs for text-to-speech (TTS) allows the LLM to generate the content and the style of the speech only from the text, similarly to humans, while the speaker identity is provided by the decoder of the VC model.
Results show that LLMs trained over speaker-disentangled self-supervised representations provide an improvement of 4.7pp
arXiv Detail & Related papers (2024-02-05T15:08:19Z) - On decoder-only architecture for speech-to-text and large language model
integration [59.49886892602309]
Speech-LLaMA is a novel approach that effectively incorporates acoustic information into text-based large language models.
We conduct experiments on multilingual speech-to-text translation tasks and demonstrate a significant improvement over strong baselines.
arXiv Detail & Related papers (2023-07-08T06:47:58Z) - Code-Switching Text Generation and Injection in Mandarin-English ASR [57.57570417273262]
We investigate text generation and injection for improving the performance of an industry commonly-used streaming model, Transformer-Transducer (T-T)
We first propose a strategy to generate code-switching text data and then investigate injecting generated text into T-T model explicitly by Text-To-Speech (TTS) conversion or implicitly by tying speech and text latent spaces.
Experimental results on the T-T model trained with a dataset containing 1,800 hours of real Mandarin-English code-switched speech show that our approaches to inject generated code-switching text significantly boost the performance of T-T models.
arXiv Detail & Related papers (2023-03-20T09:13:27Z) - A Vector Quantized Approach for Text to Speech Synthesis on Real-World
Spontaneous Speech [94.64927912924087]
We train TTS systems using real-world speech from YouTube and podcasts.
Recent Text-to-Speech architecture is designed for multiple code generation and monotonic alignment.
We show thatRecent Text-to-Speech architecture outperforms existing TTS systems in several objective and subjective measures.
arXiv Detail & Related papers (2023-02-08T17:34:32Z) - Towards Language Modelling in the Speech Domain Using Sub-word
Linguistic Units [56.52704348773307]
We propose a novel LSTM-based generative speech LM based on linguistic units including syllables and phonemes.
With a limited dataset, orders of magnitude smaller than that required by contemporary generative models, our model closely approximates babbling speech.
We show the effect of training with auxiliary text LMs, multitask learning objectives, and auxiliary articulatory features.
arXiv Detail & Related papers (2021-10-31T22:48:30Z) - SLAM: A Unified Encoder for Speech and Language Modeling via Speech-Text
Joint Pre-Training [33.02912456062474]
We build a single encoder with the BERT objective on unlabeled text together with the w2v-BERT objective on unlabeled speech.
We demonstrate that incorporating both speech and text data during pre-training can significantly improve downstream quality on CoVoST2 speech translation.
arXiv Detail & Related papers (2021-10-20T00:59:36Z)
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.