# Fast, Compact, and High Quality LSTM-RNN Based Statistical Parametric Speech Synthesizers for Mobile Devices

Source: [Fast, Compact, and High Quality LSTM-RNN Based Statistical Parametric Speech Synthesizers for Mobile Devices](<https://www.sota2.com/research/paper/68fa629d1e209131ce9fa36c>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T20:22:26.903Z

Authors: Heiga Zen, Yannis Agiomyrgiannakis, Niels Egberts, Fergus Henderson, Przemys{\\l}aw Szczepaniak

Published: 2016-06-20T10:54:51.000Z

## Abstract

Acoustic models based on long short-term memory recurrent neural networks (LSTM-RNNs) were applied to statistical parametric speech synthesis (SPSS) and showed significant improvements in naturalness and latency over those based on hidden Markov models (HMMs). This paper describes further optimizations of LSTM-RNN-based SPSS for deployment on mobile devices; weight quantization, multi-frame inference, and robust inference using an {\\epsilon}-contaminated Gaussian loss function. Experimental results in subjective listening tests show that these optimizations can make LSTM-RNN-based SPSS comparable to HMM-based SPSS in runtime speed while maintaining naturalness. Evaluations between LSTM-RNN- based SPSS and HMM-driven unit selection speech synthesis are also presented.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Text-to-Speech on Mandarin Chinese TTS dataset single-speaker (test)](<https://www.sota2.com/research/sota/text-to-speech-on-mandarin-chinese-tts-dataset-single-speaker-test.md>): Text-to-Speech; Mandarin Chinese TTS dataset single-speaker (test); MOS: 3.79.

- [Text-to-Speech on North American English TTS dataset single-speaker (test)](<https://www.sota2.com/research/sota/text-to-speech-on-north-american-english-tts-dataset-single-speaker-test.md>): Text-to-Speech; North American English TTS dataset single-speaker (test); MOS: 3.67.

- [Text-to-speech on Internal US English dataset (test)](<https://www.sota2.com/research/sota/text-to-speech-on-internal-us-english-dataset-test.md>): Text-to-Speech; Internal US English dataset (test); MOS: 3.492.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1606.06061>)

[PDF](<https://arxiv.org/pdf/1606.06061.pdf>)