Latest Articles for Speech Technology

Innovative techniques for improving TTS models and reducing knowledge loss.

This study reviews how batch size influences speech model performance and training.

A new method enhances speech model performance and efficiency in noisy environments.

A study on improving TTS systems with diverse voice samples.

Research identifies and classifies Sorani Kurdish dialects using extensive audio recordings.

RALL-E enhances text-to-speech synthesis for clearer, more natural speech.

New methods improve audio representation through self-supervised learning techniques.

New model allows precise control of voice qualities while retaining content.

A new framework for assessing foundation models in speech tasks.

Study reveals users prefer static speech agents over adaptive ones.

FlashSpeech offers rapid, high-quality speech synthesis solutions.

SEANet improves speaker isolation by reducing noise in audio processing.

A two-stage active learning method enhances speech recognition accuracy with less data.

This study evaluates ASR systems' performance with individuals who stutter.

This article investigates vulnerabilities in speech models and ways to enhance their security.

New methods improve how machines recognize emotions in speech.

Seed-TTS creates lifelike speech from text for various applications.

New model ARDiT improves text-to-speech synthesis and speech editing.

mHuBERT-147 processes speech in multiple languages efficiently.

New methods enhance speech recognition in noisy environments using adaptive techniques.

A novel method optimizing speech analysis and synthesis using vocal tract movements.

A study on enhancing audio segmentation by integrating speaker embeddings.

New efforts aim to support Yoruba dialects in language technology.

This article discusses how Wav2Vec2.0 processes speech sounds using phonology.

This study evaluates speech technology in low-resource languages like Tunisian Arabic.

Enhancing speech synthesis for more natural and expressive voice generation.

Introducing a method for better control in speech editing.

Emilia provides a diverse dataset for improving speech generation models.

Mamba shows promise against transformers in speech tasks, especially for long inputs.

A new method enhances stuttering detection by combining audio, video, and text data.

Research presents new methods for evaluating speech recognition systems in Polish.

A new dataset enhances machine speech for Mandarin, aiming for natural expression.

Explore the growing importance of speech editing for content creators.

New methods improve speech systems for underrepresented languages.

Research combines speech enhancement and transfer learning for better anti-spoofing systems.

New methods enhance emotional expression in machine speech synthesis.

Speech-MASSIVE aims to enhance spoken language understanding in various languages.

Innovative techniques protect sensitive speech data while maintaining processing accuracy.

OpenOmni builds flexible tools for creating and testing conversation agents.

SSL-TTS simplifies voice synthesis using minimal training data for high-quality results.