Latest Articles for Speech Processing

DUSTED improves efficiency in identifying spoken words by analyzing phonetic patterns.

DualSpeech model improves TTS clarity and speaker resemblance.

Research improves speech recognition for Hindi with diverse accents.

X-Codec improves audio generation by integrating semantic understanding into processing.

This study enhances SER through improved preprocessing and efficient attention models.

Research focuses on enhancing language learning through visually grounded speech models.

New methods improve voice clarity in noisy environments for hearables.

A new method improves speech clarity in loud environments.

A novel method combines meaning and sound for improved emotion detection in speech.

An overview of audio-visual speaker diarization methods, challenges, and systems.

This research analyzes Mamba's performance in speech tasks, emphasizing sound reconstruction and recognition.

SSR-Speech offers new solutions for speech generation and editing.

Researchers develop a dataset to improve speech recognition and analysis techniques.

A study revealing how deep learning models recognize emotions in speech.

A new method improves machine voice recognition for speaker verification.

Study highlights advances in robot emotion recognition using Vision Transformers.

A new framework simplifies speech recognition in busy environments.

A new loss function boosts audio quality by aligning phase and magnitude.

ESPnet-Codec enhances training and evaluation of neural codecs for audio and speech.

A new method improves speech and audio processing across multiple tasks.

This study analyzes how audio, video, and text work together in speech recognition.

Exploring new methods for recognizing emotions in speech using advanced models.

Discover how TDA enhances understanding in language analysis.

A new method aims to detect the origin of synthetic voices.

New methods improve speech separation using neural audio codecs for clearer communication.

New methods improve speech recognition while maintaining past knowledge.

New methods improve how machines recognize spoken language.

Voice cloning technology is advancing, creating lifelike speech that mimics human conversation.

Research explores how speech enhancement models maintain syllable stress amidst noise.

Researchers improve speech processing using Libri2Vox and synthetic data techniques.

A new method improves lip synchrony in dubbed videos for a natural viewing experience.