Latest Articles for Audio Processing

PLCMOS offers a new way to evaluate speech quality without human listeners.

A new method combines speech recognition and speaker identification for overlapping speech.

A new method for voice conversion improves clarity and adaptation.

Explore how diffusion models transform noise into valuable data outputs.

A new model improves voice isolation in noisy environments.

DeCoR helps machines learn new sounds without forgetting old ones.

A new method enhances the naturalness and variety of text-to-speech output.

Treff adapter improves audio classification with limited labeled data.

Research highlights effective methods for recognizing emotions in speech using embeddings.

This research analyzes dialects using audio recordings to reveal their similarities.

A novel method enhances audio classification by learning new sounds efficiently.

A new method aligns disfluent speech with text efficiently.

A new method for training keyword spotting models using weak supervision in noisy environments.

MERT addresses music modeling challenges through innovative self-supervised learning techniques.

AVLIT model combines sound and video for better speech clarity in noisy settings.

Discover how SVVAD improves voice activity detection for better speaker verification.

UnDiff enhances audio quality using innovative speech restoration techniques.

Discover the innovative Multi-Window Masked Autoencoder method for enhanced audio processing.

A novel method merges audio and visual data to repair missing speech.

SingNet improves beat tracking in singing voices using past data.

A fresh look at speaker anonymization and the crucial role of vocoders.

A new method aims to improve fake audio detection without losing past knowledge.

New model LinDiff improves speech synthesis speed and quality.

Techniques to improve speech recognition amidst background noise.

HiddenSinger improves singing voice quality using advanced AI techniques.

New methods improve speech clarity for electrolarynx users.

Recent research improves ASR models for Norwegian, enhancing performance in Bokmål and Nynorsk.

Gesper framework enhances speech clarity in noisy environments.

This article discusses a new method for building efficient ASR systems.

New algorithms enhance audio processing performance across varying sample rates.

A new model improves music transcription accuracy for multiple instruments.

A guide to using AI models for music on the Bela platform.

A new model improves voice conversion by simplifying speech separation techniques.

A new method transforms mono signals into engaging stereo experiences.

A new system enhances detection of manipulated audio through innovative techniques.

LyricWhiz combines advanced models to improve lyric transcription accuracy across languages.

This article discusses challenges and techniques for managing dataset imbalance in audio classification.

Whisper-AT combines speech recognition and audio tagging for improved performance.

A new method enhances speaker identification in film and TV localization.

New method improves accuracy in turning piano audio into sheet music.