Research focuses on detecting deepfake audio through improved techniques and data expansion.
― 5 min read
Computer Science - Sound
RSSA new approach focuses on subtle inconsistencies in deepfake detection.
― 6 min read
Examining how utterance length and social factors influence speech rate.
― 5 min read
Introducing PeriodWave, a model improving audio generation speed and quality.
― 5 min read
Learn how to prepare and submit your scientific paper effectively.
― 7 min read
A look at how sound characteristics in popular music have changed over decades.
― 4 min read
A new system improves guitar tablature creation using deep learning methods.
― 5 min read
A new system enhances speech recognition by using contextual keywords for better accuracy.
― 5 min read
PeriodWave-Turbo improves sound generation speed and quality across various applications.
― 5 min read
Research reveals how to make speech models smaller and more efficient.
― 5 min read
Dialogue separation helps viewers hear conversations clearly amidst background noise.
― 6 min read
MAT-SED uses a novel Transformer model for effective sound event detection.
― 5 min read
Combining heart sounds and echocardiography to improve diagnosing congenital heart disease.
― 5 min read
A rich dataset of guitar recordings linked to music scores for research and analysis.
― 4 min read
Auptimize enhances audio cue placement for better user interaction in XR.
― 6 min read
Malacopula challenges the reliability of automatic speaker verification technologies.
― 6 min read
A new method for more realistic 3D face animations adjusting to personal speaking styles.
― 5 min read
Adversarial training enhances keyword spotting accuracy in synthetic and real speech.
― 5 min read
This piece discusses few-shot learning and its impact on audio tasks.
― 6 min read
New technology links facial features to voice, aiding communication for those without a voice.
― 5 min read
A new method enhances audio separation and generation without labeled data.
― 6 min read
Addressing the challenges of fake audio and speaker verification.
― 5 min read
Analyzing rage music features through machine learning for better genre classification.
― 5 min read
Fake audio clips are a serious concern; effective detection methods are essential.
― 6 min read
A new method improves the accuracy of detecting synthetic audio.
― 5 min read
A new method for separating and manipulating musical sounds.
― 5 min read
SSL-TTS simplifies voice synthesis using minimal training data for high-quality results.
― 6 min read
New methods enhance ASR models for multiple languages, preserving past knowledge.
― 5 min read
A new approach enhances recognition of code-switched phrases in bilingual speech.
― 5 min read
An innovative system automates sound generation for films and games.
― 8 min read
New methods improve speaker recognition in noisy environments.
― 5 min read
New model improves voice conversion, especially for whispered speech and real-time applications.
― 6 min read
Exploring a new digital approach to guitar amplifier sound modeling.
― 5 min read
Introducing a groundbreaking system to generate Hindustani vocal music.
― 6 min read
A new method for accurately modeling optical compressors using neural networks.
― 7 min read
WhisperMask captures voice clearly in noisy places, enhancing communication.
― 6 min read
New methods improve voice quality assessments for patients with vocal system issues.
― 6 min read
VoiceX simplifies the process of creating personalized voices for various applications.
― 4 min read
Examining how voice patterns affect meaning and technology performance.
― 4 min read
NEST offers a faster, more efficient approach to self-supervised speech tasks.
― 5 min read