Combining audio and visual signals enhances speech recognition in challenging environments.
― 4 min read
Computer Science - Sound
RSSLatest Articles
Latest Articles
A new model allows musicians to control sound synthesis more effectively.
― 5 min read
Combining audio and visual data for better keyword detection in voice assistants.
― 5 min read
New methods reveal how speech can indicate depression severity.
― 6 min read
New method improves machine learning for audio tasks while retaining prior knowledge.
― 5 min read
A new framework improves multilingual ASR by merging language-specific features with efficiency.
― 5 min read
New methods improve the accuracy of voice-based identity checks.
― 6 min read
ClArTTS database enhances Arabic TTS with quality recordings.
― 5 min read
A new method improves audio matching for design documents using a unique dataset.
― 5 min read
The 2022 NIST evaluation focused on language recognition advancements, particularly for African languages.
― 5 min read
New deHuBERT model enhances speech recognition accuracy in challenging noise conditions.
― 4 min read
ParrotTTS revolutionizes speech generation with less transcribed data.
― 6 min read
A new system enhances transcription of lengthy audio recordings with improved accuracy.
― 5 min read
Introducing READ Avatars for lifelike emotional expression in digital characters.
― 5 min read
SpeechPrompt v2 enhances speech classification with efficient techniques and improved accuracy.
― 5 min read
audb simplifies handling and sharing audio datasets efficiently.
― 5 min read
This study enhances speech recognition through ensemble knowledge distillation and elitist sampling.
― 5 min read
New method improves speaker verification accuracy from far-field recordings.
― 6 min read
End-to-end models simplify speech recognition, improving accuracy and efficiency.
― 6 min read
New techniques enhance speech processing efficiency with fewer resources and better performance.
― 5 min read
LooperGP aids musicians in generating customizable loops for live performances.
― 5 min read
New methods improve emotional depth in TTS, enhancing user interactions.
― 5 min read
Self-distillation boosts detection systems against fake speech technologies.
― 5 min read
New techniques improve detection of fake voices in voice recognition systems.
― 5 min read
Innovative techniques reduce model size while maintaining performance in speaker verification.
― 5 min read
New insights into identifying emotions in speech using sound and word data.
― 5 min read
A study on capturing emotions in music through pianist performances.
― 4 min read
Improvements in TTS technology enhance personalization and speech quality.
― 5 min read
New models improve efficiency for mobile voice assistants.
― 6 min read
ProVE enhances procedural audio generation, improving sound quality and user control.
― 6 min read
A new method improves speaker recognition by combining time and frequency features.
― 5 min read
A new method enhances machine understanding of speech and text connections.
― 6 min read
This article explores the latest methods for audio representation and their implications.
― 5 min read
FoundationTTS improves naturalness and diversity in speech synthesis.
― 4 min read
New techniques for keyword spotting using small models and self-supervised learning.
― 6 min read
New method enhances sound estimation across environments using adaptive techniques.
― 5 min read
This study presents a fast method for audio data labeling and classification.
― 5 min read
Learn how images can be concealed within audio using advanced techniques.
― 5 min read
New models improve the efficiency and accuracy of piano transcription.
― 5 min read
New dataset tackles real-world challenges in active speaker detection technology.
― 5 min read
A new metric enhances ASR performance evaluation for medical transcription accuracy.
― 5 min read