🤖 Audio - Research Papers
Mirelo is now a Power in Kiro, the agentic IDE by AWS. Hand your footage to Kiro and get perfectly synced edit-ready sound effects, all in the same workflow. Try it now.
Key Points:
• Mirelo is integrated into Kiro, an agentic IDE by AWS.
• Kiro provides perfectly synced edit-ready sound effects.
• The integration allows for a seamless workflow.
🔗 Resources:
• Original post ↗ - Original source
• Mirelo AI https://x.com/MireloAI ↗ - Mirelo AI
• Kiro https://x.com/Kiro ↗ - Kiro IDE
🚀 Audio - Research Papers
VAANI Noise Event Dataset: A curated spontaneous speech dataset annotated with timestamps for noise events
Key Points:
• VAANI Noise Event Dataset is a curated spontaneous speech dataset.
• The dataset is annotated with timestamps for noise events.
• The dataset is suitable for research on noise event detection.
🔗 Resources:
• Original post ↗ - Original source
• VAANI Noise Event Dataset https://x.com/ArxivSound ↗ - VAANI Noise Event Dataset
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
Efficient Passive Acoustic Monitoring of Killer Whales Using a Two-Stage Detection and Ecotype Classification Cascade
Key Points:
• The paper proposes a two-stage detection and ecotype classification cascade.
• The method is used for efficient passive acoustic monitoring of killer whales.
• The approach is suitable for real-world applications.
🔗 Resources:
• Original post ↗ - Original source
• Efficient Passive Acoustic Monitoring https://x.com/ArxivSound ↗ - Efficient Passive Acoustic Monitoring
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
Hearing the Whispers: Black-Box Membership Inference Attacks on Finetuned TTS Models
Key Points:
• The paper proposes a black-box membership inference attack on finetuned TTS models.
• The attack is based on the idea of "hearing the whispers" of the model.
• The approach is suitable for evaluating the security of TTS models.
🔗 Resources:
• Original post ↗ - Original source
• Hearing the Whispers https://x.com/ArxivSound ↗ - Hearing the Whispers
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
Multimodal Digital Biomarker for Asthma: Complementary Roles of Vocal, Clinical and Demographic Factors
Key Points:
• The paper proposes a multimodal digital biomarker for asthma.
• The biomarker is based on vocal, clinical, and demographic factors.
• The approach is suitable for evaluating the severity of asthma.
🔗 Resources:
• Original post ↗ - Original source
• Multimodal Digital Biomarker https://x.com/ArxivSound ↗ - Multimodal Digital Biomarker
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
Backdoor Attacks on Speech Emotion Recognition via TTS-Generated Poisoning
Key Points:
• The paper proposes a backdoor attack on speech emotion recognition models.
• The attack is based on TTS-generated poisoning.
• The approach is suitable for evaluating the security of speech emotion recognition models.
🔗 Resources:
• Original post ↗ - Original source
• Backdoor Attacks https://x.com/ArxivSound ↗ - Backdoor Attacks
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
Half-Truth Audio Detection and Localisation: A Lightweight Cross-Attentive Architecture and a Cross-Corpus Diagnostic Study
Key Points:
• The paper proposes a lightweight cross-attentive architecture for half-truth audio detection.
• The approach is suitable for real-world applications.
• The paper also presents a cross-corpus diagnostic study.
🔗 Resources:
• Original post ↗ - Original source
• Half-Truth Audio Detection https://x.com/ArxivSound ↗ - Half-Truth Audio Detection
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
DuoGesture: Motion-Grounded Semantic Conditioning and Biomechanical Beat Priors for Co-Speech Gesture Generation
Key Points:
• The paper proposes a motion-grounded semantic conditioning approach for co-speech gesture generation.
• The approach is based on biomechanical beat priors.
• The method is suitable for generating realistic co-speech gestures.
🔗 Resources:
• Original post ↗ - Original source
• DuoGesture https://x.com/ArxivSound ↗ - DuoGesture
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
Binaural Sound Event Localization and Detection based on HRTF Cues for Humanoid Robots
Key Points:
• The paper proposes a binaural sound event localization and detection approach.
• The approach is based on HRTF cues.
• The method is suitable for humanoid robots.
🔗 Resources:
• Original post ↗ - Original source
• Binaural Sound Event Localization https://x.com/ArxivSound ↗ - Binaural Sound Event Localization
• Arxiv Sound https://x.com/ArxivSound ↗ - Arxiv Sound
🚀 Audio - Research Papers
Fees on pools aren't distributed right away. They are streamed over 24 hours to holders of the token. Every trade in a https:// pools.fun pool pays a 1% fee. Buys pay it in the quote asset (WETH, USDG, or the stock the token is paired with). Sells pay it in the token.
Key Points:
• Fees on pools are streamed over 24 hours to holders of the token.
• The fee is 1% of every trade in a pool.
• The fee is paid in the quote asset for buys and in the token for sells.
🔗 Resources:
• Original post ↗ - Original source
• Pools.fun https://x.com/pools_dot_fun ↗ - Pools.fun
• 0xDeployer https://x.com/0xDeployer ↗ - 0xDeployer
