Search posts, papers, and topics
StepAudio 3 Gen is a general‑purpose audio generation model that replaces diffusion with a discrete autoregressive generator over residual vector quantization tokens. Using a 16‑layer RVQ tokenizer and progressive pretraining, it reaches state‑of‑the‑art zero‑shot TTS and voice‑design performance while handling speech, vocals, sound effects, and music.
A marketing‑focused announcement for Flat.social, a browser‑based 3‑D spatial meeting platform that lets users walk around, talk via spatial audio/video, and run virtual events. The post contains only product descriptions and user praise, with no technical implementation details, architecture, or performance data.
OmniVChat defines native audio‑visual dialogue where a model consumes raw audio and video streams and replies in text. The authors build OmniVChat‑Studio, a multi‑agent simulator that generates single‑ and multi‑turn audio‑visual conversations, and use it to create OmniVChat‑Bench, a benchmark covering five dialogue abilities. They also propose OmniVChat‑RL, a reinforcement‑learning reward that b…
The author records 46 classic DOS game soundtracks on seven MIDI modules (Roland SC‑55, SC‑VA, Yamaha MU80, S‑YXG50, etc.), shares 322 lossless recordings plus REAPER project files, and documents a repeatable workflow for extracting MIDI from DOSBox, cleaning it, and rendering audio on hardware and software synths.
StepAudio 3 Realtime is an audio‑language foundation model that runs a continuous listen‑converse‑think‑act loop. It introduces Deep Perception for rich acoustic cue extraction, Seamless Duplex for handling pauses/back‑channels, and a Think‑While‑Speaking mechanism that lets the model reason in parallel with speech output. On benchmarks it scores 73.0 macro avg on StepAudioChat, 90.6 on MMSU, 98.…
The paper introduces TeleAntiFraud 2.0, a monthly refreshed, frozen audio benchmark of 900 Chinese telecom calls (600 fraud, 300 near‑domain non‑fraud) built with a Mixed‑Tree generation pipeline. Experiments show models that score perfectly on unrelated negatives fall to ~0.66 Macro‑F1 on near‑domain cases, exposing shortcut learning and prediction collapse.
FRAUDSkill is a framework that keeps a pretrained audio‑language model frozen and learns an external layer of skill programs, routing policies, and decision rules to meet a structured anti‑fraud detection protocol. On the TeleAntiFraud benchmark it reaches 73.5% Macro‑F1 (≈32% improvement) while cutting invalid predictions to 1.94%.
