Technical Salon Summary: Mobile Live Streaming, P2P Acceleration, Overseas Live Streaming Architecture, RTC‑PSTN Integration, and Video AI Practices

The event, held on June 29 by Cloud+ Community, gathered nearly 300 developers to discuss the latest advances in audio‑video and integrated communication technologies. Tencent Cloud presented a series of talks covering practical solutions, best practices, and architectural designs.

1. Mobile Live Streaming with Interactive (Co‑hosting) Technology – The speaker described challenges such as unstable networks and high‑quality requirements for co‑hosting. Solutions included using UDP‑based upload nodes, jitter buffers for smoothing, and server‑side echo cancellation (AEC) integrated in LiteAVSDK. Cloud‑side mixing supports up to 16 streams, and the MLVBLiveRoom component combines LiteAVSDK, IMSDK, and AI‑enhanced effects for rapid integration.

2. Tencent Cloud Live PCDN Acceleration – XP2P technology leverages P2P nodes to reduce CDN load and latency. It uses UDP for low‑delay transport, supports seamless segment slicing without altering codec metadata, and employs NAT traversal (STUN) and pacing mechanisms (XNTP) to improve bandwidth utilization.

3. Overseas Live Streaming Architecture – A decentralized design deploys source nodes across global regions (North America, APAC, Europe) to minimize latency. The system uses dual‑active state synchronization, tiered ISP classification, and adaptive routing to ensure stable cross‑region streaming, with DRM support for Widevine and FairPlay.

4. Real‑time Audio‑Video and PSTN Integration – The talk covered differences between RTC (UDP‑based, sub‑second latency) and PSTN (SIP/RTP, G.711/G.729). A middle‑layer adapts signaling and media streams, handling echo cancellation, jitter buffering, and adaptive bitrate control. Server‑side mixing enables multi‑party audio for mobile clients.

5. Video AI Applications – Tencent’s Zhimo platform provides AI‑driven services such as intelligent recognition (faces, logos, OCR), content moderation, and automated editing. Use cases include high‑efficiency compression (speedy‑HD), cloud‑based video editing, and automated highlight extraction using models like R‑C3D.

The salon concluded with a lively Q&A, showcasing strong developer interest and emphasizing the growing impact of audio‑video and communication technologies in everyday applications.