I have a standard-definition live broadcast feed that has to be converted to crisp 1080p in real time and then pushed out live. Latency must stay low enough for smooth, interactive streaming, and picture quality should be visibly sharper than basic bilinear or bicubic scaling. The core deliverable is a working upscaling workflow—hardware, software or a hybrid—that ingests my SD signal, upscales it cleanly to Full HD, and outputs an RTMP (or similar) stream I can point to the usual live destinations such as YouTube, Twitch or Facebook Live. If you prefer another protocol that keeps delay minimal, I’m open. I am comfortable running Linux or Windows boxes and already have a decent NVIDIA GPU available, but I’ll take your guidance if a different card or an FPGA/ASIC approach will produce better real-time results. Please provide: • the upscaling method or model you plan to use (e.g., VapourSynth with AI models, FFmpeg with CUDA filters, TensorRT, Real-ESRGAN, Topaz—whatever you trust for live performance) • step-by-step setup instructions and any required scripts/config files • recommended bitrate and encoder settings for stable Full HD streaming at 30 or 60 fps • a quick remote test session so we can verify latency and picture quality on a live signal I’ll supply a sample SD feed for you to tune against. Once we confirm the pipeline works end-to-end, the project is done.