I have a library of Full-HD (1080p) TV-show episodes that I’d like to watch in real-time on my 1050p HDR display without losing detail or color accuracy. I need a proven deep-learning approach that can: • take the 1080p source frames, • upscale and HDR-tone-map them on the fly, • output a clean 1050p signal with minimal latency. Whether you lean on Topaz Video AI, Real-ESRGAN, an mpv/VapourSynth chain, TensorRT, or a custom PyTorch model is up to you—as long as the end result is a plug-and-play workflow I can launch before each viewing session. Deliverables - A working script or real-time player configuration (Windows or Linux) that performs the deep-learning upscaling and HDR conversion. - Clear, step-by-step setup instructions so I can replicate the environment, install the required models, and update them later. - A short demonstration clip or screen capture proving the pipeline maintains smooth playback and visibly sharper, more vibrant output. If you’ve engineered live upscaling solutions before, I’d like to see a quick sample or description of what you achieved. Lift my TV-show nights to the next visual level, and let’s get started.