A basketball player leaps for a dunk in a crowded arena

I have been meaning to do POV videos and thought why not come up with a modular prompt for these types of shots. Here is an attempt: 

2x2 grid, 16:9, ai inferred subjects: input = [athlete / sport_or_activity / action_beat / venue / camera_source] run extreme_pov_sports_extraction: sport_dna := infer rules, gear, surface, movement style, and athlete role athlete_dna := infer body type, uniform, equipment, posture, effort, and emotional intensity action_state := infer exact frozen moment: launch, strike, landing, sprint, collision, dodge, slide, swing, flip, or follow-through camera_position := infer ground-level, object-mounted, helmet-mounted, court-edge, underwater, snow-level, wheel-level, or player-facing pov lens_behavior := infer fisheye distortion, wide-angle stretch, close-focus blur, lens flare, dirt, water drops, grass, snow, dust, or scratches foreground_exaggeration := infer closest limb, equipment, wheel, board, glove, racket, hoof, fin, machine part, or object venue_geometry := infer stadium ring, court lines, arena lights, street barriers, ocean horizon, mountain slope, track curve, or gym ceiling surface_physics := infer turf compression, dust spray, court reflection, ice scrape, water splash, sand kickup, snow burst, or pavement grit lighting := infer sun glare, floodlights, backlight, harsh noon, neon, indoor spotlight, or natural overcast motion_physics := infer blur direction, body momentum, particles, shadow stretch, and speed cues render: a dramatic extreme pov action frame where the viewer feels physically inside the sport. forbidden: hard-coded sport, hard-coded cleat, hard-coded stadium, hard-coded grass, hard-coded ball, perfect clean framing, static pose, generic sports poster.