PHOENIX STORY WORKS — B-ROLL PROMPT BUILDER ==================================== PROMPT ------ Act as a resourceful video director planning B-roll for a small production. The main video is about [TOPIC]. It is being filmed in [LOCATION] with access to [PEOPLE / OBJECTS / TOOLS AVAILABLE]. The style should feel [STYLE], and the video will be used on [PLATFORM]. Suggest 20 useful B-roll shots grouped into: 1. Establishing shots that show place and context. 2. Action shots that show the process. 3. Detail shots that make the video feel tactile. 4. Human reaction shots that add emotion. 5. Transition shots that help connect scenes. For each suggestion, include: - Shot size - Subject and action - Simple camera movement - Best angle or lighting note - What sentence or idea the shot could cover Prioritise shots that one person can film with a phone. Avoid repeating the same angle. End with a 10-shot “must capture” checklist for a rushed filming day. AI VIDEO GENERATOR ADD-ON ------------------------- Convert the best 5 ideas into generation prompts. Each prompt should state the subject, action, location, lighting, camera angle, camera movement, mood, and realistic visual style. Do not include text or logos in the generated scene. B-ROLL STRATEGY WORKSHEET ------------------------- Main story or message: __________________________________ Key spoken ideas that need visual coverage: ______________ Available location(s): ___________________________________ Available people: ________________________________________ Objects, tools, or products: ______________________________ Time of day and light: ___________________________________ Target platform and aspect ratio: _________________________ Visual mood: _____________________________________________ Footage that already exists: ______________________________ Footage or imagery that must not be used: _________________ WHAT GOOD B-ROLL DOES --------------------- Useful B-roll should perform at least one job: ORIENT — Show where, when, or with whom the story takes place. EXPLAIN — Make a process or abstract statement visible. PROVE — Support a claim with a real detail, result, screen, or demonstration. FEEL — Add human reaction, atmosphere, texture, or consequence. CONNECT — Cover an edit or move the viewer into the next idea. PACE — Create variation after a long presenter or screen-recording section. If a shot performs no clear job, it may be decoration rather than useful coverage. THE FIVE-SHOT COVERAGE PATTERN ------------------------------ For each important action, capture: 1. A wide shot of the whole action. 2. A close-up of the hands or most important detail. 3. A face or reaction shot. 4. A point-of-view or over-the-shoulder shot. 5. A creative detail that communicates mood or transition. Hold each shot steady for at least five seconds before and after the action. Film important actions more than once from different sizes rather than moving the camera during every take. AI VIDEO PROMPT ANATOMY ----------------------- A controllable generation prompt normally includes: SUBJECT — Who or what is the focus? Be specific without naming a real person. ACTION — What single visible action happens during the clip? SETTING — Where and when does it happen? COMPOSITION — Wide, medium, close-up, overhead, point of view, or another angle. CAMERA — Static, slow push, gentle track, or another motivated move. LIGHT — Soft window light, overcast daylight, warm practical lamps, and so on. MOOD — Calm, focused, playful, tense, reflective, energetic. STYLE — Realistic documentary, polished commercial, handmade stop motion, etc. CONTINUITY — Clothing, object placement, colours, direction, and time of day. CONSTRAINTS — No text, logos, extra fingers, sudden object changes, or other known problems relevant to the tool. MASTER GENERATION TEMPLATE -------------------------- Create a [DURATION]-second [SHOT SIZE] video of [SUBJECT] [CLEAR ACTION] in [SETTING AND TIME]. The camera [CAMERA POSITION AND MOVEMENT]. Use [LIGHTING] with a [COLOUR / MOOD] atmosphere. The visual style is [STYLE], with believable materials, natural motion, and realistic physical interaction. Keep [CONTINUITY DETAILS] consistent throughout the clip. Compose for [ASPECT RATIO], leaving clear space on [SIDE] for optional captions. Do not include [UNWANTED ELEMENTS]. CONTINUITY BLOCK FOR A SERIES OF GENERATED SHOTS ------------------------------------------------ Copy the same block into every related prompt: "Continuity: the subject wears [CLOTHING]; the main object is [DESCRIPTION]; the room has [KEY FEATURES]; light enters from [DIRECTION]; the palette is [COLOURS]; all movement follows left-to-right screen direction; realistic proportions and consistent object placement." Generate simple shots separately. Trying to create multiple camera angles, locations, or story beats in one short clip often reduces consistency. WORKED EXAMPLE: HOME BAKERY STORY --------------------------------- Voiceover idea: "Every loaf starts the night before, when flour, water, and time begin doing the quiet work." Real footage list: 1. Wide, static kitchen before sunrise; practical lights switching on. 2. Extreme close-up of bubbles in fermented dough. 3. Over-the-shoulder shot of hands folding the dough. 4. Detail of flour falling across a wooden worktop. 5. Medium side shot of shaping one loaf. 6. Close shot of the blade scoring the surface. 7. Oven light coming on; tray moving into frame. 8. Baker's reaction while listening to the crust crackle. 9. Wide shot of finished loaves on the cooling rack. 10. Close-up of one slice revealing the interior. Example AI generation prompt for an unavailable establishing shot: "Six-second wide documentary-style video of a small independent bakery kitchen just before sunrise. A baker in a dark green apron switches on warm work lights and crosses towards a wooden preparation table. Static camera at counter height, soft blue dawn visible through one rear window, warm practical light inside, quiet and focused mood, realistic materials and natural human movement, 16:9. No visible brand names, labels, captions, or camera shake." EDITING MAP ----------- Before filming, paste the voiceover into a table and assign coverage: VOICEOVER LINE | PRIMARY SHOT | BACKUP SHOT | NATURAL SOUND | STATUS Aim for at least two usable visual options over every edit point. Do not cover an entire section with random footage that has no relationship to the spoken idea. NATURAL-SOUND CHECKLIST ----------------------- [ ] Doors, footsteps, fabric, tools, keyboards, packaging, water, or machinery. [ ] Ten seconds of clean ambience at each location. [ ] A close microphone take of important small sounds. [ ] Permission to record conversations or recognisable people. [ ] Reduced background music during sounds that carry story information. FOLLOW-UP PROMPTS ----------------- - "Match every B-roll suggestion to the exact voiceover sentence it supports." - "Remove shots that are decorative or repeat the same information." - "Create an indoor backup for every weather-dependent shot." - "Convert the list into a one-person smartphone filming plan." - "Group the shot list by location, subject, and lighting setup." - "Write separate AI generation prompts for the five shots we cannot film." - "Create a continuity block shared by all generated clips." - "Identify shots that require consent, licences, safety preparation, or proof." FINAL REVIEW ------------ [ ] Every major spoken idea has relevant visual coverage. [ ] Shot sizes and angles vary for a reason. [ ] The list includes context, action, detail, reaction, and transitions. [ ] Generated shots do not imply false evidence or real events that never happened. [ ] Real people, brands, property, artwork, and music are used with permission. [ ] Continuity details are repeated across related generation prompts. [ ] Important clips have simpler backup options. [ ] The editor can identify where each shot belongs. Phoenix Story Works — Better prompts. Better videos.