The Infinite Toolbox, redone

Make me a short film for my product the same way zippr made one particular film: "The Infinite Toolbox, redone" (zippr.to). In it, a person drawn in a few ink lines asks for four ordinary jobs, and each one pulls more tools out of a toolbox: 4, then 7, then 11, then 19, until the room is a warehouse of racks and crates. Then they ask for the first job again, and the same four tools start to come out again, until everything freezes: "Or…". In the second room, each job they have done once is one small ink box with one red-orange button: four asks, four clicks, four bells, and the warehouse is left behind as faint dotted lines while a small blue mascot jumps for joy. It is drawn in thin ink on cream paper, with one red-orange accent used only on the buttons; the mascot is the only solid, lit thing. The sound counts the tools: one clank per tool, faster and faster. Below is the exact recipe zippr used: the tools, the steps, the settings and the timings. Follow it as closely as you can. Only the parts about my product change.

WHAT IT COSTS

- The code tools in this recipe are free; all but GSAP are open source, and GSAP is free under its own no-charge licence. - zippr made it with Claude (Opus 5.5) on a paid Claude plan. Any AI app that can run code can follow it. - The idea came from a chat with ChatGPT; no cost for that chat is recorded. - One step was paid: zippr's mascot was drawn once with Google's Nano Banana Pro image model, about $0.14 a picture: a standing picture, and a jump edited from it, the two this film shows (about $0.28). You can skip it: use a picture you already have, or have the hero drawn in code for free. Everything else in the film is ink lines drawn in code.

STEP 1: ASK ME FIRST

Before you make anything, ask me these questions. Use your question tool if you have one. Otherwise ask them one at a time in the chat. If I skip one, suggest an answer and tell me what you picked. 1. What is your product or business, in a sentence or two? 2. What is the one thing a viewer should take away? (zippr's: "Teach it once." then "The second time, just ask.") 3. Who is going to watch it, and where: reels, a website, a talk? 4. Length and shape. zippr's film is 28.0 seconds, portrait 1080x1920, on a grid of 90 beats a minute (a beat is 20 frames). Keep 28 s? Portrait only, or a landscape cut too (zippr made portrait only)? 5. Your brand colours (hex codes if you have them), your product's name as the wordmark, and the web address under it. 6. The words on screen. zippr's: the four asks ("Combine these spreadsheets.", "Merge these PDFs.", "Resize these photos.", "Turn these photos into a video."); the counts ("4 tools", "7 tools", "11 tools", "19 tools", counting up as each lands); "4 tools. Again."; "Or…"; the end lines "Teach it once." and "The second time, just ask."; the wordmark "zippr" and "zippr.to". Which words are yours? 7. Your story in zippr's shape: - four everyday jobs your customer asks for, each bigger than the last, ideally jobs your product has really done (zippr's four are jobs it had run); - what each one drags out (zippr's: 4, 7, 11 and 19 hand tools, a joke count, not a measurement); - the first job asked again, and the same tools starting again, until the freeze; - the turn (zippr's: "Or…"); - your version: each job done once becomes one small box with one button (zippr's: a small ink box with a red-orange button; the output pops out of it like toast). The button comes on the SECOND ask of a job, not on a brand-new one: only claim what your product really does. 8. The hero's picture: (a) you have one: a mascot, product photo or logo, ideally 2 poses (standing facing us, jumping for joy) as PNGs with transparent backgrounds; (b) you want one generated like zippr's with Nano Banana Pro (paid, about $0.14 a picture; you run it, or give me access); (c) I draw it in code (free). 9. The ink and paper. zippr's: near-black ink on cream paper, a muted grey for the second-strongest words, and ONE red-orange accent on the buttons only. Put your brand colours on the colour roles in the recipe below, or keep zippr's?

STEP 2: THE RECIPE (exactly how zippr made this film)

Where the method came from The idea is ChatGPT's, from a chat, word for word: "Start with a tiny toolbox. Every time the person wants something, another tool comes out. Email requires 4 tools. Meme requires 7. Deck requires 11. Video requires 19. The toolbox eventually becomes a giant ridiculous warehouse. Then: 'Or…' Zippr. One request. Done." zippr changed two things on purpose. Read literally, "One request. Done." for a 19-tool video says the product does any new job in one go, which zippr cannot claim; so the button arrives on the second ask of a job, which is zippr's actual promise. The film was made storyboard first: six frames of the finished look (each with what is on screen, what moves next and what is heard), a full-length sound sketch, and an animatic, approved before a single frame of the film was built. The look is the thin-ink-on-cream paper world from an earlier test in which Claude directed the art: tools are line drawings that can multiply for free, the warehouse can be hatched heavy, and a known job is a small solid ink box with one red-orange button. It was approved to be built. Its sound was picked blind from two options.

The tools - ChatGPT: the idea, the counts 4, 7, 11 and 19, and "Or…". - Claude (Opus 5.5): wrote the storyboard, the sound sketch and all the code, built the film and checked it. - A web page: every tool, the toolbox, the table, the window, the person, the racks, the boxes and the dotted "ghost" warehouse are SVG ink lines built in JavaScript; the words are HTML type; the hero is two cut-out pictures. GSAP 3.13.0 (the animation library, under GSAP's standard no-charge licence) supplies only its easing curves; every frame is drawn by seek(t) from one plan file. - Python: one small script writes the plan file (every frame of every event), which the picture and the sound both read. - Playwright for Python driving headless Chromium: opens the page, calls seek(t) for each frame and screenshots it. - Python with numpy: makes every sound from scratch, at a fixed seed (1397). There are no recordings and no voice. - ffmpeg with libx264 and AAC: joins the frames and the sound, and measures loudness. - The font: Inter (weights 500 to 800), bundled as a file. - Paid, once: Nano Banana Pro (Google's gemini-3-pro-image, through Vertex AI, at 2K) drew the mascot's poses, about $0.14 each by usage.

The steps, in order 1. The storyboard: the film in one line, the style in one line, six frames rendered by the film's own page (0:04 the first job, 0:11 the third, 0:14 the warehouse, 0:17 the same job again, 0:20 the button, 0:27 the end card), each with what is on screen, what moves next and what is heard; the story changes and the source of every line. Then the sound sketch and an animatic under it. Wait for a yes. 2. The plan file: 30 fps and 90 beats a minute, so a beat is 20 frames and a bar 80. The tool rate doubles each job (1, 2, 4, then 8 per beat). Every event (every tool's landing frame, every bell, click and tick) is written once, and nothing else holds a time. 3. The page: seek(t) takes the frame and draws the whole scene from the plan alone. 4. Render: Playwright screenshots each of the 840 frames; ffmpeg joins them. 5. The sound: numpy builds each sound at the plan's frames. Mix, set the loudness. Then move the whole finished soundtrack 2 frames earlier (below). 6. Measure the finished file: every hit's attack against its picture frame, every picture event against its sound's frame, the loudness by section. Count what passes, what fails and what could not be told, out loud. 7. Check the finished file (below), then export.

The look (every value is zippr's) - Paper #EDE6D3, ink #1D1D1D, every line a round-capped 3.2 px stroke; shapes filled with paper so lines behind them are hidden. A fine paper grain (SVG fractal noise, base frequency 0.9, 2 octaves, fixed seed, a warm dark at 6%) over the whole frame. Heavy things (the warehouse crates) are hatched: diagonal ink lines 14 px apart at 45 degrees, 2 px, at 55%. - The room (world pixels, where camera scale 1 is the 1080x1920 frame): the floor line at y 1700; a table top 660 x 20 px at x 320, y 1380, on two legs; a window 250 x 240 at (560, 700) with a cross; the person, a few lines on a stool at the right (head a 48 px circle at (908, 1106), one eye, one eyebrow, one mouth line, an arm on the table). The toolbox on the floor at x 340: a box 180 x 92 with a latch, a lid 188 x 20 with a handle. - Fourteen hand tools, each drawn upright about 70 wide and 140 tall: screwdriver, hammer, wrench, saw, ruler, pliers, brush, magnifier, scissors, tape measure, pencil, clamp, spirit level, paint roller. They are handed out in that order, round and round. - Where the tools land: job 1 stands its four tools in a row on the table (x 400, 488, 576, 664), tilted up to 7 degrees; job 2 hangs seven on a pegboard behind the table (660 x 190 at (140, 960)), 85% size; job 3 hangs eleven on a second, higher board (990 x 240 at (50, 450)), 95%; job 4 spreads nineteen across the warehouse racks and a high shelf, 125% to 130%, tilted up to 25 degrees. - How a tool flies: from the toolbox's mouth in an arc 170 px above its landing spot, spinning 300 degrees and growing from 35% to full size, over 8 to 16 frames, landing exactly on its frame. - The toolbox grows with each job (1.0, then 1.5, 2.2 and 5.0 times, each with a springy overshoot over 14 frames); its lid swings open over 3 frames and lands on the ask's send frame, and swings shut over 3 frames landing on the job's done frame. Each job's output (a spreadsheet, a stack of PDFs, two photos, a film strip) drops 120 px onto the table over 6 frames on its done frame. - The warehouse draws itself in on job 4: two tall racks with shelves, hatched crates, a high beam with more crates, a ladder; every line is revealed along its length over 16 frames, staggered. - The camera scales about the point (540, 1560): 1.15 at the start, 1.0 for job 2, 0.72 for job 3, 0.40 for the warehouse, back to 1.0 for the repeat (power2 in and out between keys), then 1.25 in the button room (moved 21 px left and 80 px down, as tight as it goes with the hero and the person both whole), and 0.75 for the end. - The button room: the same table, window and person; no toolbox, no tools. Each job is a solid ink box 100 x 72 px with 10 px corners and a thin paper line across its top, at x 400, 520, 640 and 760 on the table, with a red-orange button (32 x 18 px, 6 px corners, an ink outline) on top. A box springs in from 60% size over 6 frames; its button goes down 6 px for 6 frames on its click; its output pops up 80 px out of the box over 8 frames, starting 2 frames after the click. - The ghost: at the end the whole warehouse, every rack, crate and tool, and the giant toolbox fade back in to 50% as faint dotted lines (a dash of 3, gap of 9 to 11): what did not need to happen. - The person: blinks six times (3 frames each); the eyebrow goes up 9 px on the repeat; smiles once the hero is delighted. Nothing else moves. - The words: a cream band at the top of the frame (solid to y 470, fading out by 600) keeps them on clean paper. The ask, Inter 500 at 58 px, ink, at (96, 176), typed one letter at a time with a 4 px muted caret blinking every 12 frames; the count under it, Inter 800 at 116 px, letters -0.035em, at (92, 268) ("Again." in weight 700, muted). The end lines, Inter 700 at 78 px, -0.03em: "Teach it once." muted at y 190, "The second time, just ask." ink, weight 800, at y 284. The wordmark "zippr", Inter 800 at 120 px, -0.05em, centred at y 1712, with "zippr.to" in Inter 500 at 40 px, muted, under it. Each fades in over 8 frames, already showing on its sound's frame. - The colours are ROLES. My brand colours go onto the roles; zippr's value stays for any role my colours don't fill. - THE PAPER, the ground and every filled shape: cream #EDE6D3. Mine: my lightest warm neutral. - THE INK, every line, the hatching, the boxes and the strongest words: #1D1D1D. Mine: my darkest colour. It must read as thin lines on the paper at phone size. - MUTED, the first end line, "Again." and the caret: #5E5850. Mine: a softer tone of my ink. - THE ONE ACCENT, the buttons: red-orange #E4572E. Mine: my brand's accent. Nothing else in the frame may use it: it marks the job done once. - THE HERO'S GLOW on the paper: its screen's warm white #FEF1DF, a soft circle 420 px across, 55% at its centre, behind its screen. Mine: my hero's own light colour, or none.

The hero's pictures (the one paid step; follow this only if I chose (b)) - The model: Nano Banana Pro (Google's gemini-3-pro-image), through Vertex AI, at 2K, 16:9. It cost about $0.14 a picture by usage; Google's price page said $0.134 for a 1K or 2K picture. - What kept the character identical in 33 of 34 pictures: - one locked description, pasted word for word into every prompt; - the character's clean reference picture passed in with every call; - each new pose made as an EDIT of an existing picture, with one change per edit. - zippr's locked description: "a small chunky soft-vinyl toy creature: one rounded bean-shaped piece for head and body in matte cornflower-blue vinyl (#5387D8), with a pale periwinkle (#9DACD0) oval belly patch, two short rounded stub arms and two short round stub feet in the body's blue. The upper front of the bean is a large, softly rounded-rectangle screen set into the body, glowing a soft warm white (#FBE5CA). On that screen, and only on it, sit two small round glossy black eyes, each in a thin clear glossy rim and each carrying one round white catchlight. Two soft floppy dog ears in warm cream vinyl (#E6C28E) hang from the top of the head, one on each side. No mouth, no nose, no neck, and no marks, letters or numbers on the body." - The two pose prompts. Each ends with: "The face-screen shows only its two eyes: no mouth, no nose, no symbol, no logo. No words, letters or numbers anywhere in the picture." 1. Standing, an edit of the character's first picture: "Edit this image. Keep the identical character: its shape, proportions, pose, colours, matte soft vinyl, face-screen glow, eyes and catchlights, and the same camera angle and framing. Only change: the setting. [The old background] is replaced by a plain, seamless, evenly lit light-grey studio backdrop that curves smoothly into the floor: one smooth neutral grey (about #C8C8C8), with no texture, no props, no doodles and no horizon line. The character stands on it with only a soft contact shadow under its feet. Soft, even studio light from the upper left, so its colours read true." 2. The jump, an edit of picture 1 with the reference passed in: "The first image is the picture to edit. The second image shows a character, for reference only; copy its exact design, proportions, colours and material. Edit the first image. Keep the identical character (exactly as in the second image: its shape, proportions, colours and matte soft vinyl), the plain light-grey studio backdrop, the light and the framing; only change: its pose and expression, to a delighted jump. It is in mid-air, both feet well off the floor, a small happy hop; both stub arms thrown up high; both floppy ears flung up and outward by the jump. It faces the camera. The face-screen glows a brighter lime green, almost white-hot at its centre; the two eyes open wide and round, each with two big round catchlights, bright and glossy. A soft shadow on the floor below it, separate from its feet." - Then, free, in code: - cut each picture out of its grey backdrop (a model of the grey per row, plus a colour gate); - measure each pose once: its feet centre and its screen box (the glow sits behind the screen); - size the two poses so the hero keeps the same body area as the one it replaced, and anchor them by the figure's centre so the ears stay about 20 px inside the frame edge under the camera's zoom; - keep the jump's pale lime screen as it came. - Designing the character in the first place took zippr several earlier rounds of pictures. Those are not in the price above. - zippr swapped this blue mascot into the finished film afterwards: the page loads the hero from two picture files, so the files were changed and the frames re-rendered, with the sound untouched. - In the button room the hero stands on the floor at x 262, where the toolbox stood (standing about 292 px tall, the jump about 333), with a soft floor shadow (an ellipse 112 x 15 px, ink at 22%, blurred) that shrinks as it hops.

zippr's story, beat by beat (28.0 s at 90 beats a minute; frame = seconds x 30) - 0.13 to 1.07 s: "Combine these spreadsheets." types in. 1.33 s (frame 40): sent; the toolbox lid clacks open. - 2.00, 2.67, 3.33 and 4.00 s (frames 60, 80, 100, 120): four tools fly out and land upright on the table, one per beat; the count reads "1 tool" to "4 tools". 4.67 s (frame 140): the spreadsheet drops onto the table, the lid shuts, the job's bell (E). - 4.87 to 5.73 s: "Merge these PDFs." 6.00 s (frame 180): sent; the toolbox swells to a chest, the pegboard appears, and the camera backs off to 1.0. 6.33 to 8.33 s: seven tools, two per beat (every 10 frames). 8.67 s (frame 260): the PDFs, the bell (G). - 9.33 s (frame 280): "Resize these photos." sent; the chest swells, the second board, the camera backs off to 0.72. Eleven tools, four per beat (every 5 frames), from frame 285. 11.33 s (frame 340): the photos, the bell (A). - 12.00 s (frame 360): "Turn these photos into a video." sent; the toolbox swells to 5 times, a low boom, and the camera pulls right back to 0.40 as the warehouse racks draw themselves in; nineteen tools, eight per beat (every 2 or 3 frames), from frame 363. 14.00 s (frame 420): the film strip lands on the big bell (C). The warehouse holds a beat, then the camera pushes back in to the table (14.67 to 15.93 s). - 14.67 to 15.73 s: the very first ask types again. 16.00 s (frame 480): sent; the giant toolbox's lid jolts up 12 px and lands; "4 tools. Again."; the person's eyebrow goes up. 16.67 and 17.33 s (frames 500 and 520): the first four tools lift 78 px, tilting 15 degrees, and slam down again in the same order, at the same pace. - 17.67 s (frame 530): FREEZE, with the third tool in the air. Room tone only. - 18.00 s (frame 540): the ask line becomes "Or…". - 18.67 s (frame 560): hard cut to the button room, on the bar. The first box is already there. - 18.80 to 19.73 s: "Combine these spreadsheets." types. 20.00 s (frame 600): the button goes down by itself, and the spreadsheet pops out of the box. 20.33 s: that job's bell (E) alone; the hero hops 10 px. - 20.67, 21.33 and 22.00 s (frames 620, 640 and 660): the other three asks, each a new box, a click and its output; each bell half a beat later (G, A, C), the same four notes that ended the four jobs, now a small tune on their own. - 22.67 s (frame 680): the hero switches to its jump, rising 18 px over 4 frames. Short arms raised beside the head read as ears: throw them out to the sides, clear of the head's outline. The camera pulls back to 0.75 (to 23.87 s), and the whole warehouse fades back in as dotted lines. The person smiles. - 24.00 s (frame 720): "Teach it once." 24.80 s (frame 744): "The second time, just ask." 25.60 s (frame 768): "zippr" and "zippr.to". Hold to 28.0 s.

Fixes zippr made between the storyboard and the film, which you should keep - The toolbox stands far enough from the left edge that its open lid never crosses it (the lid opens to 92 degrees, not 100). - The lid swings over 3 frames and LANDS on its clack's frame; before, it flipped open 4 frames early with no motion. - The repeat's hop is 78 px, not 150, so the hopping tool clears the pegboard. - Every picture event with a sound has something to see on that frame, and every event you can see has a sound.

The sound (no music; the tools' own sounds are the count) - Made in numpy at 48 kHz, in stereo, with a fixed random seed (1397). The idea: the ear counts the tools. Every tool that lands is one clank, and each kind of tool always rings its own note, so 4, 7, 11 and 19 tools are 4, 7, 11 and 19 sounds, and the pile becomes a roll; each job ends on its own bell. The repeat plays job 1's four clanks again at job 1's pace and is cut dead. The button room plays no clanks: one soft click and that job's bell, four times. - Under everything: room tone (noise low-passed at 900 Hz) at -52 dB, plus a copy a third of a second later, 3 dB lower, panned right. There is never digital silence; the quietest 50 ms is about -49 dBFS. - The sounds, what they are made of, and their gains in dB: - a pulse on every beat (a 62 Hz knock with its octave, 0.25 s), -20, from the start to the freeze (resting while the warehouse holds); - each typed frame: one tick (10 ms of noise, 1.8 to 6 kHz) on each frame where new letters appear, -20 (-21 in the button room), up to 3 dB louder when several letters appear on one frame; - the lid opening: a wooden clack (noise below 2.5 kHz with a 180 and 420 Hz body), -9; the lid shutting on each done frame, -17 (a lower, heavier one for the giant box, -13); - the toolbox growing: a low creak (a 62 to 46 Hz body with its octave, a creak sliding from 260 to 150 Hz), -6; - the pull-back to the warehouse: a boom (41, 82 and 123 Hz, with low noise, 2.2 s), -5; - a clank per tool: a struck tone with partials at 2.76x and 5.4x, over a tap of noise above 1.2 kHz and a small low thud (110 to 150 Hz). Each tool kind has its own note from a pentatonic set (G4 392, A4 440, C5 523.25, D5 587.33, E5 659.26, G5 783.99, A5 880 Hz, times 1.5, the last seven kinds an octave lower). The ring gets shorter as the jobs get denser (full, 0.8, 0.6, 0.45) and the level falls (-8, -10, -12, -14 dB for jobs 1 to 4), panned at random up to half way either side; - each job's bell: E5 (659.26 Hz), G5 (783.99), A5 (880), then a big C6 (1046.5) with a low 55 and 110 Hz body, partials at 2x, 2.76x and 5.4x, ringing 3.6 s: -11 (the big one -8); - the freeze: everything the old way was sounding stops in 6 ms, and ten frames of room tone only; - "Or…": one key tick, -18; - the cut: a soft paper "page" (noise 150 Hz to 1.4 kHz over a 160 and 330 Hz body), -19; - each button: one soft ROUND click (a 480 Hz body with 1250 and 2300 Hz, under a quieter tick), -10; 480 Hz sits below every bell, so a bell still ringing does not cover it; - the button room's bells: the same four notes, -10; - the end: a paper tick (noise 0.9 to 5 kHz) under each end line and the wordmark, -23. - Bigger hits (the lid, the growing, the boom, the bells) get a soft 2-frame swell of noise before them, while their own attack starts on their frame. - The hits and their frames (at 30 fps): lid 40; tools 60, 80, 100, 120; bells 140, 260, 340, 420; grow 180, 280, 360; boom 360; tools of job 2 every 10 frames from 190, of job 3 every 5 from 285, of job 4 every 2 or 3 from 363; lid jolt 480; slams 500, 520; freeze 530; "Or…" 540; page 560; clicks 600, 620, 640, 660; bells 610, 630, 650, 670; end ticks 720, 744, 768. - Build every attack exactly ON its frame. Then move the whole finished soundtrack 2 frames earlier: trim 3,200 samples (2 frames at 48 kHz) from the start and pad the end with the same amount of silence. Every hit then starts 2 frames BEFORE its picture; the ear hears that as exactly on time. - Master: normalise, then a soft knee (untouched below half of full scale, eased so it never passes -2.2 dBFS), with the gain set by measuring (ffmpeg's ebur128 filter) until the whole is -16 LUFS integrated. zippr's came out at -16.1 LUFS, true peak -1.9 dBTP. By section, the loudness rises with the pile (job 1 about -19.8 LUFS, job 4 about -13.1), the stop sits at about -45, the button room about -13.6 and the end card about -21.4.

The finish - 1080x1920, 30 fps, 28.0 s (840 frames). Chromium flags --force-color-profile=srgb, --hide-scrollbars, --font-render-hinting=none and --disable-lcd-text, device scale 1. Encode: H.264 High, level 4.0, yuv420p, crf 17, preset slow, converted with out_range=tv and out_color_matrix=bt709 and tagged BT.709 (primaries, transfer and matrix, set with ffmpeg's setparams as well). AAC 192 kbit/s at 48 kHz stereo, with +faststart. zippr's file came out at about 7 MB.

STEP 3: SWAP IN MY PRODUCT

Keep the ink-on-paper look, the one-accent rule, the doubling tool rate, the warehouse, the repeat and the freeze, the button room, the dotted ghost, the timings and the sound design exactly. Change only these: - The four asks become my customer's four jobs, and the tools become what each of those jobs drags out today. - The counts become my numbers (or keep them as a joke count, labelled as one to me). - "Or…" stays or becomes my turn word. - The ink boxes and their buttons become my product's one-step version of each job, in my accent colour and nowhere else. - zippr's mascot becomes my hero, from question 8. - The end lines, the wordmark and the address become mine. - My colours go on the colour roles (question 9).

STEP 4: SHOW ME THE PLAN, THEN WAIT

Write my story in two sentences, then my beat table with times and frames in zippr's format, then my colours on their roles. If you can run code, show me 4 stills in the finished look (for example at 4.1 s, 14.1 s, 17.7 s and 26.7 s), and a sound sketch if you can. Wait for my yes before you build the whole film.

STEP 5: MAKE IT

If you are an AI app that can run code and save files, follow the recipe and give me the MP4 with sound. Ask me before installing anything: GSAP and Inter are downloads unless they are already here. GSAP only gives the easing curves here, so if it can't be had, write the same eases (power2 in, out and in-out, and a back-out overshoot) as plain functions and tell me. If Inter isn't there, use any free sans-serif and tell me. If you can't run code, build it as one self-contained web page that plays in the chat. Draw every ink line as SVG from the frame number, and make the sounds live in the browser. Add a big Play button (browsers only play sound after a tap), a Replay button, and a "Save as video" button if the browser allows. Tell me how to screen-record it with sound on a phone and on a computer.

STEP 6: CHECK IT BEFORE YOU GIVE IT TO ME

- Put one frame from every 0.5 s on a contact sheet and look at it at phone size. Could someone who never saw this message say what the film means from the frames alone? Fix what doesn't read. - Check the accent colour appears only on the buttons, and that the hero's face is visible in every frame it is in. - On the file: it plays start to end; each hit sound starts 2 frames before its picture and is clearly audible (the 19-tool roll and the clicks under a ringing bell are the ones most easily buried); the stop is room tone only; the loudness is -16 LUFS and the peaks are below -1 dB. - Then give me the file, one still, and one honest line on anything you couldn't do. Never use copyrighted music.