Help & Documentation

A seated suspect in close-up and a standing detective behind them. With the reference camera (full frame, 85 mm at T1.9) the suspect sits at 1.80 m and the detective stands at 4.50 m; on the suspect the depth of field is under 5 cm.

• The page is the lobby: pick the level on the map and press 'Enter the simulator' — Quick start plays the level's own setup (see the lobby section below). The simulator always opens full screen; the 3D scene loads only then.
• The take starts Ready: press 'Roll the take' or Space.
• The cue at the top says whose eyes to find: the suspect, then the detective, then the suspect again.
• A beat counts as found the moment the eyes are inside the depth of field. Then hold focus there for 3 seconds. A beat you never find within 8 seconds is a miss.
• At the end, the results show how fast you got sharp on each beat and how precisely you held it. 'Go again' restarts the take; 'Exit' returns to the lobby.

Under the campaign map, the clapperboard shows the selected level, its best take and the setup 'Enter the simulator' plays. Everything else is folded away until you want it.

• Quick start: on the level's own setup (its lens wide open, full frame, TV & streaming, 24 fps at 180°) the clapperboard says 'Quick start' — press 'Enter the simulator' and play. Picking another level on the map brings that level's lens with it.
• Change anything and the line says 'Your camera package' with a one-line summary, for example 'Venice 2 · 6K 3:2 · 85 mm T1.9 · CoC 0.029 mm'. 'Back to the level's setup' puts it all back.
• Camera package: lens, stop, camera, sensor mode, frame rate, shutter, delivery format, the depth guide, peaking, frame lines, render quality and, in browsers with WebGPU, the renderer, folded into one section whose header is the summary. The settings in the simulator (M) are the same controls.
• Depth of field on the first mark: a live diagram of the shot as it starts — camera, subjects, the sharp band — with the depth of field and the score multiplier the take would carry. Change the stop, the lens or the delivery format and watch it move.
• Cast: pick the suspect and the detective by their portraits. Each role only lists the characters that have the animation it needs.
• Input check: click the strip, then turn the mouse wheel over it or press ↑ ↓; with a hand unit connected, turn its wheel. Each input that moves the marker is ticked. The best take shows your stars and your best score on the level.

Everything you need during a take is on screen, and every control has a key.

• Play / Pause (Space) rolls or pauses the take. While paused, the lens and the take clock stand still.
• Restart (R) starts the take over from the first beat and rolls.
• Rehearse (H) walks the blocking at half speed without a score: the take starts over paused, cues and timeline run as usual, and the results card says 'Rehearsal done — not scored'. Press H again for a scored take.
• Settings (M) opens every setting in a full-height panel on the right, one tab per group — Lens, Camera, Cast, Picture and Hand unit — and reopens on the tab you left; the take pauses while it is open. Click the dimmed picture, press M or Esc to close it.
• Esc first leaves full screen (the button in the top bar, or F, brings it back); Esc again or 'Exit' returns to the lobby. On an iPhone, which has no full-screen mode for pages, the simulator fills the window.

The ring turns from anywhere in the simulator except inside the settings panel.

• ↑ and ↓ turn the ring by one step (↑ toward infinity). Hold a key to pull continuously; it speeds up the longer you hold.
• Alt with an arrow is a fine step for the last millimetres; Shift is a coarse step for big racks. Page Up and Page Down jump further; Home and End go to the near and far stops.
• The mouse wheel or trackpad turns the ring too: scrolling down focuses farther. Ctrl with the wheel still zooms the page.
• On a touch screen, drag the round focus wheel beside the timeline like a hand unit's knob: clockwise focuses farther.

A Tilta Nucleus hand wheel can turn the focus ring, in Chrome or Edge on a computer or an Android phone. Connect it on the card beside the input check in the lobby or in the settings (M); the link stays up while you go in and out of the simulator.

• Switch the hand unit on, choose Connect hand unit, and pick the NUCLEUS device in the browser's list. Then turn the wheel once: the unit only reports when it moves.
• Calibrate once per lens: turn the wheel to both end stops, then choose whether that span drives the whole lens throw or only the marked section (close-focus mark to ∞). The calibration is saved with your settings.
• If the link drops, the simulator reconnects by itself for two minutes and tells you; the arrow keys, the mouse wheel and the on-screen wheel keep working meanwhile.
• The card shows the link's rate and the delay inside the app (median and 95th percentile) so you can judge the feel. Nothing is smoothed: the unit already smooths its own reading.
• Safari, the iPhone and the Cine Power Planner apps cannot connect a hand unit yet; the card says so instead of offering a button.

Two aids under the picture, updated live while you pull.

• The depth guide shows the camera on the left and both people on a distance scale. The blue line is where the lens is focused; the green band is what is sharp, with its near and far limits.
• The ringed person is the current target. A person turns green when their eyes are inside the band; an arrow beside the blue line points the way to turn. Switch the guide off in the settings once you pull by eye.
• The timeline shows the three beats — done with its grade, current with a bar that fills while you search and turns green while you hold, and to come — over a trace of the take: the mark in yellow, your focus in white (green while sharp).
• Grades: Critical within half a circle of confusion, Sharp within one, Soft within two, Miss beyond. The readouts show the target, lens focus (MOD at the near stop, ∞ at infinity), how far in front of or behind the eyes you are, and precision.
• Distances follow your account's units (Settings › General › Units): in imperial the marks read in feet and inches (5′ 11″), the depth guide is marked 5′, 8′, 10′ …, the focus error is in inches and the depth of field in inches. The custom circle of confusion always stays in millimetres.
• Hide diagrams in the bar under the picture (or D) turns the depth guide and the focus trace off together, and the picture grows into their room; the beat bars and the readouts stay. Settings › Picture switches each one on its own. Each diagram hidden from the moment the take rolls until it ends multiplies the score by ×1.05 (both ×1.10); showing it for a moment, even while paused, loses that factor for the take.

Every finished take gets a score, and the results card shows what it is made of.

• Each beat scores up to 100: 60 % accuracy (how sharp you held the eyes), 20 % speed (full marks when sharp within 1 s of the cue, none at the end of the beat's window — the 8 s timeout on the studio lot, the beat's own slot on the level clock elsewhere), 20 % stability (the share of the hold inside the depth of field), minus 5 points per overshoot, when focus swings right through the eyes and out the other side (at most 3 are counted). A missed beat scores 0; the take's base score is the average.
• The base score is multiplied by the difficulty. It depends on how deep the depth of field is at the marks compared with the level's own setup (85 mm on full frame at T1.9 for Level 1). A shallower setup, from a wider stop, a bigger sensor or the Cinema delivery format, multiplies it by up to ×2; a deeper one never earns less than ×1. Changing only the focal length hardly matters, because the shot size stays the same.
• Frame rate and shutter add at most ×1.1, for a staler, more smeared picture (for example 24 fps at 360° with motion blur). Moving shots add a movement factor for their world: ×1.05 when the camera moves, ×1.1 when the camera and the actors move, ×1.15 on a zoom (the finale counts as a zoom). A static shot stays at ×1.
• The take is scored on the easiest setup it saw: stopping down in the middle of a take and opening up before the end earns nothing.
• Your best score for each level and setup (lens, stop, sensor, delivery format, frame rate, shutter and motion blur) is kept with your preferences and syncs to your other devices. The results card says when you beat it, and shows it otherwise.
• Hidden diagrams add ×1.05 each for the depth guide and the focus trace when they stayed hidden for the whole take; the results card names the factor. The leaderboard replays the take from its inputs and cannot see your screen, so it ranks the take without this factor.

Where the points are, in the order they pay off. After every take the results card names the one lever that take left.

• Rehearse first (H): the blocking plays at half speed and nothing is scored, so you learn where each mark is and when each cue comes before a take counts.
• Find every beat. A missed beat scores 0, so a late find is always worth more than giving up on it.
• Accuracy is 60 % of each beat: keep the eyes Critical, within half a circle of confusion. Rack with Shift for the long part, then hold Alt for the last fine steps.
• Speed is 20 %: full marks when you are sharp within 1 s of the cue. Start turning on the cue — the line under the picture says which way and to which mark.
• Stability is 20 %: the share of the hold spent inside the depth of field. On a moving shot keep following the subject until the next cue.
• Every overshoot costs 5 points (at most 3 per beat). Slow down near the mark and land on it from one side instead of swinging through the eyes.
• Set the multipliers before you roll; a change mid-take only ever counts the easier setup. A shallower depth of field pays most, up to ×2: a wider stop, a bigger sensor or the Cinema delivery format (the focal length hardly matters). Hiding the depth guide and the focus trace adds ×1.05 each, and 24 fps with a 360° shutter and motion blur up to ×1.1. The world's movement factor is fixed.
• Peaking, frame lines and render quality never change the score — use whatever helps you see focus.

An opt-in board per level where your takes are ranked against other players. It is off until you join, and every posted take is replayed and scored on the server.

• Join under Public leaderboard in the lobby: choose a display name (3–24 characters — letters, digits, spaces, dots, hyphens or underscores; a pseudonym — better not your real name). Nothing is sent until you post a take. No e-mail addresses, web addresses, phone numbers, reserved or offensive names.
• After a scored take, press Post as <your name> under the results. The server replays your take from its recording on the same simulation and scores it itself; the board holds the server's score. A take whose replay does not reach your score is not posted.
• A take can be posted only when the lens, stop and shutter stayed the same while it ran, on a lens from the built-in library (an imported lens file never leaves your device), and on a level with a board. The results card says why when a take cannot be posted.
• Boards: All setups ranks each player's best take on any setup; Level's own setup only takes on the level's lens and stop, full frame, TV delivery and 24 fps or crisper — like for like. All time or This month; Everyone or Team (you, people you share a project with, and your connected crew). Your own row is marked and always shown.
• The board keeps your best take per setup per month. When the scoring rules change, the boards start fresh.
• Change name renames all your posted takes. Leave the leaderboard deletes every take you posted, for all levels and months. Withdrawing your cloud consent or deleting your account deletes them too; the personal-data download lists them.
• A small crown with four dots next to a name is Mr Watts' crown: one dot per world where that player posted a ★★★ take of every level. The server rates the posted takes itself, so only posted takes count — a ★★★ you never posted earns no dot here.

Change the camera and the shot stays the same — the distances change instead.

• Pick any camera and sensor mode from the device catalogue; none means full frame. On a longer lens or a smaller sensor the people stand farther away, on a wider one closer, so each face fills the frame the same way. The lobby shows where the marks land.
• The delivery format sets the circle of confusion, i.e. what counts as sharp: Cinema (big screen, sensor diagonal ÷ 3000 — strictest), TV & streaming (÷ 1500, the app standard), Online & mobile (÷ 1000 — most forgiving) or a custom value in millimetres. It changes the grade and the green band, not the picture.
• Frame rate (23.976–60 fps) and shutter angle (45°–360°) set the exposure — 1/48 s at 24 fps and 180°. The picture updates on the camera's cadence, and with motion blur on a rack pulled during the exposure smears through focus.
• Brightness follows the shutter (a narrower shutter is darker) unless you switch that off, as you would compensate with ISO or ND.

Set these under Camera package and Cast in the lobby (folded away under the campaign) or in the settings panel (M).

• Lens: every prime in the bundled Tilta lens library. Each turns the ring on its own engraved focus scale.
• Stop: the lens's own iris stops, wide open by default. Stopping down makes the level easier.
• Import a Tilta lens file (.tl) to practise on your own calibration. It is read on your device, never uploaded, and kept for this visit only. Zoom lenses are not supported yet.
• Focus peaking colours the edges that are sharp on screen: choose its colour, strength and threshold (lower marks softer edges too). Frame lines show 2.39:1 (the default) or 16:9, or are off. Render quality is explained in its own section below.
• Cast: choose, by their portraits, who plays the suspect (characters with a seated animation) and the detective (characters with a standing one). Where the characters cannot load — offline, or in the mobile and desktop apps, which do not include them — the simulator uses simple built-in figures.

Settings (M) → Picture → Render quality. Phones and tablets start on Low and offer Low and Medium (High and Ultra are too slow there); computers start on High. Every tier keeps the eyes exactly as sharp at critical focus — the extras are light and texture, never a softer picture.

• Low: the lightest picture, for phones, 960 pixels wide. Contact shadows under the people and the furniture, nothing else extra.
• Medium: 1280 pixels wide, and adds light bounced from the room (image-based lighting), a soft sheen on skin and clothes, soft shadows from the two key lights, and film grain that changes once per frame of your chosen frame rate.
• High (computers only): 1920 pixels wide; the practical nearest the suspect casts a shadow too, and the gate weaves slightly, like film in a camera gate. The weave stays off when your system asks for reduced motion.
• Ultra (desktop), computers only: 2560 pixels wide and larger shadow maps. It needs a strong graphics card.
• Temporal anti-aliasing (High and Ultra, off by default): smooths jagged edges over several frames, but also softens the finest detail you judge focus on. Leave it off while you train.
• Renderer (only in browsers that offer WebGPU; elsewhere the setting is not shown and the simulator runs on WebGL 2): WebGL 2 is the standard, and every visit starts on it. To try WebGPU (experimental), pick it beside Render quality, in the camera package or on the Picture tab. It draws the same optics with more bokeh samples, bright highlights such as bulbs and catchlights as discs in the shape of the iris, and ambient occlusion in creases and corners from Medium up. Switching restarts the take, paused. Where WebGPU cannot start on the device, the picture stays on WebGL 2 and the setting says so; a WebGPU device lost mid-take switches back to WebGL 2.

The page opens on the campaign: a film set with four worlds, and the selected level on a clapperboard with its brief, its best take and Enter the simulator. Every scored take earns up to three stars, and ★★ on a level opens the next one.

• The four worlds — 1 The studio lot (static shots), 2 The dolly track (the camera moves), 3 The street (camera and actors move), 4 The zoom tower (moves plus zoom). Each has four levels. All sixteen levels are playable, from the Interrogation Room to the Crash Zoom.
• Stars — ★ for finding every beat, ★★ from 75 points, ★★★ from 90 points with no beat graded Miss. The stars use the take’s base score times D★, which compares your setup with the level’s own: more depth of field than the level’s reference lowers the stars, a harder setup never raises them. The score and the personal best keep the full difficulty multiplier.
• The results card counts the stars up, shows your best on the level and says what the next star needs. When your setup cost stars, it says so and names both depths of field.
• Unlocking — ★★ on a level opens the next level of its world. A world opens once every level of the world before has ★★ or more. When a take opens a level, “Next level unlocked — to the map” closes the simulator and Mr Watts (Miss Watts in the pink theme) jumps to the new level.
• The map — select a level on the map or in the list under it; the clapperboard above the worlds shows and plays the selected level. With the keyboard, Tab reaches the list, the arrow keys move between levels, Home and End jump to the ends, and Enter plays the level. Mr Watts stands on the selected level and walks, runs or jumps to the next one; with reduced motion the mascot appears there at once.
• Star total and the finale — the map counts your stars out of 48. At 36 stars the bonus level after world 4 opens: Level 17, “The Oner”, one take on the station platform where the traveller walks in while the dolly moves and the zoom tightens, then a rack to the conductor and back. Free play opens it too. Its own stars are shown on its node and do not count toward the 48.
• Free play — for trainers and film schools: the switch on the map opens every level. Stars are still recorded, and switching it off brings back the unlocks you earned.
• Stars and unlocks are kept with your account settings and sync to your other devices. Your best stars on a level are never lowered by a worse take.

World 1 holds four static levels. After the Interrogation Room come three that each teach one thing a tape measure will not tell you.

• Level 2, The Evidence Table — a tabletop shot at close focus: the phone’s screen, a shell casing, the seated witness’s eyes, then the signature on a statement. At 85 mm T1.9 the depth of field on the phone is about 1.2 cm, and the phone sits just outside the lens’s close focus (MOD). A lens that cannot focus that close frames the shot looser, and the camera card says so.
• Level 3, Rain on the Window — a raindrop on the café window, a stranger outside under an awning, and a lit window in the skyline. The skyline is 120 m away, short of the hyperfocal distance (about 132 m at 85 mm T1.9), so parking the lens on the ∞ stop leaves it soft. Focus just short of ∞ — at the hyperfocal distance everything from about 65 m out is sharp.
• Level 4, The Mirror — a note stuck on a mirror and a reflection in it. The reflection is not at the mirror’s distance: the lens focuses along the light path, camera to mirror and on to the face (here 2.20 m + 1.90 m = 4.10 m). The note on the glass is at 2.20 m; the face in it is almost twice as far. A lamp seen in the mirror is at its own light-path distance too.
• Objects as marks — a prop’s mark is one point on it (the screen, the casing’s mouth, the signature, the drop, the lit window, the note). The depth guide shows a prop with a cube instead of a person, and the telemetry names it.
• Cast — each level casts from the same two picks: the suspect’s actor plays every seated part (the witness), the detective’s every standing part (the stranger, the reflection) — except the street’s lead walkers, whom the suspect’s actor plays standing up.

World 2 holds four levels where the camera moves and the actors stay put. The beats run on the level clock: a cue comes at its moment, whether you are ready or not, and a hold is scored while the dolly moves.

• Level 5, The Push-In — the dolly eases in on a seated writer, from 3.90 m to 1.60 m. Find the eyes while the camera is still, then ride the focus with the move: the distance falls the whole way, and the depth of field shrinks from about 23 cm to 4 cm at 85 mm T1.9. Hold the close-up when the dolly lands.
• Level 6, The Pull-Out Reveal — a close-up on a guest at 1.45 m; the dolly pulls back to 3.45 m and reveals the concierge by the desk at 5.41 m. Riding a pull-out is easier than a push-in (the depth of field grows), but the wheel still has to move. Then rack to the concierge and back.
• Level 7, Along the Bar — the camera trucks along a bar and eases in a little. A sideways move alone changes no focus distance; the ease-in does (the patron goes from 3.80 m to 3.30 m). A bottle on the counter crosses the frame close to the lens: rack to its label while it passes, then back to the patron once it has wiped past the face.
• Level 8, The Arc — the camera circles the diner and pans to keep the face centred. On a centred arc the distance does not change (2.40 m all the way round): leave the wheel alone. On the off-centre arc that follows, the circle’s centre is in front of the diner and the distance falls to 1.97 m: follow it.
• Rehearsal — every move runs at half speed with no score in rehearsal (Rehearse, or H, in the session), so you can learn where the marks go before a take.
• Other lenses — on a longer or shorter lens the whole move scales with the shot, so each framing stays as authored. On the arc the circle grows or shrinks too, so the centred arc still keeps one distance.

World 3 holds four levels where the actors walk and the camera moves too. A walk is faster than a dolly and changes pace, so lead it: turn the wheel for where the actor will be, not where they were. Tape on the floor marks where each actor starts and stops.

• Level 9, Walk and Talk — a Steadicam leads an editor and a reporter down an office corridor at a steady pace, but they don’t keep it: brisk, then slower while they argue, then brisk again. The editor’s distance breathes in from 3.00 m to 1.99 m, out to 2.87 m and in to 1.83 m before they stop at 1.97 m. Rack to the reporter half a metre behind, and back.
• Level 10, The Approach — a courier walks toward a camera that dollies back almost as fast, so the distance closes at only 0.2 m/s. When the dolly stops, the courier keeps coming at 1.25 m/s and the distance falls to a close-up at 1.50 m, where the depth of field is about 3 cm at 85 mm T1.9. The motor needs a moment to follow: lead the walk.
• Level 11, The Cross — on a station platform a conductor ambles toward you while a traveller hurries in from the back and overtakes. At the moment the traveller passes, both are 3.75 m away: hand off right there and there is nothing to rack. Then rack between them as they stop.
• Level 12, The Doorway — hold the room number on the door frame while a visitor walks past you through the door, then rack through the door to the host in the lit room. Once the visitor has turned round (stepping on the spot, when their character has a turn animation), hand off to them and ride them back to the doorway. Nobody is scored with their back to the lens.
• Marks and walks — every actor who moves has a tape mark in their own colour where they start and wherever they stop, as a second AC lays them in rehearsal. A walk’s stride follows the distance walked, so the feet stay on the floor whatever the pace. Rehearsal (H) walks the whole blocking at half speed.
• Cast — the suspect’s actor stands up and leads every walk on the street, and the detective’s actor plays the other walker. Both default characters walk with their own walk animation; a character without one glides on the standing idle. Recast either in the lobby or in the settings.

World 4 holds four levels where the operator zooms and you pull focus. Each level is shot on its own zoom, so the lens is locked while one is selected; the stop is still yours. The zooms teach what a zoom does to focus: the depth of field changes with the focal length, a varifocal zoom moves the focus, a ramping zoom closes its iris, and every zoom breathes a little.

• Level 13, The Speech — a parfocal cine zoom (24–70 mm T2.8) creeps from 28 mm to 70 mm while a minister speaks behind a lectern, leans in, turns and steps closer. At 28 mm those small moves disappear in the depth of field; at 70 mm it is about a quarter of a metre either side of the eyes, so the same moves have to be ridden. Rack to the aide while the shot is still wide.
• Level 14, The Varifocal Trap — a stills zoom (24–105 mm T4) is varifocal: zooming in pulls the focus toward the lens. The busker stands still, but a ring left alone goes soft as the zoom goes to 105 mm. Turn toward ∞ as the zoom tightens, by the picture, not the marks: the motor readout shows the ring as engraved, and on this zoom the marks lie.
• Level 15, The Vertigo — a dolly zoom: the camera dollies in while the zoom widens at exactly the rate that keeps the witness the same size. On the monitor only the corridor behind changes, while the distance falls from 5.60 m to 2.00 m. Ride the distance, not the picture; then rack to the watcher at the far end.
• Level 16, The Crash Zoom — a broadcast zoom (28–200 mm T3.5–5.6) whose iris closes above 70 mm. At 28 mm most of the platform looks sharp; be on the traveller anyway, because the snap to 150 mm shows where focus really is. The picture darkens as the iris closes. Then the traveller walks in while the zoom widens and the stop opens again.
• Breathing — a zoom's picture grows a little as focus comes closer. It changes the framing, never the depth of field or the score.
• The lens — a zoom level shows its zoom's range in the camera package (for example 24–70 mm T2.8) and its current focal length beside the lens name during the take. These are training zooms modelled on each type, not a maker's product.

Earn every star in a world and Mr Watts (Miss Watts in the pink theme) gets a gem for the crown. The crown appears with the first gem.

• A gem per world — a world’s gem is set when all four of its levels have ★★★. Each gem has its world’s colour: blue for the studio lot, orange for the dolly track, teal for the street, purple for the zoom tower.
• The crown — Mr Watts wears no crown until the first gem; a single ★★★ level does not give one. The crown has four sockets, and the ones you have not earned stay empty. The map’s header says how many gems it holds, for example “Crown: 2 of 4 gems”.
• The moment — when a take earns a gem, the results card still shows the crown as it was. Back on the map Mr Watts cheers: the first gem brings the crown down onto the cap with the gem already set, and each later gem flies into its socket. The fourth gem completes the crown with a short sparkle. With reduced motion the crown and its gems simply appear.
• Where it shows — only in the Focus Pulling Simulator: on the map, on the results card and, as a small badge beside names, on the public leaderboard. Mr Watts elsewhere in the app is unchanged.
• The crown is worked out from your best stars, so it syncs with your account like the stars and can never disagree with them. A gem needs all four levels of its world to be built; until they are, the crown stays out of reach.