Reference D · Failures by symptom
Reference
Every failure that the tool sections and the stage chapters tabulate, sorted by what you see on the screen or hear in the speakers. Each row gives the likely cause and the fix, worded as its chapter words them, and the section that teaches it.
In this reference
D.0: three questions to ask before you look a row up
Thirteen headings, D.1 to D.13: face; hands; product; text; camera; light, look and world; motion; performance; voice; sync; music; sound; file
232 rows: what you see or hear, the likely cause, the fix, and where the fix is taught
How to read a row. The fix cell gives the first lever; where the chapter gives a next step, an arrow leads to it. A row from a tool section applies to that tool; a row from a stage chapter (Chapters 1, 2, 7, 9, 10, 13, 14, 15, 22 and 25) names a fault whatever the tool. A dash in the cause cell means the chapter's table gives the fix without a cause. Where a chapter's table adds a stop condition, it follows the fix ("stop prompting when"). The same fault can appear under two tools with different fixes: read the row for the tool you use.
D.0 Before you look anything up
Most wasted credits come from repairing the wrong layer. Ask three questions first (29.12):
Is it already in the source? Open the keyframe, or listen to the take alone. A defect at frame 0 or in the take is fixed in the still or the line, not in the clip.
Does every attempt do it? The same defect every time points to the shot, the route or a sentence in the prompt; a different defect each time is variance.
Does the route own this behaviour? A mouth moving in a pause, a line re-timed by a re-performance, a lip-sync output shortened to the shorter of its inputs: no wording will move them.
Then climb the smallest lever first: a different select or trim in the cut, a local fix in post, a one-change edit of the still, a new generation with one variable changed, another route, and last a new plan for the shot (29.12). The rungs for a failed clip are in 14.16.
D.1 Face
A face that is not the person's or that drifts, skin that looks wrong, a face that acts, a twin or an extra person.
You see or hear | Cause | Fix | Where |
The same person twice: a face drifts between shots, and further along a chain of edits; OpenAI's guide says to restate what must be preserved when a chained result drifts | — | sign one still of each character; compare every later frame with it at full size; return to the approved parent instead of editing a copy | 1.3 The camera is a model (Chapters 9, 10) |
The uncanny human: a face reads as wrong at once, at the size it will be seen | — | judge faces at viewing size and at full size; design subjects and actions the models render truthfully | 1.3 The camera is a model (Chapters 10, 15) |
The wrong identity in a good scene | — | give the identity picture its authority again; remove competing faces → stop prompting when the face still differs in a view that matters; then a new accepted frame, or another route | 2.9 Directing in words |
A wrong profile the model has never seen | — | make and approve that view first → stop prompting when a real actor's likeness cannot be reached; then an approved side view, before any motion | 2.9 Directing in words |
Skin looks plastic or aged | no skin facts, or pores everywhere | the character's own marks; shine following the light; judge at viewing size | 7.9 Production design |
Plastic skin | no skin facts | one to three specific ones: pores, sheen on the nose, the lines from nose to mouth | 9.8 The finished frame |
The face is acting | an emotion adjective | describe the muscles, or write "her expression does not change" | 9.8 The finished frame |
A younger, prettier face than you wrote | the face was written from a type, not from asymmetries | put the fixed facts early, asymmetries first (element 3) | 9.8 The finished frame |
The wrong face in an otherwise right scene | — | go back to the identity picture; remove competing faces → stop prompting when the face still differs in a view that matters; then a new accepted frame, or another tool | 9.9 The finished frame |
Plastic skin | no skin facts | add one to three: pores on the nose and cheeks, a sheen on the forehead, the lines from nose to mouth | 11.8 Nano Banana Pro |
The face is acting | an emotion word | write the muscles, or "her expression does not change" | 11.8 Nano Banana Pro |
A younger, prettier face than you wrote | the face was written from a type | put the fixed facts early, asymmetries first | 11.8 Nano Banana Pro |
The face drifts from the photograph over several frames | the clause was retyped, the master was not attached, or an edit was made of an edit | re-attach the master, paste the clause, and start again from the last approved frame | 11.8 Nano Banana Pro |
A face or hand moved in an edit | edited from a damaged frame, or a chain of edits | return to the approved parent; composite the region | |
Costume or face drifts | too long, or the still already drifts | a shorter clip; protect by name; fix the still first | 14.16 The moving shot |
A face warps on a turn | A turn past about sixty degrees | A small turn, or two shots cut on the look | 16.8 Kling |
A twin, or an extra person | a picture with several faces, or near-duplicate references | separate clean files; the count and the one-instance line | |
One character takes another's face or voice | the upload order does not match the numbers | re-attach in order of first appearance and renumber | |
Glowing eyes | strong emotion words | calmer words; "normal human eyes; no glowing eyes" as the top-priority negative | |
A reference's identity drifts | one reference did two jobs | one job per reference | 19.8 Veo and MiniMax |
Skin waxy, a label or badge changed | creative or generative upscaling | precise mode or a lower factor; failing that, native and composite the text in post | 32.8 Premiere and finishing |
D.2 Hands
Grips, contacts, fingers, and who holds what.
You see or hear | Cause | Fix | Where |
Hands and physical states: hands holding small objects are among the hardest cases; what rests on what is easily missed | — | describe the support and the contact; attach a reference photograph of the pose | 1.3 The camera is a model (Chapters 9, 10) |
A wrong grip or contact | — | fix what supports what; add a blocking photograph → stop prompting when a reworded attempt returns the same grip; then a new keyframe, or different coverage | 2.9 Directing in words |
A wrong grip or contact | — | correct what supports what; add a blocking photo → stop prompting when two attempts repeat the same contact; then a new keyframe, or different coverage | 9.9 The finished frame |
The event is wrong (who holds what) | the instant was implied | positions and contacts | |
Fingers melt at the touch | contact rushed, or the hand hidden in the still | name and slow the contact; path, grasp, rest; a cleaner grip keyframe; split at the touch | 14.16 The moving shot |
A duplicated prop, or a phantom grip | no single instance, no release gate | "one [prop]"; release only after the grip | 14.16 The moving shot |
The prop stays with the giver, or floats | — | "takes its weight, and only then … lets go"; the giver empty-handed after | 15.10 Directing performance |
D.3 Product
A product whose shape, size or badge is wrong, or that does not sit in its scene.
You see or hear | Cause | Fix | Where |
A product that keeps its wrong shape | — | supply real photographs of its front and side → stop prompting when the model keeps redesigning it; then a photographed or composited product | 2.9 Directing in words |
The product floats | it has its own light and no contact | the scene's light, a contact shadow, an overlap, a hand for scale | 7.9 Production design |
A product the wrong shape | — | attach real front and side photographs → stop prompting when the model keeps redesigning it; then the photographed or composited product (10.11) | 9.9 The finished frame |
The product grows or shrinks between shots | no scale lock | the forearm test, and "smaller, never larger" | 10.11 Identity and references |
Cap or shoulders redesigned | the model's idea of the product | attach the side view again "only to resolve that geometry"; after two controlled attempts, a photographed or composited product | 10.11 Identity and references |
A second bottle, badge or seal nobody asked for | the model's habit | name it in the exclusions | 10.11 Identity and references |
A product's shape or label is wrong | words cannot hold a product | attach front and side photographs; then composite the photographed product | 11.8 Nano Banana Pro |
D.4 Text
Lettering, logos, Arabic and captions: in the frame, on a plain object and on the timeline.
You see or hear | Cause | Fix | Where |
Words in the picture: misspelled text, invented logos, faint printing on plain objects | — | two layers: in-world text may be generated and checked letter by letter; logos, titles and legal lines are always set in post | 1.3 The camera is a model (Chapters 10, 30) |
Stray printing or logos on plain objects | the model's habit | clean it in post (Chapter 10) | 7.9 Production design |
Fine type wrong | — | correct it once or twice → stop prompting when the delivery needs exact text, or two tries failed; then real artwork composited on a clean surface (10.12) | 9.9 The finished frame |
Logos, faint printing or extra text on plain objects | the model adds print to surfaces | describe the surface as plain and unmarked; clean it in post | 11.8 Nano Banana Pro |
In-frame text is misspelt, or Arabic letters are wrong | small text and fine detail may not come out right | quotation marks and the exact string; two tries at most; then a blank panel and the real text set in post (Chapter 30) | 11.8 Nano Banana Pro |
Captions or watermark-like text appear | the closing line is missing | end on "no overlaid text, subtitles, logos, or interface graphics" | 11.8 Nano Banana Pro |
Marks or letters on a plain object | the model adds plausible detail | "blank label panel, no lettering or marks"; remove in an edit, or paint out | |
Misspelt or extra words | placement and clarity are documented weaknesses | quote once, fence the frame, raise quality; after two controlled attempts set the words in post | |
Arabic unjoined, wrongly dotted or reversed | OpenAI claims nothing for Arabic | blank the panel and set the line from the real font | |
The type block off the brief, or the quiet zone filled | layout placement is a documented weakness | write the zone as a surface, cut elements, or set the layout in post and keep the frame for mood | |
A label smears or re-letters mid-move | text under motion | a shorter move; the label composited from the artwork | 14.16 The moving shot |
Lettering misspelt | text generated in the frame | composite it in post; the maker's fix is separate letters, or the text supplied as an image | |
Subtitles appear | dialogue repeated in the prompt, business tied to words, or subtitles in a reference | say the line once, time the business by order, use "Character's line (emotion): content"; clean references | |
Arabic caption characters reversed | the caption track's Text Engine is not the South Asian and Middle Eastern one (typical after an SRT import) | set it in Track Settings, re-create the captions, never reverse by hand | 32.8 Premiere and finishing |
Full stop or comma on the wrong end of a second Arabic line | typing direction | switch on right-to-left typing in Essential Graphics | 32.8 Premiere and finishing |
No transcript for Arabic dialogue | Speech to Text lists no Arabic | use the approved script, or import a JSON transcript from another tool; check its timings against the waveform | 32.8 Premiere and finishing |
D.5 Camera
A camera that does not move as written, framing that differs from the frame it started from, borrowed or mirrored geography, edges and depth of field.
You see or hear | Cause | Fix | Where |
The edges look accidental | an object half-clipped, a bright leak | move the object in or out; darken the leak; add a foreground witness | 7.9 Production design |
The plate brings its framing into a new shot | it was fed without a fence | "take the room; take nothing of its camera position" | 10.10 Identity and references |
The room comes back mirrored | landmarks were described as screen sides | state them in world terms first, then as seen from this station | 10.10 Identity and references |
The depth of field or the scale does not match the numbers | numbers are cues, not physics | write the result beside the number; recheck the distance with the formula (11.3); judge the frame | 11.8 Nano Banana Pro |
The camera floats | a move without a mechanism | give the move a body, or declare the camera locked | 14.16 The moving shot |
The camera zooms and does not move | The far feature grew as much as the near one | Name the mechanism and the destination in the rig sentence; measure again | 16.8 Kling |
Wider framing than asked, or the shot opens empty | a reference does not lock the frame | write the shot size, camera height and who is in frame at the start | |
The camera moves, or a cut you did not ask for | no fixed-shot sentence; no single-shot line | the fixed-shot sentence first inside the shot; "One continuous shot" on the first line | |
The framing differs from the keyframe | a reference does not fix the frame | write the shot size and camera position; if it must match, a Seedance or Kling start frame |
D.6 Light, look and world
Light in the wrong state, objects nobody asked for, a room that looks staged, colour that has slid, a reference that leaks its ground, an edit that spreads.
You see or hear | Cause | Fix | Where |
Physics: shadows that disagree with their source; liquids, reflections and fast motion; the tells cluster around physics, not texture | — | name each light by what it is and what it does; describe what rests on what; cut before a fast move peaks | 1.3 The camera is a model (Chapters 6, 7, 14) |
Crowds and scale: crowds default to uniformity; the scale of objects to one another breaks | — | clusters rather than exact counts; test a count before a client shot | 1.3 The camera is a model (Chapters 10, 14) |
The generic look: "cinematic" pulls a room towards teal and orange; "8K" and "ultra-realistic" set nothing | — | name sources, surfaces and colours; no quality words | 1.3 The camera is a model (Chapters 2, 6, 7) |
A scene drifting after several edits | — | return to the last approved frame → stop prompting when you are editing the damaged copy; then one new edit from the accepted source | 2.9 Directing in words |
The room looks new or staged | wear was written as a mood, or not at all | name the touch points and their marks; one exclusion against the cliché | 7.9 Production design |
Everything is equally detailed | too many materials; density written as a count | three or four material families; clusters by owner; fewer layers | 7.9 Production design |
Clothes look airbrushed | colour and mood, no construction | fabric, seams, closures and wear state | 7.9 Production design |
One object from the wrong world | the genre's or the present's default | an inventory against the world note or the period | 7.9 Production design |
A light is on that should be off | the state was mentioned once, and as a negative | an edit of that frame that says it firmly and names the only light | 9.8 The finished frame |
An object you never asked for | a surface was left unnamed | name every surface and what sits on it; describe emptiness as a surface ("the oilcloth between the laptop and her is bare") | 9.8 The finished frame |
The room goes teal and orange | a grade word, or "cinematic" | delete it; name three real surfaces and their colours | 9.8 The finished frame |
A small unrelated mark on a plain wall | — | clean it in post, or a narrow edit → stop prompting when repeated edits start moving the actor or the product; then the original with an approved patch | 9.9 The finished frame |
A figure or object left in the "empty" room | the removal list was incomplete | name everything that goes, and what the surface becomes | 10.10 Identity and references |
Surfaces smooth and plastic after an edit | the edit softened a crisp plate | composite the original texture back into the untouched regions | 10.10 Identity and references |
A light is on that should be off | the state was mentioned once, and as a negative | an edit that says it firmly and names the only light: "the only light in the room is the low warm bulb" | 11.8 Nano Banana Pro |
An object you never asked for | a surface was left unnamed | name every surface and what sits on it; describe emptiness as a surface: "the tabletop between the glass and the tray is bare" | 11.8 Nano Banana Pro |
The frame goes teal and orange | "cinematic" or a grade word | delete it; name three real surfaces and their colors | 11.8 Nano Banana Pro |
The reference's room, light or clothes came along; a second face; a grey ground | the picture owns everything it shows and no fence was written | write the fence ("take nothing of its ground, light, clothes or pose; no face"), or crop the picture to its one job | 11.8 Nano Banana Pro |
An edit moved something you did not mention | no mask: the whole frame is redrawn | repeat the preserve list and name the thing that moved, or composite only the changed area onto the parent | 11.8 Nano Banana Pro |
A relit frame looks pasted in | Google notes that complex edits, lighting changes among them, can produce unnatural artifacts | a one-change edit that names the new light's source, direction and Kelvin and what it does to each named surface; inspect shadows and reflections | 11.8 Nano Banana Pro |
The finish looks like damage: dirt, scratches, heavy noise | the finish block asked for age or damage | describe capture only; add delivery grain and halation in post (Chapter 30) | 11.8 Nano Banana Pro |
The edit changed more than asked | the change named no boundary | name consequences and protected regions; one change; composite | |
The reference's moment came back | the picture's state beat the written state | make them agree, or say what not to copy | |
An edit came back as a new shot, restyled throughout | wording that reads as a new shot, or no task word | task word first, the change from A to B, the protected list; else composite in post | 13.7 Choosing the route |
A reference job copied the source clip's person or room | edit wording in a reference job | remove "replace" and "instead of"; say the clip supplies only its camera path | 13.7 Choosing the route |
A detail flips (a bangle's wrist, a shirt's colour) | The prompt contradicts the frame | Correct the prompt; the frame does not always win | 16.8 Kling |
A reference's backdrop, light or grade leaks in; sheet panels or grey ground appear | no fence | "take nothing of …"; crop the file to its job; "do not show the sheet, its panels or its backdrop" | |
Fingerprint-like texture in grass, foliage or fabric | a high-resolution AI-made reference | resize the reference to no larger than the output | |
An edit comes back a new shot, or spreads, or half applies | no task word, the wrong mode, a loose scope, or two changes in one job | "Edit video:" first, the mode set, when and where, one change, the protected list; after two that spread, composite in post | |
People appear in an empty plate | the model fills the space | "No people", and what makes emptiness plausible | |
A reference's backdrop or light leaks in | no fence | "take nothing of its framing or background" | |
The look changed | the prompt restyled the frame | remove style words; the frame owns the look | 19.8 Veo and MiniMax |
D.7 Motion
An action that does not happen, overshoots or is skipped; warping and drift inside a clip; extensions and end frames.
You see or hear | Cause | Fix | Where |
An action that never happens: a start frame that already shows the action finished leaves the clip nothing to do | — | freeze the keyframe early in the action, each mover's direction written as positions | 1.3 The camera is a model (Chapters 9, 14) |
The clip performs no action | the keyframe shows the end state | an edit back to the first instant, before any video re-roll | 9.10 The finished frame |
Blurred limbs or a warped frame in the clip | an open verb, or an impossible contact, survived in the still | rewrite it as positions; make the contact possible | 9.10 The finished frame |
A clip started from this frame begins with a photograph waking up | the instant carried no evidence of direction | write the phase of each moving system as positions | 11.8 Nano Banana Pro |
An extension restarted the action | no stated state at the join | the extension mode with its direction; the join state read from the file | 13.7 Choosing the route |
The prop never moves, or the action is skipped | the keyframe shows the move finished | a state-0 keyframe (a one-change edit), then regenerate | 14.16 The moving shot |
The move overshoots (a lid closes fully) | no stopping point | write where it stops "and stays", and the composition it lands on | 14.16 The moving shot |
A dead first second, then the photograph wakes up | a statue keyframe, or an action that starts late | a keyframe just before the action; motion at once | 14.16 The moving shot |
The background warps | stacked or fast moves | one device; slower; a simpler background | 14.16 The moving shot |
An object appears from the edge of frame | something off frame was named | describe its light or its sound, not the object | 14.16 The moving shot |
A head swings through profile | only the landing was written | the short way, never past three-quarter; an end frame | 14.16 The moving shot |
Right ends, an invented middle | the frames are not exact relatives, or the path is not bounded | rebuild the pair; orientation in positive words; shorten the travel | 14.16 The moving shot |
A reversed clip gives itself away | something in frame moves on its own | remove it from the still, or go forward with an end frame | 14.16 The moving shot |
An exclusion is ignored ("does not tip or turn") | There is no negative field, and prose negatives are weak | State the positive state held all the way | 16.8 Kling |
One action is missing and the rest is good | The model skipped one event | Keep the take; generate the missing event alone from the same start frame and composite it (14.13) | 16.8 Kling |
An extension restarts the action, jumps at the join or steps in volume | a recap; the boundary state not matched; the maker's volume shift | open on the boundary state with no recap; trim at the join; cut on a sound; even the level in the mix | |
An extension goes the wrong way | direction words read the other way from Seedance | "Continue from the last frame of Video 1", or "Before the first frame of Video 1" | |
An extension drifts from its clip | Generative Extend invents the frames | shorten it, choose another handle, or generate longer | 32.8 Premiere and finishing |
D.8 Performance
A performance that shows the feeling, times a reaction wrongly, or resets at every cut.
You see or hear | Cause | Fix | Where |
Faces that act: an emotion word invites the face to perform | — | give the character a task, then visible cues in the body and face | 1.3 The camera is a model (Chapter 15) |
The wrong person speaks | a second face in the keyframe | move the listener outside the frame, or make them still | 9.10 The finished frame |
The actor advertises the emotion | — | give a practical task and cut the feeling word | 15.10 Directing performance |
The outcome shows before its trigger | — | limit the character to what they know at that moment; move the trigger earlier in the text | 15.10 Directing performance |
The listener waits mechanically | — | let understanding begin during the other line | 15.10 Directing performance |
The gestures illustrate every word | — | give the hands their own business | 15.10 Directing performance |
An empty pause | — | make it assessment, concealment or decision; otherwise shorten it | 15.10 Directing performance |
Every cut resets the actor | — | carry the previous action and its residue (Chapter 29) | 15.10 Directing performance |
The ensemble moves in unison | — | a different target and timing for each person | 15.10 Directing performance |
The whole body turns for every glance | — | separate the torso from the gaze | 15.10 Directing performance |
The wrong person reacts first | — | the responder keeps a task until the cue | 15.10 Directing performance |
Identities merge in a two-shot | — | simplify reflections and crowding; props that tell them apart; singles | 15.10 Directing performance |
D.9 Voice
Wrong words, dialect, numbers and accents in a voice line; a clone that differs; a line that will not lock.
You see or hear | Cause | Fix | Where |
Arabic spoken by a video model: no video model's own speech is a delivery route for Egyptian or Saudi Arabic | — | the voice comes from a voice actor or ElevenLabs; each project decides whether the exact take must ship or a new performance will do | 1.3 The camera is a model (Chapters 21, 22) |
Arabic speech, or any language but English | not evaluated by Google; not documented by MiniMax | never; an actor or ElevenLabs (Chapters 21 and 22) | 19.8 Veo and MiniMax |
A word sounds wrong | the take | re-lock the take | 22.5 The speaking shot |
A word changed or added | low stability, or a long line | one step toward higher stability; one sentence per take | 23.8 ElevenLabs |
The bracket is ignored, or the read is flat | the voice cannot play it, or stability is high | recast the voice; one simpler bracket; one step lower stability | 23.8 ElevenLabs |
A tag or direction is spoken aloud, or turns into a noise | the tag read as a sound cue; direction written as prose | describe the voice quality; fewer tags; direction to the picture prompt | 23.8 ElevenLabs |
Foreign or wrong dialect | the voice, not the text | recast from samples in the variety; audition the other model | 23.8 ElevenLabs |
One word fails: ق, ع, a vowel | this voice on this word | the ladder of 21.7 | 23.8 ElevenLabs |
IPA is ignored or wrong | IPA "can still vary by voice and phrase"; Arabic is not documented | generate again; another voice; an alias | 23.8 ElevenLabs |
A number or date misread | the normaliser | the spoken form, in words (21.9) | 23.8 ElevenLabs |
The accent drifts, the level shifts | a long line, or a voice not native to the language | one breath group per take | 23.8 ElevenLabs |
A clone sounds different on v4 than on v3 | v4 follows the source more closely | audition both; the ear decides; keep the model on the row | 23.8 ElevenLabs |
A newly generated line no longer matches a locked one | the model "may shift over time" | never regenerate a locked line; use the lock, and re-audition new lines | 23.8 ElevenLabs |
The performer's slip survives a conversion | the input decides the output | re-record the input | 23.8 ElevenLabs |
A word sounds wrong | the take | re-lock the take with the spelling repaired (Chapter 21); no more video credits | 24.8 Sync Lipsync 3 |
D.10 Sync
Lips against sound: the onset, the hold, a blurred mouth, a plate that talks too long or too little, a mouth that opens uninvited.
You see or hear | Cause | Fix | Where |
The mouth is already speaking in the first frame | a speech mouth was baked in | lips neutral or barely parted; the line begins in the clip | 9.10 The finished frame |
The mouth is lost in shadow | the lower face is unevenly lit | light the lower face evenly, and name the surface that bounces the light | 9.10 The finished frame |
A mouth opens uninvited on a silent clip | — | write the mouth: lips together and relaxed; give the listener no line that invites speech | 15.10 Directing performance |
A mouth opens uninvited on a silent plate, or speech comes back in English | A listening line invites speech; dialogue outside the five languages | Name the mouth ("lips together and relaxed"); never ask for dialogue; sound off if the plate is silent | 16.8 Kling |
The mouth does not follow; the line starts at once; a long hold after it | the sound is off; nothing is written before the line; the model re-timed the line | sound on; write the business first, or extend backward; trim the hold and buy fewer seconds next time | |
The mouth drifts in the second half | a three-quarter view, or two pauses | a frontal keyframe, one pause, a shorter line | |
Speech where there should be none, or extra words | "No dialogue" or "says only this line" missing | write it; sound off for a wordless plate | |
Sync right, face dead | the plate | re-plate (22.4) | 22.5 The speaking shot |
The first sound is missing | the audio window or its trim | restore the head silence and the onset | 22.5 The speaking shot |
Head and jaw fight the new line | the approved clip spoke the old line | a talking plate of the same shot | 22.5 The speaking shot |
The mouth is blurred or looks pasted on | the face is small, turned or occluded | a tighter, nearer-frontal plate with nothing at the mouth; else redesign the shot (22.3) | 22.5 The speaking shot |
Sync right, face dead | the plate | re-plate with more business and a talking mouth (22.4) | 24.8 Sync Lipsync 3 |
The first sound is clipped | the audio window | restore 200 ms of head silence | 24.8 Sync Lipsync 3 |
Lips start before or after the sound | the take's head does not match the plate's talk onset | pad the head to the measured onset, or trim the plate (22.4) | 24.8 Sync Lipsync 3 |
The mouth moves in a pause or after the last word | the plate talked longer than the speech | shorten the plate's talk; end the take after its tail | 24.8 Sync Lipsync 3 |
Motion reversed, repeated or sped up | bounce (the default), loop or remap | set cut_off | 24.8 Sync Lipsync 3 |
The line is clipped at the end | the take outlasts the plate | a plate 0.3 to 0.5 s longer than the padded take | 24.8 Sync Lipsync 3 |
Head and jaw fight the new line | the approved clip spoke the old line | a talking plate of the same shot | 24.8 Sync Lipsync 3 |
The mouth is blurred or pasted on | the face is small, turned or occluded | a tighter, nearer-frontal plate; else redesign the shot (22.3) | 24.8 Sync Lipsync 3 |
D.11 Music
A cue that enters wrongly, fades instead of ending, masks the voice, drifts in time or arrives with voices.
You see or hear | Cause | Fix | Where |
Music under "No BGM" | a sound effect turned into a score | an audio policy at the head and the tail listing the music terms (BGM, score, instrumental, melody, synth effects, ambient pad); or strip it in the mix | |
The cue enters too early | its placement | move the entry; regenerate only if no interval works | 25.5 Music and sound |
A jump at the seam between two phrases | the interval chosen | replace that region in the tool, or choose another interval | 25.5 Music and sound |
The music masks the dialogue | density in the speech range | duck it; then a reduced section, or "sparse under speech" | 25.5 Music and sound |
It fades instead of ending | no ending in the file | cut to a phrase with a real end; replace the ending | 25.5 Music and sound |
The export sounds different from the tool's player | gain, or stems doubled under the mix | compare the two aligned before changing the music | 25.5 Music and sound |
A muddy, crushed sound | a second limiter on a file that arrived loud | one limiter only, in the finished mix | 25.5 Music and sound |
Every candidate is the same generic "cinematic" bed | the brief names a genre and a mood, not a piece | rewrite as function, figure, roles and arc; move "cinematic, epic" to Exclude Styles | 26.8 Suno |
The figure is lost under the colour | too many instruments; the figure sentence is late | figure sentence first; cut the list to the lead and two supports | 26.8 Suno |
Voices, a choir or a hum appear | the switch alone, or an implied singing word | all three layers (26.3); add "backing vocals, ad-libs, vocal pad" to Exclude Styles; replace the section | 26.8 Suno |
It fades out, loops, or builds when it should not | no written ending, or "does not build" is missing or late | write the ending as behaviour and [End] in Lyrics; move "does not build" to the first sentence; make the lift harmony, not volume; replace only the ending | 26.8 Suno |
The cue fights the voice-over | the middle register is filled | "sparse under speech"; thin that section by arrangement; the level is set in the edit | 26.8 Suno |
A section you did not ask for, or a tag sung aloud | structure words in Style; a tag that is too long | move structure to Lyrics as tags; keep each tag to one to three instructions | 26.8 Suno |
Wrong words in one sung section | Suno drops or substitutes words | edit that section's lyrics and replace the section; check every sung word | 26.8 Suno |
The generation is refused | a name, a brand or a trademarked term in any field | remove it; describe the sound; check Title and Lyrics too | 26.8 Suno |
A click or jump at the seam | the boundary is mid-phrase | drag the boundary to a phrase edge or a decay | 26.8 Suno |
The downbeat has moved | a seam changed the timing | re-spot the edit; widen the region if needed | 26.8 Suno |
Stems do not add up to the full mix, pump, or carry a ghost voice | they are separations | use them as parts; keep the full mix as reference; use the original mix outside the dialogue and a reduced section under it | 26.8 Suno |
Drift late in the cue, or a locked cue that sounds stiff | the tempo is approximate; or a breathing cue was forced onto a grid | Manual BPM for a grid cue, or cut at phrases; return a breathing cue to its own timing | 26.8 Suno |
A regional palette reads as a postcard | a reflex instrument in the roles | remove that role; keep pulse and harmony; compare | 26.8 Suno |
It sounds older, or ignores the chunks | it ran on the default music_v1 | name music_v2_5 | 27.8 Eleven Music |
A voice appears | force_instrumental does not apply to plans | "vocals" in the negative styles of every chunk | 27.8 Eleven Music |
The ending cannot be repaired | it was not stored at generation | upload the approved file to store it, then repair | 27.8 Eleven Music |
Chunk 1's palette leaks into a changed ending | the first chunk shapes all that follow | an audio-reference chunk for what must stay; lower context adherence on the chunk that should change | 27.8 Eleven Music |
A click at a chunk joint, or a refusal | the boundary sits mid-phrase; or a name in a style | move the reference range to a phrase edge; describe the sound instead of naming it | 27.8 Eleven Music |
a generic bed | a vague prompt | instruments, BPM, key, mood and structure | 28.8 Sound generators (Lyria) |
voices appear | a hint of singing in the brief | "Instrumental only, no vocals", remove any sung wording, audition another run | 28.8 Sound generators (Lyria) |
blocked prompt | the safety filter (an artist's voice, copyrighted lyrics) | describe the sound, name no one | 28.8 Sound generators (Lyria) |
D.12 Sound
Effects and beds: timing to the contact, size and material, tails and loops, and a track you did not ask for.
You see or hear | Cause | Fix | Where |
A voice or music appears | the track cannot be switched off and the sound line left room | name each sound and refuse the rest; strip the track in post; or use a tool with a switch (Chapters 16 to 18) | 19.8 Veo and MiniMax |
The sound drifts off the contact | the event was vague | write the event and its sound as one moment; shorten the clip | 19.8 Veo and MiniMax |
A sound is the wrong size or material | a vague prompt | name what struck what, with force, scale and space (25.7) | 23.8 ElevenLabs |
A sound effect carries music or a voice | the prompt invited it | end with "No music, no voices"; make the event alone | 23.8 ElevenLabs |
Hits late or early | timing | slide the transient to the contact frame | 25.13 Music and sound |
Several contacts | repetition | a single-hit candidate, or trim → record | 25.13 Music and sound |
Ring or tail too long | tail | trim or fade; "damped, short dry decay" → a fuller glass, a smaller room | 25.13 Music and sound |
Too heavy | perspective | reduce the force and weight words → record it lighter | 25.13 Music and sound |
A close sound on a wide shot | perspective | distance and room in post → or was it a chosen subjective emphasis? | 25.13 Music and sound |
Loop bumps | repetition | a longer file → a library bed | 25.13 Music and sound |
A musical hum in a bed | source | a simpler request: the fan alone, "no tones" → a library or recorded bed | 25.13 Music and sound |
A trailer thud | source | give "impact" a force and a scale; remove "cinematic" → record | 25.13 Music and sound |
The bridge starts on the cut | placement | start it under the outgoing frames | 25.13 Music and sound |
An effect masks a line | level | lower or move the effect → a different take | 25.13 Music and sound |
contacts off the action | the guide was performed watching the face | record the guide again, watching the object; keep the words | 28.8 Sound generators (Firefly) |
the wrong material, or an ambience like drumming on pots | the words | rewrite the words; keep the guide | 28.8 Sound generators (Firefly) |
several sounds mixed | more than one sound in the prompt | one sound per generation, layered on tracks | 28.8 Sound generators (Firefly) |
several taps or a long tail | duration too long, or "impact" without a scale | a shorter duration, "once", a size | 28.8 Sound generators (Stable Audio) |
the attack moves or doubles | noise level too high | bring it back towards the recording, or keep the recording | 28.8 Sound generators (Stable Audio) |
a region repeats the original | mask too small | widen the mask | 28.8 Sound generators (Stable Audio) |
A "silent" clip has sound | sound was left on, or the model is Kling Turbo | type sound off on Kling 3.0, or strip the track in the edit | 33.8 Higgsfield |
D.13 File
Sizes, aspects, frame counts, uploads, charges and job records; import and export.
You see or hear | Cause | Fix | Where |
Repeatability: no seed on the video models; every take is new | — | keep every original file; record every job the day it runs | 1.3 The camera is a model (Chapters 8, 33) |
The frame comes back in the wrong shape | an aspect written in the prompt fights the setting | keep the aspect out of the words | 9.8 The finished frame |
The frame comes back in the wrong shape | an aspect written in the prompt fights the setting | set it in the interface and keep it out of the words | 11.8 Nano Banana Pro |
A small, soft frame | the defaults (low, 1k) were left | write all four settings and re-run | |
A fringe on black or white | a translucent edge made against another ground | denser material in the words; choke the matte in post | |
A floor, shadow or checkerboard baked in | transparency not asked for in both places | set the background and say it in the words; try 1:1; check the file | |
A preset suggestion instead of a job | The platform intercepted the request | Resubmit with the decline field; no charge for the interception | 16.8 Kling |
An edit is a few frames short | the source is not 8n+1 frames | trim the source; or use a reference job with a set duration | |
A complex job is unstable, or a changed aspect fills with invention | a reference task and an edit in one prompt; an edit cannot change the canvas | a reference pass, then an edit pass; for an aspect change, a reference job with a frame-by-frame plot description | |
A rejected upload (a 400 error) | a JPG with unusual chroma subsampling | convert with ffmpeg, or use PNG | |
The job is rejected | a start or end frame went in with a reference | remove the frame, or drop the references | |
The clip is 8 s and the price doubled | the default | set 4 | 19.8 Veo and MiniMax |
Soft detail at 2K | the 2K may not be real detail | judge at 200 %; choose another route for that plate | 19.8 Veo and MiniMax |
Four videos billed | batch above 1 | batch 1 | 19.8 Veo and MiniMax |
Open Project refuses the XML | it is an import, not a project | File > Import | 32.8 Premiere and finishing |
Source out-point drifts by a frame | end typed, not computed | end = start + (out − in) | 32.8 Premiere and finishing |
An effect or fade is missing after import | it did not translate | read the translation log; rebuild it natively | 32.8 Premiere and finishing |
Charged double | ByteDance above 30 fps | keep the source rate | 32.8 Premiere and finishing |
A panel result cannot be previewed | the Topaz panel has no real-time preview | run a 2-second test first | 32.8 Premiere and finishing |
Timeout, no job number; a wait that ran out | the outcome is unknown; the job may still be running | ledger and history first; look it up by number; never resubmit | 33.8 Higgsfield |
Rate-limit error, or part of a batch failed | some requests may have run | ledger and history; keep the numbers you got; resend only the missing items, in a smaller batch | 33.8 Higgsfield |
"Completed", no usable file | an error behind the status | read the job record; report it; no silent resend | 33.8 Higgsfield |
A preset suggestion, an "unlimited" question or a billing card | not a job | record it; for a preset, resend with the decline field; for the others, stop and ask | 33.8 Higgsfield |
A setting rejected or ignored; a role renamed in adjustments | the contract changed, or the platform mapped your role | re-read the contract; record the role actually used; never retype a setting from memory | 33.8 Higgsfield |
The model "hears nothing" | the upload's type did not match, or it was never confirmed | upload the matching type; confirm; use the new id | 33.8 Higgsfield |
The route refuses the approved frame | the wrong route for the shot | stop and ask (a lock); never switch models yourself | 33.8 Higgsfield |
A wrong checksum; a frame size a few pixels off | the file was read mid-write; the platform's transcode | probe after the download finishes; conform in the edit | 33.8 Higgsfield |




Comments