Chapter 26 · Suno
Part 5 · Music and sound
Suno is the music generator for score cues, temp tracks and rhythm references. It behaves like a session band that has read your page once: it plays what you describe and invents whatever you leave open. This chapter teaches the brief in Suno's own fields, as one directed paragraph built on a theme, with the refusals in their own box and no name anywhere Suno reads. It then teaches the controls, the repairs and the limits, and ends with five cues at the standard.
In this chapter
What Suno is today, and when to reach for a composer, a library or another tool
The fields Suno reads, and what it does with a name
The Style paragraph in eight parts, and which field each part goes in
Five cues at the standard: beauty, food, telecom, automotive, and Mastorna's opening cue in corrected form
The settings, the repair operations, the checks, the failures, and the limits, prices and rights
Before you start. Chapter 25 decides whether music enters, what it does, and whether the cut follows the cue or the cue follows the cut. This chapter is how to say it to Suno. Chapter 34 covers rights across the whole job.
26.1 What it is
Suno makes a complete, mixed piece of music from a written description: instruments, arrangement and, unless you switch it off, a singing voice. Its current model is v6, released on 9 Sep 2026, in three variants (checked 30 Sep 2026). Suno's help describes v6 as "our most advanced model", v6-wild as "built for experimentation", and v6-mini as "a faster, more efficient way to create". v6 and v6-wild need a Pro or Premier plan; v6-mini is open to every user, free included. All three make pieces up to eight minutes long. Every generation returns two songs and costs 10 credits.
It runs on suno.com, on your own paid account, with its own credits. It is not on Higgsfield, and Higgsfield's audio catalogue is not part of any music route in this book (Chapter 33).
Reach for Suno when the job is a score cue that meets picture, a temp track laid under a rough cut so the cut can be judged, a rhythm reference that sets cut points before picture exists (Chapter 25), or a sketch to find a direction in minutes.
Reach for something else when:
The job | Use | Where |
The client needs an exclusive score, guaranteed originality, or timing exact to the bar | a composer or a licensed library track | |
Timed sections such as a change at second 24, online delivery only | Eleven Music, written as a plan | |
A single sound event or a room bed | a sound-effect generator or recorded Foley | |
A sung line in Egyptian or Saudi dialect that must be right | a singer, or an actor and a recording session |
Suno also has a Sounds tool (a beta, edited 9 Sep 2026) for one-shot samples and loops; sound events follow the rules of Chapter 25.
26.2 How it reads what you give it
Suno reads fields, not a manifest. Nothing in it binds one input to another by number or by label. The field a sentence sits in decides what Suno does with it, so most failures begin with a sentence in the wrong field.
Input | What it owns | Notes (checked 30 Sep 2026) |
Style of Music | the identity of the whole piece | one paragraph; about 1,000 characters on v6 (26.9) |
Lyrics | the running order, as section tags in square brackets, and any words to sing | up to about 5,000 characters; tags are a convention |
Instrumental | whether a voice may appear | a switch in the interface, not a sentence |
Exclude Styles | what must not be in the piece | Suno's help: "Enter any information (instruments, etc) that you do not want in your track" (edited 19 Dec 2025) |
Title | the file name | not sung; keep it neutral |
Weirdness, Style Influence | how far Suno departs from, or holds to, your text | Weirdness runs from "Safe" to "Chaos" with 50 % the normal result; Style Influence from "Loose" to "Strong" (help edited 3 Jun 2025) |
Variety | how much Suno rewrites your style tags | new on v6; it introduces variety "by adjusting and updating your style prompts", and Suno says to "reduce the Variety slider to 0" to retain full control of your style tags |
Audio Influence | how far uploaded audio steers the result | appears only when audio is attached |
Source audio and references | melody, timbre, arrangement | v6 accepts songs, playlists, audio, images and video as references in one prompt; how it weighs each is not documented |
Names are blocked. Suno's moderation help lists what can stop a song generating: "The names of well known artists or people", "Copyrighted or trademarked terms", derogatory or defamatory terms, excessive profanity, and other inappropriate topics (edited 19 Dec 2025). Treat every field as covered, Title and Lyrics included. A composer, a singer, a film, a product and a brand never appear in anything Suno reads. The brand goes in the cue sheet, and the description does the work the name would have done (26.3).
Refusals live in one field. Prose refusals inside the Style text are unreliable; Exclude Styles is the field Suno built for them. Only refusals that hold for the whole cue go there: "choir" excluded from the box is excluded from the ending too. A refusal that holds for one section stays in that section's tag.
Tags and numbers are guidance. Suno documents no tag parser; creators report that short tags lift control and that a long tag may be sung aloud, so confirm every tag by ear. A tempo in the prompt is approximate (Suno's tempo help says a generated tempo drifts "just like live music"), and seconds are not obeyed reliably. Timing lives in the edit (Chapter 25).
What it ignores or misreads. "Cinematic", "epic" and "trailer music" name the average, and Suno delivers the average. A negative inside a paragraph is weak. Vocals appear unless the switch is off, and some genre words invite them: gospel brings a choir, "cinematic" brings chant. Without a written ending a cue tends to fade or loop.
Figure 26.1 — The brief stays in the cue sheet; only its translation is pasted, and everything the interface controls is set there. The windows in seconds and the brand never reach Suno.
26.3 The prompt, in this tool's dialect
Chapter 25 decides what the cue does: whether music enters, its function in story terms, the windows measured on the picture, the entry beat and the ending. This section turns that brief into Suno's fields. Use Custom mode for every cue that meets picture. Simple mode, one description box, is for a sketch and nothing else.
The spine: a theme, not a mood
Build every cue around a theme, as 25.3 teaches: one figure and what it does in this cue, never a mood, and never "trailer music", "cinematic suspense", "epic orchestral", "haunting score" or "in the style of" a composer, because each names the average of a genre and Suno delivers the average. A 15 or 30 second spot carries one small figure, three to five notes, that survives the cutdown. Describe it by contour, rhythm and instrument ("a rising step, then a falling fourth, on celesta"), and say what it does in this cue: stated alone, broken into cells, deformed, doubled, thinned under a voice. No maker states that Suno holds a named motif across a cue, so listen for whether it survives (26.7).
The order of the paragraph
Write the Style field as one directed paragraph, the most important content first, in the eight parts of the brief (25.3): form and function; the figure and what it does; roles (one lead, three to five instruments each in the singular, what accompanies, what is absent); arc, entry and ending; harmonic identity; performance and room; speech space; and last, the tempo as an approximate number. Example 26.1 shows the order at work.
Count the paragraph and keep it under about 950 characters. Every refusal has a positive twin in the text: "no sentimental romantic sweep" in Exclude Styles pairs with "the melody presented plainly, with emotional reserve" in the paragraph. Feeling is carried by behaviour and by the picture, never by an adjective alone.
Write the ending. Say how the cue stops: a clean stop on the beat, a held note that decays to near silence, a tail of a stated length for editorial overlap. Say what it must not do: "never a fade". A cutdown is a new structure that keeps the figure and the ending, so brief a 15 second cue as its own cue.
Leave the speech range open by naming where the voice sits ("middle register nearly empty for a close female voice-over"). The level is set in the edit.
Regional sound is a set of roles (25.3). To imitate a singer's manner, describe it (register, grain, ornament on phrase endings, era, instruments, rhythm) and never name the singer. If the client's brief cites a singer, that reference stays in your document.
The Lyrics field
For an instrumental cue the Lyrics field holds the running order. Open with [Instrumental], write one tag for each picture section with one to three instructions in it, and close with [End]. Keep key changes, solos and section events here, not in Style, where they are often ignored. Turn any arrow-and-slash structure into separate tag lines. A conditional refusal ("no choir during the verse") belongs in its section's tag.
Instrumental means three layers
Set the Instrumental switch on, open the Lyrics field with [Instrumental], and put the vocal words in Exclude Styles. An empty Lyrics field alone can still produce voices. Keep the brief free of implied singing ("a choir sings the slogan", "chant-like").
The capitals in the examples name Suno's fields. Paste only what follows each label, never the label: a pasted "STYLE:" becomes part of the prompt. Settings go in the controls, never in the text.
26.4 Templates
Written for this book · template; not run · finished grade, a score cue in Custom mode
Template — the finished cue
STYLE: A [adjectives of behaviour] [form] for [what it does for the picture], built on one [size] figure, [contour and rhythm], first stated [alone / by instrument]. The figure [what it does: returns / breaks into cells / rises an octave / thins under the voice] and is never [what it must not become]. Roles: [lead] leads; [two or three supports]; [what accompanies only after the figure is established]; no [what is absent]. [Mode or key], [one characteristic interval], [cadence behaviour]. Recorded [close / wide], [dry / roomy], [what detail is audible]; [steady / rubato], about [n] BPM. The [middle] register stays [nearly empty] for [the voice]. It ends [on the tonic, a held note, a decay to near silence, a clean stop on the beat], never a fade.
LYRICS: [Instrumental] [Intro: ...] [Phrase: ...] [Lift: ...] [Outro: ...] [End]
INSTRUMENTAL: on
EXCLUDE STYLES: [vocals, choir, and up to three sounds this cue could turn into]
TITLE: [neutral file name, no brand]
CONTROLS (set in the interface, not typed): model [v6] · Weirdness [n] · Style Influence [n] · Variety [0] · plan [Pro / Premier]Written for this book · template; not run · test grade, a temp cue or a direction check
Template — the test cue
An instrumental [character] piece for [film use]: [lead] over [support], [how it starts], [how it moves], ending on [clean stop / tail]. Two sentences at most.
CUSTOM MODE, INSTRUMENTAL: on, EXCLUDE STYLES: vocals. No tags, one generation. Label the file TEMP.A test cue proves a direction in two minutes. It is labelled TEMP in the cue sheet, the rough cut it sits under carries the slate "DRAFT — NOT FOR APPROVAL" (Chapter 31), and it never goes to a client as a cue. Rebuild the winner as a finished cue.
Written for this book · template; not run · the repair and continuation lines, one per operation
Template — Replace Section, Extend, rights line
REPLACE SECTION (select the region with a bar of lead-in): Keep the existing tempo, harmony, [lead], and the transitions into and out of this region. [One change]. Keep the ending as it is and keep the cue instrumental.
EXTEND (from the last approved bar): Next: [what happens musically], then [the ending]. Name the figure: [the figure, in one phrase].
LYRICS from the extension point: [Section: ...] [End: ...]
RIGHTS LINE (34.10): [cue id] v[n] (a stem: separated by Suno [date]) · kind cue · source class generated cue · maker and product Suno · plan [plan] · model [as recorded] · medium [online / TV / film / radio / games] · generated [date] · downloaded [date, on which plan] · uploads [none / file, origin] · terms read [page, date, saved copy] · consent [n/a, or record, scope] · ships [yes / no / pending, and what it waits for]26.5 Examples at the standard
Each cue is complete against the eight parts of 26.3: a Style paragraph under 950 characters, the refusals in Exclude Styles, the running order in Lyrics, and the controls in the label. The picture windows and the brand stay in the cue sheet. The slider values are start points for a cue that must hold its plan, not maker defaults (26.6).
Written for this book · beauty and personal care · a 30-second premium hair-repair film, Cairo studio, voice-over from 3 s · Suno v6, Custom mode, Instrumental on, Weirdness 30, Style Influence 80, Variety 0 · written for this book; not run
Example 26.1 — a three-note figure under a close voice
STYLE: A poised, minimal chamber cue for a thirty-second premium hair-repair film, built on one small three-note figure, a rising step then a falling fourth, first stated alone on a glass-toned celesta. The figure returns twice and is never stretched into a tune: once over a soft felt-piano bass line, once an octave higher. Roles: celesta leads; felt piano and one warm sustained cello note support; a single brush on a snare marks the pulse only after the figure is established; no pad, no drum kit. F major, one gentle lift to D minor at the second statement, a plain resolved cadence. Recorded close and dry in a small room, hammer and finger noise audible; steady and unhurried, about 90 BPM. The middle register stays nearly empty for a close female voice-over. It ends on the tonic, one held cello note and a short natural decay into room tone, never a fade.
LYRICS: [Instrumental] [Intro: celesta alone states the figure] [Phrase: felt piano bass enters, brush after the figure] [Lift: figure an octave higher, one lift to D minor, same ensemble] [Outro: tonic, held cello note, short decay] [End]
INSTRUMENTAL: on
EXCLUDE STYLES: vocals, choir, orchestral swell, heavy drums, long fade
TITLE: haircare_30s_A_v1The figure is described, and kept small: contour and instrument, then "never stretched into a tune". The lift is harmony, not volume or a drum fill, so Suno has nothing to build with.
The roles write the arc: the brush enters only after the figure is established, so no seconds are needed.
Speech space is named in the Style text; the voice-over window (3 to 24 s), the product hold (26 to 30 s) and the 15 second cutdown, a second cue with the figure and the ending, stay in the cue sheet.
Written for this book · food and drink · a 30-second film for a cold hibiscus drink, Cairo rooftop, cut on the beat, voice-over from 5 s · Suno v6, Custom mode, Instrumental on, Weirdness 30, Style Influence 80, Variety 0; Manual BPM in Studio · written for this book; not run
Example 26.2 — a grid cue with a hummable figure and a button
STYLE: A warm, bright street-café cue for a thirty-second film about a cold hibiscus drink on a Cairo rooftop, built on one hummable four-note accordion figure that rises a third, repeats its top note and falls back. Roles: accordion leads; a maqsum pattern on darbuka and riq carries the groove; upright bass plays pizzicato; hand-claps join only for the final statement. It starts with riq and finger clicks alone, adds darbuka and bass when the first ice hits the glass, states the figure on the pour, drops to one held chord over the groove while a voice speaks, then states the figure in full once more. A major mode with one flattened seventh turn at the second statement; the cadence resolves plainly on the tonic. Live and slightly loose, recorded close in a small room, no drum machine. Steady, about 100 BPM. The last bar is one chord and one clap on the beat, a clean stop.
LYRICS: [Instrumental] [Intro: riq and finger clicks alone] [Groove: darbuka and bass enter, accordion states the figure] [Under the voice: one held chord, groove only] [Final statement: accordion full, claps join] [End: one chord, one clap, clean stop]
INSTRUMENTAL: on
EXCLUDE STYLES: vocals, drum machine, EDM drop, big-band brass, long fade
TITLE: hibiscus_30s_A_v1The regional sound is written as jobs: a maqsum pattern on darbuka and riq carries the groove. No singer, song or brand is named.
The arc is told against picture events ("when the first ice hits the glass", "on the pour"), so the edit sets the times.
The ending is a button: "one chord and one clap on the beat, a clean stop". The cut follows this cue bar by bar, so lock it with Manual BPM (Studio, so Premier) and split the percussion for a cue map (26.6, Chapter 25).
Written for this book · telecom and tech · a 30-second launch of an unlimited data plan, Riyadh, Saudi voice-over from 6 s · Suno v6, Custom mode, Instrumental on, Weirdness 30, Style Influence 80, Variety 0 · written for this book; not run
Example 26.3 — a five-note riff that thins under the voice and comes back
STYLE: A bright, confident electronic-pop cue with a Gulf hand-clap groove, for a thirty-second film launching an unlimited data plan, built on one five-note oud riff that climbs three steps and falls back. Roles: a close oud states the riff; claps and one hand drum keep a swinging six-beat pulse felt in two groups of three; a round sub-bass and one plucked synth double the riff only in the final statement; no pad. The riff is stated alone over claps, doubled when the phone lights up, thinned to two repeated notes while a voice speaks so the middle register is empty, and returns in full at the end. G minor with a raised seventh, a nahawand colour; the cadence lands plainly on G. Tight and dry, crisp claps, no lush reverb. About 104 BPM. The last clap is a clean stop with a half-second tail for a sound logo.
LYRICS: [Instrumental] [Intro: claps alone, then the oud riff] [Build: sub-bass enters, one pluck doubles the riff] [Under the voice: riff thinned to two notes, claps continue] [Final: full riff with pluck, hand drum] [End: last clap, half-second tail]
INSTRUMENTAL: on
EXCLUDE STYLES: vocals, trailer risers, EDM drop, orchestral strings, long fade
TITLE: dataplan_30s_A_v1The modern and the regional sound are both roles: claps and a six-beat pulse for the Gulf, sub-bass and one pluck for the tech, each with a job and an entry.
Speech space is an arrangement decision ("thinned to two repeated notes while a voice speaks"), which is stronger than asking for "space for a voice".
The tail is stated in length. The brand's sound logo is added in the mix, never asked from Suno.
Written for this book · automotive and luxury · a 45-second film of an electric luxury SUV crossing dunes at dawn, sparse Saudi voice-over · Suno v6, Custom mode, Instrumental on, Weirdness 25, Style Influence 85, Variety 0; a breathing cue, no Manual BPM · written for this book; not run
Example 26.4 — a slow cue with a figure of three long notes
STYLE: A slow, spare, restrained cue for a forty-five-second film of an electric luxury SUV crossing dunes at dawn, built on one figure of three long notes, a rising fifth then a step down, stated alone at first. Roles: felt piano leads; low bowed strings hold beneath it; a deep sub-bass breathes under the last third; one frame drum strikes once per bar, but only after the car first appears; before that the piano and strings have the space to themselves. The figure returns over the strings when the car appears and lifts an octave at the reveal. D Dorian, open fifths, a suspended fourth left unresolved until the final chord. Recorded wide, dry and quiet, felt hammers audible; a breathing tempo near 60 BPM with rubato allowed. The middle register stays empty for a low male voice-over. It ends on one chord held for four seconds that decays to near silence.
LYRICS: [Instrumental] [Intro: felt piano alone states the figure once] [Approach: low strings hold, the figure returns] [Reveal: figure an octave higher, sub-bass breathes, frame drum once per bar] [Outro: final chord held, decays to near silence] [End]
INSTRUMENTAL: on
EXCLUDE STYLES: vocals, braams, oud swell, epic drums, long fade
TITLE: suv_45s_A_v1Luxury is restraint: few roles, a wide and dry recording, a suspended fourth that waits for the last chord, and an ending that decays instead of swelling.
Percussion is withheld until a picture event, which gives Suno the arc without a time stamp.
The refusals name only the reflexes this film could fall into: trailer braams, an oud swell on the reveal, epic drums.
From Mastorna, corrected · the opening cue of the trailer, a restrained chamber processional that introduces the theme before the rupture · Suno v6, Custom mode, Instrumental on, Weirdness 30, Style Influence 80, Variety 0; a breathing cue, no Manual BPM · corrected for this book; not run
Example 26.5 — the theme introduced plainly, with an ending and room for speech
STYLE: A restrained 1960s Italian chamber-orchestral processional that states its theme plainly before the rupture, built around one recurring singable theme first stated by solo cello in D minor, 3/4, with a childlike upward minor-third rise and a single chromatic wound that destabilizes the cadence; supported by muted low strings, bass clarinet, bassoon, and very soft brushed timpani, with the melody presented plainly and with emotional reserve, as if public travel ritual is already shaded by death; harmony avoids heroic cadences and uses suspended dominant color, modal minor, and one recurring semitone abrasion; performance stays dry, close, and period-acoustic, in a lightly worn studio room with minimal reverb. It opens on one unaccompanied cello note, states the theme once without building, and ends on a held low note that decays to near silence. The middle register stays open for faint speech. Unhurried, about 60 BPM, rubato allowed.
LYRICS: [Instrumental] [Theme: solo cello, plainly] [Support: muted low strings, bass clarinet, bassoon, brushed timpani] [End: held low note, decays]
INSTRUMENTAL: on
EXCLUDE STYLES: synth pulse, braams, trailer percussion, parody circus, sentimental romantic sweep, glossy symphonic swell
TITLE: mastorna_cue1_v1The theme is a musical fact (a singable cello theme, a childlike rise of a minor third, one chromatic wound), so the cue has something to keep and something to deform in later cues.
The refusals moved into Exclude Styles, and the paragraph kept their positive twins: "dry, close, and period-acoustic", "emotional reserve".
The period is described as music (chamber orchestra, a worn studio room). The period's playback character is applied in the finished mix (Chapter 35).
Kept in the cue sheet: the cue enters as the fanfare ends and is drowned by the storm, so its last phrase must be whole by then.
26.6 Settings
Set every control on purpose and write it in the cue sheet with the take (checked 30 Sep 2026).
Setting | Value to use | Why |
Plan | Pro or Premier from the first generation | commercial permission follows a paid plan and a permitted download (26.9); a cue with a free-plan ancestor cannot be made client-safe later |
Model | v6 for picture cues; v6-wild only to explore; v6-mini for sketches | write the model on every candidate |
Mode | Custom | separate fields for style, running order, switch and refusals; Simple mode cannot separate them |
Instrumental | on for every score cue | Suno sings unless told otherwise |
Weirdness | low; start at 25 to 30 | 50 is "normal"; a cue that must hold its plan wants less surprise. Suno gives no other value; record yours |
Style Influence | high; start at 80 to 85 | holds the result to your text |
Variety | 0 for a written brief | at 0 your style tags stay yours |
Audio Influence | only with source audio; raise it as a test | it appears only when audio is attached |
Max Mode | off for a 30 second cue | Suno recommends it for songs longer than two minutes, for covers that must stay close, and for consistent vocals; it costs more credits |
Exclude Styles | the cue's refusals, one to five sounds | Suno states no limit on their number; keep them to sounds this cue could turn into |
Length | a whole cue, trimmed at a phrase in the edit | up to eight minutes per generation |
Title | a neutral file name | never a brand or a song title |
Download | the candidate you choose, on the paid plan; WAV if the plan gives it | downloads are counted; MP3 on every plan, WAV on Pro and Premier (help edited 9 Sep 2026) |
Run the brief twice as written before changing a word. One generation is two songs for 10 credits, so two runs are four candidates and 20 credits. An identical repeat measures how much Suno varies; a changed prompt measures your change. Then plan three candidates: A as briefed, B and C each changing one thing (the lead, the tempo feel or the ending; never all three).
Defaults that bite. Vocals unless the switch is off. Fades and loops unless the ending is written. Downloads are counted. Free-plan songs stay non-commercial. A file named "tempo locked" proves nothing until you check it.
The operations: the smallest one that fixes it
Every operation makes a child version from an approved parent; keep the parent, name the child, log what changed. If 70 to 80 per cent of a cue works, repair it before you regenerate: a regeneration renegotiates the parts you approved.
Figure 26.2 — Choose the operation by the problem. The smallest one that fixes it is the right one: a region is repaired before a song is regenerated.
Extend keeps the part you choose and generates a continuation or an ending; "Get Whole Song" merges the new ending with the original (help edited 19 Dec 2025). Nothing promises a seamless join or an exact end beat. Extend from the last bar you approve, not from the end, and name the figure in the extension line or the extension can forget what it continues. Extend only songs you made, on a paid plan: Suno's help advises against assuming ownership of an extension of a song you did not make, or one that began on the free plan (edited 7 Jan 2026). For a long cue, generate the core and extend.
Replace Section (Pro and Premier; help edited 19 Dec 2025) regenerates a highlighted region, returns two versions to compare and, once you choose, a new whole song. Write the replacement as "keep … change one thing", naming what must survive first, and ask for a smaller ending, not a new arrangement. The Song Editor (help edited 30 May 2025, before v6; confirm each control on your screen) adds an Edits Library, a draggable boundary on the waveform, and adding, removing and rearranging sections. Drag the seam to a phrase edge or a decay, and write down which of Replace Section, the Song Editor and Studio made each repair.
Remaster (help edited 9 Sep 2026) makes variations of a clip at three strengths: Subtle ("very close to the original"), Normal, the default ("keeps duration and style the same"), and High ("more noticeable differences, including possible changes to musical elements and vocals"). It is for balance and polish, not for changing lyrics or style. Use Subtle on a cue you already approve.
Cover (help edited 27 Sep 2024, a beta) makes a new arrangement and works only on songs you created. Judge it on what it kept before what it added: if the figure is lost, it is a new song. Suno recommends Max Mode for covers that must stay close. What it keeps on v6 is not documented.
Stems and tempo. Suno calls its stems separations of the finished mix (help edited 12 Jun 2026). They are not the original tracks, so they will not add back to exactly the full mix and a part can carry bleed. Auto Split gives up to 12 stems for 50 credits; Split from Mix extracts one part and the remainder for 10 credits per extraction (20 for the pair); Advanced Split lets Premier choose from nearly 100 instruments. Extracting an instrument that is not in the song can spend credits and return nothing. Name every stem file "separated (Suno, date)" and audition each alone and recombined against the full mix. For a cue the cut will follow on a grid, set Manual BPM in Studio's Project Tempo and the receiving program to the same tempo (help edited 8 Oct 2025, before v6), then compare the locked version with the original. A breathing cue keeps its own timing.
Upload, Add Vocals, Voices, Personas, Custom Models. An upload starts a cue from audio you own. Suno takes a broad licence over what you upload (26.9), so upload only neutral material you own or are licensed to use: never client-confidential music, and never anyone's voice without written consent. Add Vocals (a beta, edited 21 Jul 2025, Pro and Premier) works with uploaded instrumentals, generated instrumentals and stems, and has an Audio Strength slider for how far the new song keeps the old instrumental. A Voice is your own verified singing (a recording of 15 seconds to 4 minutes, singer 18 or older, with an affirmation that you hold the rights; help edited 1 May 2026). A Persona saves the vocal and style character of a song and is "public by default" (help edited 19 Dec 2025), so never save one from client material. A Custom Model needs six songs you own the rights to, up to three per user, Pro and Premier, private (help edited 9 Sep 2026). None of the four is a scoring tool: a cue for picture is instrumental and its identity is its theme.
26.7 Checks before you sign
Listen in this order. A cue is signed only after all of it.
Listening checks
Alone, on headphones, the whole length. No voice, hum or breath under the mix, and no ghost of a vocal you excluded.
The figure. Hum it back after one hearing. It returns where the paragraph said it returns, and it is not stretched into a tune.
The ending. It stops as written: a button, a held note, a decay. Count the tail. If it fades or loops, the ending was not written strongly enough (26.8).
Against the picture, with the scratch voice-over laid on. The speech range is open; nothing masks a syllable; the cue enters and leaves on the beats the cut asked for.
Across every join. After Extend, Replace Section or a boundary drag, listen to the seam twice: no click, no jump in tone, the downbeat has not moved.
The tempo, measured from the file. Write the cue map from the file, not from the prompt's BPM: tempo, accents and section starts in seconds, checked at the start, the middle and the end.
Stems, if any. Alone and recombined against the full mix.
The file. Probe the format, sample rate, bit depth and length of the first file of each kind you export; Suno's pages disagree on Studio's export format (26.9).
The record. The rights line is filled (34.10; template in 26.4; terms in 26.9); the watermark and metadata are untouched; the title carries no brand.
Loudness is not judged here. Suno's output is not set to a delivery level. Loudness is measured in the finished mix against the specification the delivery owner names: EBU R 128 asks −23 LUFS and −1 dBTP true peak, ATSC A/85 −24 LKFS and −2 dBTP (25.14, 30.6). A cue that sounds loud beside the voice-over is fixed in the mix, not by another generation.
26.8 Failures and fixes
You hear or see | Cause | Smallest fix |
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 |
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 |
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 |
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 |
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 |
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 |
Wrong words in one sung section | Suno drops or substitutes words | edit that section's lyrics and replace the section; check every sung word |
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 |
A click or jump at the seam | the boundary is mid-phrase | drag the boundary to a phrase edge or a decay |
The downbeat has moved | a seam changed the timing | re-spot the edit; widen the region if needed |
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 |
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 |
A regional palette reads as a postcard | a reflex instrument in the roles | remove that role; keep pulse and harmony; compare |
The retry ladder, smallest lever first.
Listen again and name the fault by region.
Run the brief unchanged once more, to measure Suno's variation.
Change one sentence, or move a refusal into Exclude Styles.
Move the sliders: Style Influence up, Weirdness down, Variety 0.
Replace the failing section, or extend from the last approved bar.
Remaster at Subtle if the song is right and the balance is not.
Change the model for a different reading (v6-wild only to explore).
Rewrite the brief. Two rounds of one-change candidates that still miss the function mean the brief is wrong, not the prompt.
Stems and the edit: keep the good part, rebuild the rest in the mix.
A composer or a licensed library track.
If repairs keep spreading into neighbouring sections, stop patching and return to 26.3.
26.9 Limits, prices and rights
Limits (checked 30 Sep 2026).
Length. Up to eight minutes per generation on v6. Voices and themes drift in long pieces; generate the core and extend.
Field lengths. Suno publishes none. A third-party reading of the live v6 composer on 7 to 9 Sep 2026 reports 1,000 characters for Style and for Exclude Styles, 5,000 for Lyrics and 100 for Title. Trust the counter on your own screen; keep a Style paragraph under about 950 characters.
Uploads. The pricing page lists up to 8 minutes on the free plan and up to 30 minutes on Premier; the help article on uploads (30 May 2025) says 60 seconds on the free plan and 8 minutes on paid plans. Test the limit on the day.
Stems. Up to 12 (Auto Split), or chosen instruments (Split from Mix, Advanced Split).
Files. MP3 on all plans; WAV on Pro and Premier on suno.com; stems download without counting as extra downloads; MIDI from Studio on Premier. Studio's export help of 9 Sep 2026 says all audio exports are high-quality WAV files and gives no sample rate or bit depth; the Studio 2.0 page of 13 Aug 2026 says 32-bit WAV or MP3.
Prices (Suno's pricing page, checked 30 Sep 2026; it shows no date).
Plan | Price | Credits | Downloads a month | Adds |
Free | $0 | 50 a day | none | the free model (v6-mini); no commercial rights; no stems |
Pro | $8 a month billed yearly ("Saves $24 by billing yearly", so about $10 monthly) | 2,500 a month | 20 | v6 and v6-wild; commercial rights; Auto Split and Split from Mix |
Premier | $24 a month billed yearly ("Saves $72", so about $30 monthly) | 10,000 a month | 60 | Studio; Advanced Split; uploads up to 30 minutes |
A generation costs 10 credits for two songs, so Pro's monthly credits cover about 250 generations. Auto Split costs 50 credits; Split from Mix and Advanced Split 10 per extraction; MIDI from a stem 10. Max Mode costs more, by an amount Suno does not state. Suno does not state the cost of Extend, Replace Section, Cover, Remaster or Add Vocals in the pages read: read the price in the account before you run one. Suno credits are separate from Higgsfield credits.
Rights: what the terms say (revised 10 Aug 2026, effective 3 Sep 2026).
A paid plan and a permitted download. For a user "who has subscribed to the Pro or Premier paid tier", Suno "hereby assigns to you all of its right, title and interest in and to any Output owned by Suno". You "may commercially exploit Output solely to the extent it adheres to Suno's Conditions of Access and Use … provided you have obtained a permitted download of that Output". The rights attached to a download "are perpetual and are not affected by … the expiry, cancellation, downgrade or suspension of your subscription". Suno's help says songs downloaded while subscribed may be used "in film, tv, or video games" (edited 3 Sep 2026).
Nothing warrants copyright. "Suno makes no representation or warranty to you that any copyright will vest in any Output." Suno's help adds that "granting commercial use rights does not guarantee copyright protection", and that music made entirely with AI "would not qualify for copyright protection because a human did not write the lyrics or the music. Writing the prompt does not constitute the creation of the song" (edited 4 Sep 2026).
Free, Basic and trial. Outputs from the free plan are for "lawful, personal and non-commercial purposes". Trial downloads "are not eligible for commercial use" (help edited 3 Sep 2026). Extra downloads beyond the monthly allowance can be bought; whether a purchased download carries commercial rights is not stated, so ask first.
Marks, copies, remixes. Do not remove or obscure the fingerprint, watermark or metadata Suno appends; recording or stream-ripping an output is prohibited; editing to picture is allowed "to the extent such use is incidental". "Nothing in this paragraph permits commercial use of any Remix."
Other platforms. Use of an output elsewhere is subject to that platform's terms, and rights holders may collect revenue on distribution: permission from Suno is not permission from every place a film will play.
What you upload. Suno takes a licence to your submissions that covers "your likeness, voice rights and other indicia of your persona that may be embodied in your Content or Voice Model".
What may ship to a client. A cue downloaded on Pro or Premier, with its rights line, and with the client told in writing that a generated cue carries no copyright warranty and that an exclusive score needs a composer or a library. What may not: a cue with a free-plan ancestor, a trial download, a remix, a recorded stream, or a Voice or Persona of anyone who has not consented in writing. Never call a Suno cue "cleared". Write the rights line when you download (26.4) and re-read the terms before each client delivery (Chapter 34).
26.10 Version notes
What v6 changed (9 Sep 2026). Three variants replace the earlier models. v6 powers Custom Models, Extend, Cover, Replace Section and sound generation, takes multiple references (songs, playlists, audio, images, video) in one prompt, and adds the Variety slider and Max Mode. Several help pages predate it: Cover (27 Sep 2024), Song Editor and Audio Uploads (30 May 2025), Creative Sliders (3 Jun 2025), Add Vocals (21 Jul 2025), tempo (8 Oct 2025). Studio 2.0 (13 Aug 2026, Premier) is an editing surface, not a model.
Announced. Nothing further from Suno at the time of checking.
Re-check when the next model ships: the model selector and the plans it needs; whether Voices, Personas and Studio run on it; the Style, Exclude and Lyrics counters; the download allowances and file types; the terms (revised 10 Aug 2026); and the help pages listed above.
Open questions that change what you do
Whether Suno holds a named figure across a cue. No maker states it. Listen for it on every cue (26.7) and do not promise it to a client.
Whether Voices, Personas and Studio run on v6. The help pages do not say. Read the model selector before you plan on one.
What Cover, Add Vocals and Replace Section keep on v6, and whether stems are separated or regenerated. The pages are older than the model. Run each once on a cue you can afford to lose and compare it with its parent.
Whether Suno sings Egyptian or Saudi dialect correctly. Not established. Sung Arabic goes to a dialect reviewer word by word; a line that must be right goes to a singer.
The real field lengths, and Studio's export format. Only a third-party reading exists for the first; read the second from the first file you export.
What to remember
Brief a cue around a theme, not a mood: one figure and what it does in this cue. "Trailer music" and "epic orchestral" name the average.
Use Custom mode. Write the Style field as one paragraph in order: form and function, theme, roles, arc and ending, harmony, room, speech space, tempo last. Keep it under about 950 characters.
Never name an artist, a composer, a film or a brand anywhere Suno reads. Suno's own moderation help says names and trademarked terms may stop a song generating.
Refusals go in Exclude Styles, one to five sounds this cue could turn into, each with a positive twin in the text.
Instrumental means three layers: the switch, [Instrumental] first in Lyrics, and the vocal words in Exclude Styles.
Write the ending as behaviour and close the Lyrics with [End]; leave the speech range open by arrangement.
Run the brief twice as written, then candidates that change one thing each.
Repair the smallest region that fixes the fault and keep every parent. Stems are separations.
Work on a paid plan from the first generation, download the candidate you choose, fill the rights line, and tell the client in writing that copyright is not warranted.
Loudness is measured in the mix against the named delivery specification, never in Suno.
Comments