30 jingle candidates for 5 brands: $3.75, then trim 5 to 15 seconds

Six Music 1.0 candidates for each of five brands is 30 x $0.125 = $3.75. Trimming five winners to 15 seconds in Timeline adds $0.50. The arithmetic.

5 min readSume
All posts

Thirty jingle candidates (six per brand, five brands) cost $3.75 on Sume, because Music 1.0 charges a flat $0.125 per accepted generation. If you then trim the five winners to 15 seconds with Timeline 1.0 at $0.10 per ceil(output minute), you add $0.50 and the whole exploration costs $4.25.

Why a flat price changes how you audition

Per-minute music tools make a long exploratory run expensive. Music 1.0 does not: the price does not change with prompt length, image conditioning or the length you ask for in the prompt. So the cheapest way to find a good jingle is to ask for many short, different takes rather than to refine one prompt for a long time.

Each request goes through the Music Router (sume/music-auto, which resolves to Lyria 3.5 today). The prompt can be up to 5,000 characters. There is no duration parameter; write the length into the prompt, for example 'a 15-second jingle'.

Audition budget (as of 2026-10-08)
StepCountUnit priceCost
Candidates30$0.125$3.75
Trims to 15 s5$0.10$0.50
Total$4.25

Make the six candidates different

Music 1.0 has no seed, so repeating a prompt does not give you a controlled variation. Change the brief instead. The docs suggest a seven-axis brief: emotion, genre, tempo as a number, key and mode, two to four textured instruments, an arc with one named moment, and an era or production note. For six candidates per brand, vary tempo by at least 12 BPM between takes and change the lead instrument.

End each prompt with 'Instrumental, no vocals.' unless you want a sung tag line. Put exclusions in the positive prompt; a non-empty negative_prompt is not supported.

Budget for a bigger pitch

Scale it: ten brands with six candidates each is 60 x $0.125 = $7.50, and ten trims add $1.00. Even a 100-candidate sprint is $12.50 of music, before any trims. A flat price makes a wide search cheap and a long search no more expensive than a short one.

One caution: a fixed price per generation is not a license answer. Read the terms before you use any generated track in a paid campaign.

Trimming without re-generating

Generated length is steered, not guaranteed. A request for 15 seconds can come back a little long. Timeline 1.0 takes a Sume-hosted audio spine with an exact duration_seconds, so you can cut the winner to the second. Add soundtrack fade_out_seconds (up to 10) so the end of the jingle does not stop abruptly.

A trim-only render is rated by the output minute, so a 15-second jingle is one billed minute: $0.10. The plan call returns the billable minutes before you commit.

A short prompt template

Per brand, write one base brief with the seven axes and change only tempo and lead instrument across the six takes. For example: 'Playful, optimistic, 108 BPM, G major. Marimba and finger snaps, a bright synth pad, a short vocal-free hook at 0:08. A 15-second jingle. Instrumental, no vocals.' Then change 108 BPM to 96, 120, 132 and swap the marimba for ukulele or pizzicato strings.

Listen to all six, shortlist one per brand, and keep the rest as alternates. The alternates cost you $0.125 each and are already paid for.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume