Suno Studio clip stretching: fit AI music to a video on Sume

Sume does not time-stretch music. Suno Studio's October 8, 2026 update lets you shift-click-drag a clip so it stretches to a track's tempo. Sume's Music Router has no stretch control, no downbeat detection and no editable clip lane. To fit a track to a video on Sume, you write the tempo and length into the prompt, cut the result with a Timeline audio job, and set fades and ducking in the Timeline render.

What did Suno Studio change on October 8, 2026?

Suno's release notes list three Studio changes for October 8: better downbeat detection for clip timing, shift-click-drag clip stretching for matching samples with tracks at any tempo, and fade curves you can shape by dragging nodes into exponential and logarithmic forms. The same page lists pop-out windows for synths and plugins.

How do you make a track the right length without stretching?

Ask for the length when you generate. The Music Router takes a prompt of 1 to 5,000 characters and rejects duration and duration_seconds, so length and tempo belong in the text: "a 30-second track", "84 BPM", or section markers such as [0:00-0:30] Intro. Each generation is one flat charge of $0.125, so a second take at a different tempo costs the same as the first.

If a track runs long, cut it rather than stretch it. A Timeline audio split job takes up to 20 ranges of a Sume-hosted file and returns a durable file for each. The job is sample-domain with no re-synthesis, costs one flat credit, and writes wav by default. The same documentation advises keeping wav if you will join the file again.

Where do fades and ducking happen?

In the Timeline 1.0 render. The soundtrack field takes a bed url, a gain_db, a loop flag, a fade_out_seconds up to 10, and a duck_db from 0 to 20 that lowers the bed under the audio spine. The output also accepts fade_in_seconds and fade_out_seconds of 0 to 5. These are lengths, not curves: you cannot bend a fade into an exponential shape.

• Set loop when the bed is shorter than the video, and a fade_out_seconds so the loop does not end on a hard cut.
• Use duck_db only when the render has a real voice spine; a silent spine is refused with duck_requires_audio_spine.
• Keep the music file as wav if it goes through a split first; mp3 adds priming padding at every edge.

When is Suno Studio still the better tool?

If you want to place a finished song on a multitrack timeline and nudge individual clips by hand, Studio's stretching and downbeat snapping do something Sume does not. Sume's route suits a different job: an API call that returns a track, a trim, and a rendered video, with every step a job you can poll at GET /v1/jobs/:id/status.

A practical split: use Studio when the music is the product and you are editing it. Use Sume when the music is a bed under a video and the fit is a matter of prompt, cut, fade and duck. If a track lands at the wrong tempo for the edit, regenerate it with the tempo in the prompt instead of trying to bend the audio.