Wine tasting video music: a 2-minute refined, quiet bed

For a wine tasting video, ask for a refined 2-minute chamber-jazz bed with soft piano and brushed drums, ducked under the host's voice, ending quietly so the closing line is heard. On Sume that is one generation at $0.125 and a two-minute render at $0.20, so $0.325.

A tasting video is led by a speaker describing taste and place, so the music is a room tone, not a feature. The brief aims for subtle sophistication and avoids anything that sounds like a commercial: no big builds, no bright pop elements.

The brief to send

The sections give one slow change at 1:00 and a long quiet fade, which gives the host room for the closing line.

curl -X POST https://api.sume.com/v1/music-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: wine-tasting-video-music-a-2-minute-refi-001" \
  -d '{
      "model": "sume/music-auto",
      "prompt": "Refined wine tasting ambience, 70 BPM, D minor to F major. Soft piano, brushed drums, warm double bass, quiet muted trumpet. Relaxed and elegant, one gentle chord change at 1:00, long quiet fade from 1:40. Intimate, warm club recording. Instrumental, no vocals. A 2-minute track."
    }'

Why each axis is set that way

Getting the length right

Put the two minutes in the prompt and verify the real length. In Timeline, use the host's voice as the spine, set duck_db to dip the music while the host speaks, and use fade_out_seconds up to 10 for the ending. A silent spine would reject duck_db.

Keep alcohol-related claims and age notices in your own script and on-screen text; the music has no role in that.

Before paying for a render, call POST /v1/timeline-1.0/plan with the same body. It is unbilled and reports the billable minutes, which for this 120-second video is 2. Then send the real POST /v1/timeline-1.0/render with an Idempotency-Key, wait up to 30 seconds in sync mode, and poll GET /v1/jobs/:id/status if it is still queued or processing. Statuses are queued, processing, completed, failed and canceled, and a 429 queue_full means try again shortly.

What it costs

A two-minute render is $0.20. For a ten-wine series, ten distinct beds are $1.25 in music and $2.00 in render, $3.25 in total, whereas one bed reused across all ten is $0.125 plus $2.00, $2.125. Consistency across a series is usually the better brand choice. Scaling is simple arithmetic. One finished video here is $0.325 of audio and assembly, so ten of them are $3.25 and fifty are $16.25, before any retries. Failed jobs are not billed as accepted generations, but a regeneration you ask for is a new job, so budget one or two extra takes for a hero clip.

Check before you publish

If the tasting is long, split into chapters of two minutes each and render separately.

Keep the prompt in your channel guide.

The hand-off between the two calls is the step people miss. When the music job completes, GET /v1/jobs/:id/result returns result.artifacts[] with an audio/mpeg file on media.sume.com. Timeline only accepts workspace media URLs, so import any outside file through POST /v1/media-imports first, and send the resulting URL as soundtrack.url. The Music docs also note that job.request.routed_model names the engine that actually ran, which is worth saving with the file.

• Check local rules for alcohol advertising before you publish.
• Keep the bed under the voice; descriptions of flavour need to be heard.
• Avoid pop-sounding elements that cheapen the mood.
• Confirm usage terms for generated music with Sume before commercial use.