20 overlapping 15-second stings from one Lyria track for $0.135
One Timeline audio split job takes up to 20 ranges, and ranges may overlap. One 2-minute bed plus one split gives 20 stings for $0.135, not $2.50.

One Timeline audio split job accepts up to 20 ranges and the ranges can overlap, so a single 2-minute Lyria track can become 20 fifteen-second stings starting every 5 seconds. The cost is $0.125 for the track plus $0.01 for the split, $0.135 in total. Generating 20 separate stings would cost 20 x $0.125 = $2.50.
The windows are 0 to 15 s, 5 to 20 s, and so on up to 95 to 110 s. The last window ends 10 seconds before the track does. Each range becomes its own file with its own audio_url, and the split is sample-domain, with no re-synthesis.
The request
Follow the rules on the Timeline audio page: operation: "split", a top-level url that is already this workspace's media.sume.com audio, ranges[] of 1 to 20, and an Idempotency-Key header. Do not send parts. The script builds the body and checks the two limits before you spend anything.
import json
ranges = [{"start": s, "end": s + 15} for s in range(0, 100, 5)]
assert len(ranges) == 20 # the split limit
assert ranges[-1]["end"] <= 120 # inside a 2-minute track
body = {
"operation": "split",
"url": "https://media.sume.com/artifacts/artf_demo/bed.wav",
"ranges": ranges,
}
print(len(ranges), "windows, last ends at", ranges[-1]["end"], "s")
print(json.dumps(body)[:90], "...")
Cost and size
The split is a worker job with no provider inference, so the rate is flat at $0.01 per job whatever the number of ranges. The table shows what changes when you cut the 2-minute bed different ways.
| Cut | Files | Jobs | Arithmetic | Total |
|---|---|---|---|---|
| Generate each sting | 20 | 20 music | 20 x $0.125 | $2.50 |
| One split, 20 overlapping windows | 20 | 1 music + 1 split | $0.125 + $0.01 | $0.135 |
| Two splits (if you need more than 20 ranges) | 40 | 1 music + 2 split | $0.125 + 2 x $0.01 | $0.145 |
| Non-overlapping 15 s windows | 8 | 1 music + 1 split | $0.125 + $0.01 | $0.135 |
Where 20 stings are used
A sting is a short musical punctuation: an ad end tag, a transition between scenes, a notification sound inside a Short. Twenty windows give you enough variety to assign a different one to each of 20 ad variants without paying for 20 generations. Pick the files by ear, because overlapping windows from the same track share a key and a tempo, which is what makes a set of stings sound like one brand.
If you need fewer, cut fewer. The price of the split job does not change with the number of ranges, so there is no saving in asking for 6 windows instead of 20, only less to review. If you need more than 20, run a second split on another range of the same file for another $0.01.
Caveats before you rely on it
Overlapping windows share audio, so ads that run together will sound alike. Cut at musical boundaries: at 120 BPM a bar is 2 seconds, so a 4-second step moves exactly two bars each time and keeps every window on a bar line, while a 5-second step moves 2.5 bars and every second window starts mid-bar. Twenty windows at a 4-second step reach 0 to 76 s for the starts and end at 91 s. The Music 1.0 page says tempo in a prompt is a direction, not a guaranteed value.
Keep the default wav output if you will join or re-cut the stings again. The docs say mp3 is smaller but adds priming padding at every edge. As a size check, if the output keeps CD-quality stereo (44.1 kHz, 16-bit, 176,400 bytes per second), each 15-second wav is about 2.6 MB and the 20 files about 53 MB; the docs do not state the output sample rate, so measure one file.
Sources
Related posts
More in Media tools
- Crop a 21:9 video to 9:16, 4:5, 1:1 and 16:9: fraction table
From a 21:9 source, a full-height crop is 0.2411 wide for 9:16, 0.3429 for 4:5, 0.4286 for 1:1, 0.7619 for 16:9. $0.02 per crop; the check is free.
- 24 character stills on gpt-image-2.5: 24c low, 48c medium, $1.68 high
If you omit quality, gpt-image-2.5 defaults to high. 24 stills (8 characters, 3 poses) cost 24c at low, 48c at medium and 144 to 168c at high on Sume.
- A 28-second bed under a 59-second Short: loop, duck, $0.225
Timeline soundtrack loop repeats a 28-second bed under a 59-second voice spine, with a fade-out up to 10 s and duck up to 20 dB. Render $0.10 plus Music $0.125.
- A 30-second music bed under a 70-second render costs $0.325
One Music Router track at $0.125 plus a 70-second Timeline render at $0.20 is $0.325. Use soundtrack loop, fade_out_seconds up to 10 and duck_db 0 to 20.
Written by Sume