Dim a full 3-minute Short in one filter job: 180 of 300 s, $0.02
Video filter accepts clips up to 300 seconds, so a whole 180-second Short dims in one $0.02 job. A 10-minute TikTok ad needs at least two.

A 3-minute Short is 180 seconds, and Sume's Video filter takes a source of up to 300 seconds, so the whole thing dims in a single job at $0.02 (as of 2026-10-08). A dim op of 0.45 multiplies the luma of the full clip and leaves black black. A TikTok non-Spark ad of 10 minutes is 600 seconds, twice the filter's limit, so it needs at least two filter jobs, or a dim applied to shorter pieces before they are assembled.
Limits on the same page
The docs give the filter source limit as 300 seconds. A longer file is refused with output_duration_exceeded. The filter keeps geometry, frame rate and audio unless the program changes them, and it returns a new MP4 while the source stays as it was.
| Clip | Length | Fits one filter job? | Filter cost |
|---|---|---|---|
| 3-minute Short (YouTube limit) | 180 s | Yes, 120 s to spare | $0.02 |
| Longest clip the filter accepts | 300 s | Yes, exactly | $0.02 |
| 5-minute and 1 second clip | 301 s | No | Split first |
| TikTok non-Spark ad at the 10-minute limit | 600 s | No; 2 x 300 s | 2 x $0.02 = $0.04 |
Why dim a Short at all
Burned-in captions lose contrast over bright footage. The dim op takes an amount in the range above 0 and up to 1. The docs example is 0.45, which makes the clip darker; 1 changes nothing, and 0 or anything above 1 is refused. You can chain up to 8 ops, and add a filters-only filtergraph of at most 2048 characters and 32 filters, applied after the ops.
The free check endpoint runs the schema, the op allowlist, the filtergraph allowlist and the host preflight, and it returns an estimate and next_action. Run it first. It does not create a job, so it costs nothing, but it also cannot show you how dark 0.45 looks. For that, you pay for one encode.
Order of operations for a captioned Short
Dim first, then caption. The caption job is a separate render that burns text onto an existing public video URL, at $0.20 per job for clips of up to 60 seconds under the current fixed estimate. So for a 59-second Short: dim at $0.02, captions at $0.20, total $0.22. For a 180-second Short, the dim is still one job, but the docs state the caption price only for up to 60 seconds, so check the live catalog before you assume a rate.
Remember the host rules differ. Video filter needs a media.sume.com clip in your workspace, while Video captions needs a public HTTPS URL that Sume can fetch. Plan the hand-off between the two, and read the YouTube page's three-minute rule once more if you plan to dim and publish a file longer than 180 seconds.
What one job does not do
A dim job does not trim, resize or add text. Cutting a range is Video trim, and the filter allowlist leaves out trim and setpts for that reason. It also leaves out drawtext and subtitles, so captions come from the separate caption job. The output keeps the source geometry, frame rate and audio, which means a 1080x1920 Short in is a 1080x1920 Short out. If you want to darken only part of the clip, split it first with trim, dim the piece, and rejoin in Timeline; each piece is a $0.02 job, plus the render at $0.10 per started minute.
Sources
Related posts
More in Media tools
- Exact or keyframe video trim: which to pick and what changes
Sume video-trim defaults to a frame-accurate exact re-encode; keyframe is a stream copy that can start a GOP early. How to choose and re-base the times.
- Extract 6 thumbnail candidates from a clip with video-frames
Send POST /v1/video-frames with an at list of up to 24 timestamps and get durable image URLs at source size. Python polling example, formats and limits.
- Extract the first 90 seconds of a video's audio: audio-detach range
Send audio-detach range {start 0, end 90} for $0.01: 17.28 MB as 48 kHz stereo wav or 1.44 MB as 128 kbps mp3, a Shorts Audio Library song's 90 s.
- Five caption looks on one clip: one transcription, four restyles
Test five caption styles on one video for $1.00: the first job transcribes, four source_caption_id restyles reuse the word timings. Request bodies included.
Written by Sume