
FFprobe's read_intervals seeks approximately, so the actual start can differ from the time you requested. When you request a relative interval duration, its end is calculated from the seek position actually found. Inspect returned timestamps before labeling the sample as an exact excerpt.
Key Takeaways
- A requested seek time and an observed first timestamp can differ.
- Relative duration follows the actual found start.
- A sampled probe is not a complete inspection of the whole movie.
Record the requested and observed interval
The FFprobe interval documentation explicitly warns that seeking is not accurate. It defines an interval with a start and optional end separated by a percent sign, and a plus sign for a relative offset.
For example, the documented pattern 10%+20 requests a seek near ten seconds, then reads for twenty seconds from the point found by seeking. It does not establish that the first returned packet or frame is exactly at ten seconds.
- Identify the stream and fields needed for your diagnostic.
- Record the interval expression before running the probe.
- Inspect the timestamps of the returned records.
- Note the first observed position and the scope actually examined.
- Use another appropriate selection workflow when the task requires an exact visual boundary.
Do not change the reported timestamps to match the requested label. That would hide the difference you need to understand.
A proposed boundary investigation
Suppose a hypothetical generated shot has a suspected glitch around thirty seconds. A probe requested near that point might return records from an earlier seek point. The earlier records can be useful context, but the report should describe the observed interval rather than claiming an exact thirty-second start.
| Record | What it establishes |
|---|---|
| Requested start | Where you asked FFprobe to seek |
| First returned timestamp | Where the observed sample begins |
| Relative duration | The requested length from the found start |
| Actual examined records | What the diagnostic can support |
This worksheet is a proposed reporting method, not output from a tested file.
Distinguish sampling from editing
Use the presentation-time selection guide when you need to express which frame timestamps pass a condition. Changing a sample request's label is not equivalent to editing a frame-accurate excerpt.
FFprobe also supports packet-count endpoints using a # prefix. Packets are not interchangeable with a duration or a decoded frame count; choose the unit for the investigation.
Accept a diagnostic only when it states the stream, requested expression, observed coverage and remaining uncertainty. If the suspected event lies outside the returned interval, extend or revise the investigation. A clean sampled interval does not prove the entire film contains no defect.
Related tasks
- How to Select FFmpeg Frames by Presentation Time
- Why Does a Lossless Video Trim Start Before My Chosen Frame?
- Why Video Reverse Prompts Can Miss a Fast Action
Sources and scope
- FFmpeg official documentation source: FFprobe interval selection: Approximate interval seeking; relative duration measured from the found start; packet-count endpoint syntax.
Documentation-based guidance. All scenarios and acceptance worksheets are proposed or hypothetical; no hands-on test, search-volume measurement or AI-citation ranking is claimed.
Prepared by Imagild Editorial, the company editorial team responsible for these guides. AI assisted research and drafting. This is documentation-based guidance with proposed examples, not a claim of hands-on testing or an independent product review. Report a correction.
Cover: AI-generated conceptual editorial illustration; not a real application screenshot, observed output or before-and-after test.