Subtitle Timestamps and Frame Rate: When Does FPS Actually Matter?
An SRT file uses elapsed times, not frame numbers. A video being labeled 24 or 30 fps does not by itself mean its subtitles need scaling. Frame rate matters when the media's playback speed, timecode interpretation or frame-to-time mapping changes. Identify the actual change before transforming the subtitle timeline.
In this guide
Separate elapsed time from frame-based timecode
SRT: 00:00:12,500 means 12.500 elapsed seconds.
A frame-based display such as HH:MM:SS:FF needs its timebase to interpret FF.Do not read the final three digits in SRT as a frame count. They represent milliseconds. Likewise, do not copy a frame-number field into a milliseconds field without conversion. Label screenshots and notes with the time format they actually show.
Aegisub's variable-frame-rate documentation discusses the mapping between video frames and time. This distinction is important when a timeline display and a subtitle timestamp appear to disagree even though they refer to the same moment.
A frame-rate conversion may or may not change duration
| Change | Likely subtitle implication |
|---|---|
| Frames were resampled while duration stayed the same | Elapsed-time subtitles may already be correct; verify playback. |
| The same frames play faster or slower | Subtitle times may require a speed-related transform. |
| A new scene was inserted or removed | The affected section needs an edit-aware correction, not simply a global ratio. |
| Only a player display setting changed | Do not alter subtitle times until an actual synchronization error is observed. |
Avoid applying a familiar ratio merely because two exports show different fps labels. Measure recognizable speech near the start and end. If they already align, an unnecessary transform creates the error you were trying to prevent.
Calculate only a known speed change
For an illustrative constant-speed change, suppose 100.000 seconds of source content becomes 96.000 seconds in the final video with no cuts. A corresponding source time would be multiplied by 0.96, plus any separately measured offset. At 50.000 source seconds, that predicts 48.000 final seconds.
That model assumes a single constant speed across the section. Speed ramps, inserted pauses and different edits violate it. Aegisub's export filters include a frame-rate transform; its existence is not evidence that every mismatch should use it.
Use measured reference points in AENural
When a known speed change produces gradual drift, use the Subtitle Sync Fixer to map real subtitle times to real video times. This avoids choosing a ratio merely from an export's fps label. Keep a subtitle master before applying a correction.
- Load your subtitles and choose an Early reference. Enter the line's existing Subtitle time and its actual video position under Should appear at.
- Choose a Late reference for a recognizable line well away from the first. Enter its existing and correct times using the same elapsed-time convention.
- Select Fix sync, inspect the correction summary and download the corrected subtitles. Replay an independent middle line as well as your two references.
In the illustrative speed-change example, a line at source 50 seconds should occur at final 48 seconds. If that middle point does not fit after matching your endpoints, investigate edits or a changing speed instead of repeatedly rescaling the file.
For one cue, the Subtitle Editor exposes explicit Start and End fields. Enter millisecond timestamps there, not a frame count copied from a different timeline display. AENural's reference fields do not automatically infer missing timebase information from a screenshot.
Choose the repair from the symptom
- Confirm the exact final video version and duration.
- Identify whether the error is constant, steadily changing or section-specific.
- Keep the original subtitle file before changing times.
- Verify the corrected result in the actual receiving player.
For measured delay or drift, follow the existing subtitle synchronization guide. This article explains why a frame-rate label alone is insufficient; the repair should follow observed timing evidence, not a guessed fps conversion.
Frame rate, offsets and drift
Captions start in sync but get worse later. Is a fixed delay enough?
A fixed delay moves every cue equally; it cannot correct a changing mismatch. Compare several reference lines and check the matching edit and any speed change before choosing a measured timeline adjustment.
Does converting a video from 24 to 30 fps mean I must rescale the SRT?
Not if the conversion keeps the same elapsed duration by changing frames rather than playback speed. Measure recognizable content times. The frame-rate label alone is insufficient evidence that subtitle times should be multiplied.
Can an embedded subtitle track still belong to the wrong cut?
Yes. Being inside the video container does not verify its text or timeline. Check the actual content. Treat statements that embedded subtitles cannot drift as an unsafe shortcut, not a reliable diagnostic rule.
Why is one section wrong while the opening and ending look correct?
A cut, inserted pause or local speed change can create a nonuniform mismatch. Check the middle and each edit boundary. A single whole-file scale cannot represent every kind of timeline change.
Should I use the FPS number listed beside a downloaded subtitle as proof?
Use it as a clue, then compare the actual release and content. Different introductions or edits can create mismatches even with matching FPS labels. A few measured reference lines are more useful than the label alone.
What is different about an SRT timestamp and a frame timecode?
SRT stores elapsed hours, minutes, seconds and milliseconds. Frame-based timecode counts frames within a timebase. Establish which representation an application is showing before treating its final number as milliseconds.
Why does the player speed control make one scene line up but another worse?
You may be changing the wrong timing model. Diagnose constant offset, steady drift and sudden jumps separately. Compare the actual recording rather than repeatedly increasing a speed parameter until one sentence happens to align.
Does variable frame rate automatically prove the subtitles are wrong?
No. It is relevant to investigate, especially when applications interpret the source differently, but it is not proof. Compare measured speech times and the intended playback before converting media or rewriting caption timings.
When can I use two reference points to correct a speed mismatch?
When the mismatch is consistent with a single linear change across that section. Use separated, unmistakable lines, then check additional middle points. Two matching endpoints alone do not validate every cue.
Why do subtitles disappear after seeking even though playback started correctly?
Check whether the subtitle file actually extends to that part of the video, then compare another player. A seek-related symptom can come from incomplete cues or player behavior; do not immediately rescale the whole file.
