The short version
- A timestamp marks a spot measured from the beginning of a file. A timecode does the same thing in a strict format that often includes frames.
- SRT uses a comma before milliseconds. VTT uses a period and lets you leave out the hour.
- Transcripts can log time by segment, speaker turn, or word. Word-level timings let you edit captions with tight control.
- Automated timings are only estimates. Check them while playing the video and fix the few spots that run early or late.
- If every subtitle is off by the same amount, shift the whole file at once. If the gap gets wider as the video plays, you are usually looking at a frame rate mismatch or a recording that was edited.
What is a timestamp in a transcript?
A timestamp shows when a piece of speech starts, counted from the beginning of the audio or video file. In a transcript, you will usually see it next to a paragraph or speaker change, written in minutes and seconds, or in hours, minutes, and seconds on longer files.
You use timestamps for three main things: jumping straight to a spot in the recording to check a quote, citing an exact moment like 41 minutes in, and turning text into subtitles. A subtitle is just text attached to a start and end time.
How is a timecode different from a timestamp?
People use the two words in overlapping ways in everyday talk. In video production, though, a timecode is a standard label given to every frame using hours, minutes, seconds, and a frame number, like 00:12:30:15. Editors rely on it to line up video, audio, and captions. A transcript timestamp is usually less strict since it only needs to point to a moment in the audio.
This difference comes up when you move files between programs. Video editors count frames, so they need to know the project's frame rate. Transcripts and subtitle files count seconds and milliseconds instead, which means they do not depend on a frame rate. When you import one into the other, the editing software handles the conversion.
How do SRT and VTT format time?
An SRT file is a list of numbered blocks. Each block has a number, a line with the start and end times joined by an arrow, one or two lines of text, and then a blank line. Times are written in hours, minutes, seconds, and milliseconds, with a comma before the milliseconds, like 00:01:02,500 followed by an arrow and 00:01:05,100.
A VTT file starts with the word WEBVTT at the top. The time format follows the same order, except it uses a period instead of a comma and lets you drop the hours when they are zero. VTT comes from a W3C specification and also supports caption placement and styling.1
When a subtitle file will not load, check the decimal separator first. Putting a comma in a VTT file or a period in an SRT file is one of the most common mistakes. KolWrite exports both directly from your edited transcript with the right punctuation already in place.
What is the difference between segment times and word times?
Segment times record when a full sentence or subtitle block starts and ends. That is enough for regular subtitles and for clicking through a transcript. Word times log the start and end of every single word, so you can highlight words as they are spoken, cut video right on a word, and place captions with much finer control.
What you pick depends on the job. For meeting notes, segment times and speaker turns are plenty. If you are making animated captions or cutting video quickly, word times help a lot. Developers can pull word timings, speaker labels, and SRT files from the KolWrite API at console.kolwrite.com.
If you are quoting a recording in a report, put the timestamp next to the quote in one consistent style, such as 00:41:12, and leave the original media file untouched. Trimming or re-encoding a copy can nudge the timings slightly, and a citation only works if someone else can land on that exact same spot.
How precise are the timings?
In an automated transcript, timings are estimates. They can miss by a fraction of a second, especially right after a pause, when people talk at the same time, or when the audio is noisy. You will not notice this while reading or searching the text, but in subtitles even a small delay stands out.
To check them, play the video with the subtitles turned on. Pay extra attention to the very first line, anything right after a long silence, and parts where speakers talk over each other. Just fix the few lines that miss the mark instead of redoing the whole file.
Why are my subtitles out of sync, and how do I fix them?
First, check the pattern of the error. If every subtitle comes in early or late by the same amount, you are dealing with a constant offset, for example when the video has an intro that was not in the audio. Shift all the timestamps by that exact amount. If you add a five-second intro, push every timestamp back five seconds.
If the delay gets worse as the video plays, either the frame rate does not match or someone edited the video after the transcript was made. Always transcribe the final cut of the video, because cutting even a few seconds early in the timeline throws off everything after it.
Once you fix the file, check it in the actual player your viewers will use. Subtitles that line up cleanly inside an editing timeline can still play differently in a web player.
Common questions
Do I need timestamps in a transcript?
For simple reading, they are optional. You do need them if you want to check quotes against the audio, cite exact moments in a long recording, or create subtitles.
Why does SRT use a comma in the timestamp?
The SRT format specifies a comma between seconds and milliseconds. VTT uses a period there instead, so you cannot swap one format for the other without changing that separator.
Can I get a timestamp for every word?
Yes. You can get word-level timings in JSON from the KolWrite API, and the KolWrite panel supports word-level captions directly in Premiere Pro and DaVinci Resolve.
How do I shift all subtitles by the same amount?
Use the time-shift tool in a subtitle editor, or change the times by hand in a text editor if the file is short. Afterwards, test the file in the video player where it will actually be watched.
Why do the subtitles drift by the end of the video?
Most of the time, either the frame rate is mismatched or the video was edited after the transcript was generated. Transcribe the final cut of the video so the timings stay lined up.