What Are Transcription Timestamps?
Transcription timestamps are timecode markers linked to specific text segments in an audio or video transcript. They indicate when a spoken line begins and ends relative to the start of the recording.
If you are learning how to transcribe audio, choosing the correct timestamp granularity is one of the most important steps. Timestamps turn raw text into an interactive index that lets readers match written quotes back to exact audio moments.
Sentence-Level Timestamps
Sentence-level timestamps group text into complete grammatical sentences, assigning a single start time and end time to each sentence.
When to Use Sentence Timestamps
- Video subtitles and video editing: perfect for SRT and VTT exports where text lines must stay readable on screen.
- Podcast episode transcripts and show notes for content creators creating subtitles and captions.
- Reading documents where grammatical flow is more important than millisecond-level word timing.
Paragraph-Level Timestamps
Paragraph-level timestamps aggregate speech segments between longer natural speech pauses, creating broader blocks of text with start and end timestamps for whole paragraphs.
When to Use Paragraph Timestamps
- Meeting notes, executive summaries, and long lectures.
- Legal depositions and oral history documentation.
- Continuous reading experiences with minimal visual clutter.
Word-Level Timestamps
Word-level timestamps track the exact start time and end time of every single spoken word down to the millisecond.
When to Use Word-Level Timestamps
- Academic research, phonetic analysis, and audio indexing.
- Precision audio cutting and automated video highlight clipping.
- JSON data exports for developer integrations and custom media players.
Which Timestamp Mode Should You Use?
Compare Konthora’s three timestamp modes across granularity, readability, editing convenience, and export suitability.
| Timestamp Mode | Granularity | Readability | Editing Convenience | Typical Use Case | Export Suitability |
|---|---|---|---|---|---|
| Sentence-Level | Medium (per sentence) | High | High | Subtitles, show notes | SRT, VTT, TXT |
| Paragraph-Level | Low (per speech block) | Very High | Medium | Lectures, meeting notes | TXT, VTT |
| Word-Level | High (per word) | Low (Dense) | Very High (Precision) | Research, audio editing | JSON |
How Timestamps Appear in Each Export Format
Understand how timestamps are formatted across different export files generated by Konthora. You can also turn text into speech using our free text-to-speech tool.
SRT Format
SubRip files use sequential line numbers followed by start and end timestamps inHH:MM:SS,mmm format.
1 00:00:01,200 --> 00:00:04,500 Welcome to Konthora's audio transcription workspace. 2 00:00:04,800 --> 00:00:08,100 Choose sentence or word timestamps for your file.
VTT Format
WebVTT headers begin with WEBVTT and use period delimiters for milliseconds:
WEBVTT 00:00:01.200 --> 00:00:04.500 Welcome to Konthora's audio transcription workspace. 00:00:04.800 --> 00:00:08.100 Choose sentence or word timestamps for your file.
TXT Format
Plain text exports strip all timecode headers, providing clean prose for reading documents.
Welcome to Konthora's audio transcription workspace. Choose sentence or word timestamps for your file.
JSON Format
Structured JSON provides segment objects containing raw floating-point second offsets and individual word objects.
[
{
"id": 0,
"start": 1.20,
"end": 4.50,
"text": "Welcome to Konthora's audio transcription workspace."
}
]