Skip to main content

AI Speech Recognition: Speaker Labels, Audio Events, and Punctuation Control

· 3 min read
QuantumSubtitle Team
QuantumSubtitle Team
Maintainer

QuantumSubtitle Speech Recognition turns speech in audio and video into timestamped subtitle text. New recognition options give you more control over who said each line, which non-speech sounds appear in the transcript, and whether punctuation is included.

What the New Speech Recognition Controls Do​

When you create a recognition task, you can choose these options:

  • Speaker recognition: Separate different voices and label their subtitle cues. You can rename speaker labels in the subtitle editor.
  • Audio events: Include detected non-speech sounds, such as laughter, applause, and background sounds. Event cues can be filtered or edited in the subtitle viewer.
  • Punctuation control: Choose whether punctuation marks appear in subtitle text. The setting applies when you copy or export the text.

Speaker recognition distinguishes voices; it does not identify a speaker's real-world name. Add names yourself by renaming the labels after reviewing the transcript.

How to Use the Options​

  1. Open Speech Recognition and upload an audio or video file.
  2. Choose the audio language, or leave it on Automatic if you are unsure.
  3. Turn on speaker recognition or audio events as needed, and choose whether to include punctuation.
  4. Start recognition, then review the timestamped transcript in the subtitle editor.
  5. Correct the text, rename speakers, and filter or edit event cues before downloading an SRT file or copying the text.

These controls are useful for interviews, podcasts, lectures, webinars, and other recordings where readers need to follow multiple voices or understand meaningful sounds beyond spoken words. Clear audio and a review pass still help catch names, overlapping speech, and recognition errors.

Frequently Asked Questions​

What is speaker recognition, or diarization?​

Diarization separates voices in a recording and assigns labels to the corresponding subtitle cues. You can rename those labels in the editor.

Which audio events can it detect?​

The feature can include non-speech events such as laughter, applause, and background sounds. You can filter or edit event cues in the subtitle viewer.

Can I remove punctuation from subtitles?​

Yes. Turn punctuation display on or off. Your selection is used when subtitle text is copied or exported.

Can I export the transcript as SRT?​

Yes. Review the transcript and download it as an SRT subtitle file.

Try AI Speech Recognition, read the speech recognition guide, or visit the Speech Recognition feature overview.