Audio Transcription (Studio)

Audio transcription takes audio as input and uses Automated Speech Recognition and Automated Speaker Identification to generate text output.

Transcription glossaries improve AI transcription accuracy for specialized or difficult terms. Create glossaries on the Settings page and select one when creating a project. To reduce false matches, limit glossaries to specialized or difficult terms that are unlikely to be confused with common words or numbers. Avoid short or generic terms, and review the transcription output for misapplied terms before finalizing.

Transcription glossaries are shared with all users of the same organization in read-only mode.

Phrase Studio consumes Video Localization Hours.

Use Cases

A 45-minute customer interview recorded as an MP4 file.

Phrase Studio generates a text transcript with speaker identification which can be used to create a case study and pull quotes for a website.

Create an Audio Transcription Project

Supported Files and Limits

Formats: MP3, WAV, M4A, MP4, and MOV
Number of files: up to 50 files in one project.
File size and length: up to 20 GB or 180 minutes per file.
Subtitle files: SRT and VTT.

To create an audio transcription project, follow these steps:

From Phrase Studio, click New Project.

The Create Project page opens.
Add the files in one of these ways:
- Drag a file onto the upload field or click Upload Files to locate a file on your system.
- Click Import and select YouTube, Google Drive, or Brightcove as the source, if enabled.
The uploaded file is displayed.
Optionally, set the number of speakers for each file. Open the Speakers drop-down list next to the file and select a value from 1 to 5. If the file includes more than five speakers, keep the default Auto-detect option.
In Project Name, enter a name for the project.
Set the project visibility:
- Public Project is selected by default. Public projects are visible to all users in the organization who have access to Studio.
- Deselect Public project to create a private project that is visible only to the project owner. A private project can still be shared with selected users if needed.
In Source Language, select the spoken language of the files.

Source Language is required. Select the language before attaching subtitle files. Changing the source language afterward removes attached subtitles.
Optionally, in Other Spoken Languages, select additional languages. Add languages only if the audio switches between languages.
Optionally, select a Subtitle profile to determine subtitle display rules.
Optionally, select a Transcription Glossary to improve the recognition of specialized terms.
Optional: To use existing subtitles instead of automatic transcription, click + Add Subtitles next to the file and select an .SRT or .VTT file.

Phrase Studio skips automatic transcription and speaker identification, and aligns the subtitles with the media. Create and assign speakers manually, because .SRT and .VTT files do not include speaker information.
Click Create.

The file is uploaded and is displayed on the My Projects page.

The dialog generates transcripts and subtitles only. Add translations and dubbing after reviewing the transcript, from the editor.

Edit and Download the Transcription

Click on the project name to open it in the editor and view it in the Transcribe and Subtitle tabs. Both texts can be edited if required.

To download the transcription, click Download, select the content and language, and click Download again. Options include the transcription, translations, subtitles, source audio, and source video. Bilingual output, timecodes, and splitting paragraphs on speaker pauses are available for document formats.

Studio Insights

Studio Insights analyze the transcript of a recording with AI models and generate a summary, key topics, safe communication checks, and custom insights.

Insights and safe communication checks created on the Settings page are shared with all users of the same organization in read-only mode.

Use cases

Summarize customer support calls or identify potentially unsafe or low-quality communication. Phrase Studio returns a summary and flags sections for review.

Generate Insights for a Recording

Studio Insights are available after the transcription is complete.

To generate insights for a recording, follow these steps:

Open the project in the editor.
Click Studio Insights in the top bar.

The Studio Insights panel opens.
Select the type of analysis:
- Summary: Generates a summary of the spoken content. Click Generate Summary.
- Key Topics: Generates a summary of key topics. Click Generate Key Topics.
- Safe Comms: Checks for potential red flags. Select one or more check types, and click Check.
- Insights: Generates custom insights with user-defined prompts. Select one or more analysis types, and click Analyze.
A confirmation message is displayed, and the result appears under Results when the analysis is complete.
Optionally, click View Full Report to open the report. Select the language in the drop-down list, and click the download icon to save the report as a Word document.

Speaker Identification

Speaker identification detects and labels different speakers in an audio file for clearer transcripts and subtitles.

Automatic speaker identification is not available for projects with imported subtitle files.

Use Cases

A podcast with multiple participants is processed and each speaker is automatically tagged (e.g., "Speaker 1", "Speaker 2").

Manage Speakers

Click Manage Speakers under the Subtitle step to open the speaker list. In the list:

Click Add New Speaker to add a speaker.
Click the Edit Speaker icon to rename a speaker. Speaker names must be unique.
Delete speakers listed under Unused Speakers (speakers not assigned to any segment). The last remaining speaker cannot be deleted.

Assign Speakers to Segments

When multiple speakers are detected, drag a segment in the Speakers view to another speaker's row to change the assigned speaker. Drag a segment within its row to change its timing. The segment keeps its length. Dragging is not available in read-only projects.

To assign a speaker to several segments at once, follow these steps:

Select the segments using the Select segment checkbox.
Click Assign speaker (N), where N is the number of selected segments, and select the speaker.

The selected segments are assigned to the speaker.
Optional: Click Clear Selection to cancel the selection.

Note

A segment cannot overlap another segment of the same speaker. If dragging or assigning a segment would cause an overlap, the message Cannot assign speaker: would overlap with another segment of the same speaker is displayed, and the speaker is not changed.

Overlapping Speech

Use the Combined/Speakers toggle at the bottom of the editor to switch between a single waveform and individual waveforms for each speaker.

To represent two speakers talking at the same time, add a new segment for the second speaker in a separate row, with a start and end time that overlaps the first segment. Segments of different speakers can overlap in time.

Speakers in Dubbing

Dubbing generates separate audio for each speaker's segments. Segments assigned to different speakers that share a time range are dubbed as simultaneous audio.

A segment generates dubbed audio in the voice of its assigned speaker only, even if the segment contains dialogue from another speaker. Dialogue that is not captured in a segment does not appear in the dub. To include the dialogue of a second speaker, add a separate segment for that speaker.