Overview

Small recording improvements lead to noticeably better transcripts.

Environment

• Choose a quiet room and reduce echo with soft surfaces.
• Turn off fans, AC, and other continuous noise sources.

Microphone setup

• Place the mic 6 to 12 inches from the speaker.
• Use a pop filter to reduce plosives.
• Monitor levels to avoid clipping.

Speaker tips

• Speak clearly and avoid talking over others.
• Introduce speakers at the start of the recording.

After upload

• Rename speakers and review key sections with timestamps.
• Use Smart Summaries to confirm the main points.

Audio quality targets

Accuracy starts before upload. The model performs best when speech is clear, room noise is controlled, and speakers stay at consistent volume. Aim for clean voice capture instead of trying to fix heavily noisy files later.

• Record in a quiet room with soft surfaces to reduce echo.
• Keep microphones 6 to 10 inches from the speaker.
• Avoid clipping by checking levels before the session starts.
• Use separate mics for multi-speaker interviews when possible.

Before-you-upload cleanup

1. Trim long silence at the beginning and end.
2. Normalize volume if one section is much quieter than others.
3. Export to a standard codec (AAC/MP3/WAV) to avoid decode issues.
4. Listen to a 30-second sample to verify speech is intelligible.