In noisy environments, ambient sounds or other people’s voices may be transcribed in your report along with your own voice. To prevent this from happening, use background noise filtering to reduce or block sounds that fall below a chosen loudness threshold before they are transcribed.
In this article:
- Prerequisites
- How background noise filtering works
- When to use background noise filtering
- Enabling background noise filtering
-
Frequently asked questions
Prerequisites
Before you start, make sure:
- You use the Rad AI Reporting classic experience.
- Background noise filtering has been enabled for your organization.
How background noise filtering works
Background noise filtering uses a loudness threshold to reduce or block sounds below a chosen level before they’re transcribed. It filters based only on how loud a sound is, not who is speaking.
Users can change the loudness threshold to suit different environments. Since this feature only considers volume, a higher threshold setting sometimes drops the user’s own quiet words.
When to use background noise filtering
Use background noise filtering when the voices of other people or loud background noises are being transcribed along with your own voice—in a noisy emergency department, for example.
If you frequently use short phrases, background noise filtering may also be helpful to ensure the short phrases aren’t dropped.
Enabling background noise filtering
-
Go to Personal Settings > Speech
>
Voice ID, and if it’s enabled, disable it. Voice ID
is available
only by request. If you don’t see it in your Personal Settings, it is
disabled.
💡 Using each feature in isolation helps us evaluate the effectiveness of each.
-
Go to Personal Settings > Speech > Background filtering.
-
Choose a level based on how much background sound you want to reduce. Higher levels reduce more background sounds, but they’re also more likely to filter out quieter parts of your own speech, especially if you speak softly or away from the microphone.
💡 Rad AI Reporting recommends starting with Medium in standard radiology reading rooms with multiple speakers.
-
In your computer’s system settings, adjust your microphone gain to make the
microphone more or less sensitive to distant voices while still capturing
yours. The best level depends on how loudly you speak and how close you sit
to the microphone.
-
In Windows, open Control Panel > Sound.
-
Click Recording, right-click your microphone, and choose Properties.
-
Open the Levels tab and set the microphone level to 60.
-
Test the setting by dictating a report.
- If your own words aren’t being picked up reliably, raise the value.
- If you are still capturing other people’s speech, lower the value.
-
- Optimize your microphone position and room setup: keep the microphone close to your mouth, point it away from other radiologists when possible, and don’t leave it positioned toward nearby conversations.
Frequently asked questions
Should I use Voice ID, microphone settings, or background noise filtering?
These features solve different problems.
| If your problem is... | Try... | Why |
|---|---|---|
| Other people’s voices are appearing in your reports. | Voice ID | Voice ID recognizes who is speaking and helps ignore speech that doesn’t match your voice. |
| Background sounds and other people's voices are appearing in your reports. | Background noise filtering | Background noise filtering reduces or blocks sounds based on how loud they are. |
| Echo, hum, or inconsistent microphone audio is affecting transcription. | Microphone settings | Microphone settings improve the quality of the audio captured by your microphone before it’s transcribed. |
For the best results, use the combination of settings that matches your environment. For example, you might use Voice ID to ignore nearby speakers, background noise filtering to reduce quieter background sounds, and microphone settings to improve microphone audio quality.
What settings are recommended when using background noise filtering?
For the best results with background noise filtering:
- Set Voice ID to Off.
- Choose the background noise filtering level that best matches your environment.
This configuration helps background noise filtering work as effectively as possible while reducing the chance of unwanted background sounds being transcribed.