Audio Routing for OPI Interpreters: A Practical Setup Guide

Audio routing is the part of OPI setup you only notice when it breaks.

The call connects. You hear the provider. The patient hears you. Then your live transcript shows nothing, or worse, it captions only your own voice. The problem usually sits between the call app, your headset, the browser, and the tool that needs to hear the audio.

Good routing gives each part of the workflow the right signal:

• The agency portal gets your microphone
• Your headset gets the call audio
• Your transcription tool gets the source audio
• Nothing else listens without approval

For Windows device setup basics, Microsoft documents how to manage sound settings.

Audio routing is invisible until it breaks the call.

Map your setup first

Before changing settings, write down your path.

Ask:

• Where does the call happen: browser, softphone, meeting app, remote desktop, desk phone?
• Which device plays call audio: headset, speakers, phone handset?
• Which microphone sends your voice?
• Which tool needs to hear the call?
• Does the setup involve protected or confidential content?

This map prevents random toggling. Random toggling creates new problems, especially when you work under time pressure.

Browser portal setup

Many agency portals run in a browser. That setup often works well with live transcription because the browser can share tab audio or system audio.

Use Chrome or Edge if your tool recommends it. Open the portal, confirm the call audio plays through your headset, then start your transcription session. When the browser asks what to share, choose the tab, window, or screen that includes the call audio. If you see an audio checkbox, enable it.

Do a test before your shift. Play audio in the portal or a safe test tab and confirm the transcript responds.

If the transcript is blank, check whether you shared a window without audio. If it captions your voice, check whether the tool selected your microphone instead of system audio.

Softphone setup

Softphones add more variables because call audio may live in a desktop app, not a browser tab.

Start with the softphone settings. Confirm the input device is your microphone and the output device is your headset. Then check the transcription tool’s input options. Look for system audio, application audio, or screen share with audio.

If the tool cannot hear desktop-app audio, ask whether your agency supports a browser version of the softphone. Many portals have both. Do not install audio-routing utilities on a work machine unless IT approves them.

If you work from your own computer, keep the setup simple. Virtual audio cables and mixers can work, but they can also create echo, double audio, or privacy issues if you route the wrong source.

Meeting app setup

Video meeting platforms come with their own caption features, but OPI interpreters may need more than meeting captions.

If you use a separate OPI tool, make sure it can hear Zoom, Teams, Meet, or Webex. In many cases, that means sharing system audio. It may also mean choosing the meeting app as the source.

Watch for echo. If your microphone picks up your speakers, the tool may hear both the original speech and your interpretation. Use a headset. Keep speakers off. Confirm your microphone input and call output are separate.

For platform limits, see Zoom captions for interpreters and Chrome Live Caption for interpreters.

Desk phone or dual-handset setup

Physical phone audio creates the hardest routing problem because the call may never reach your computer.

A few compliant options may work:

• Use an approved headset base that connects to the desk phone and computer
• Use an agency-approved softphone instead of the handset
• Use a phone system that can output audio to the computer
• Use a private speaker setup only if policy allows it and audio quality stays clear

Avoid putting sensitive calls on unapproved devices. Do not use smart speakers. Do not use personal recorders. Do not route audio through a consumer app because it is convenient.

For healthcare calls, start with HIPAA for interpreters. For privacy-first note workflows, read No-recording interpreter notes and privacy.

Avoid double capture

Double capture happens when the transcription tool hears the same speech twice. You may see repeated words, delayed duplicates, or captions that mix source speech with your interpretation.

Common causes:

• Laptop speakers feed back into the microphone
• The tool captures both system audio and microphone audio
• A headset base sends mixed audio to multiple apps
• A meeting app plays your own voice back through monitoring

Fix the physical setup first. Use headphones. Turn off speaker output. Then check software inputs. If you need the tool to transcribe the source speaker, do not feed it your microphone unless the workflow requires it.

Build a pre-shift audio check

Use the same check every shift:

1. Confirm headset input and output.
2. Confirm the agency portal or softphone uses the right devices.
3. Start the transcription tool.
4. Play or receive safe test audio from the same source as calls.
5. Confirm the transcript hears the call audio.
6. Confirm it does not capture unrelated room audio.

This habit saves cognitive load. You do not want to troubleshoot devices while interpreting discharge instructions.

Keep routing boring

The best OPI audio setup is boring. It works the same way every shift. It uses approved devices. It keeps confidential audio away from unapproved apps. It does not require ten clicks before each call.

Map the path. Test the source. Use a headset. Avoid double capture. Keep policy in the loop.

Then get back to interpreting.

---

Related reading:

• Share call audio for live interpreter captions
• Interpreter home office setup
• Low-latency transcription for interpreters
• Interpreter privacy policy