Voice to voice translation
Set up spoken translation for an online meeting, an in-person conversation, or audio playing on your computer.
Voice to voice is AI translation spoken aloud: JotMe listens, translates, and speaks the result instead of only writing it down. Use it one direction at a time, or both at once.
Before you can turn it on#
In language settings, select Fast (Audio) and the translation output before opening the voice-to-voice menu. The menu unlocks when both are set. Fast must be selected before the session starts; its usage is covered once in Usage limits.
First, choose where the audio comes from#
Use the matching section below:
- Online Meeting for Zoom, Microsoft Teams, Google Meet, or another meeting app.
- Other for an in-person conversation or event, or media playing on this computer.
- For participants when guests need to hear translated audio in their browsers.
Online Meeting#
Use this environment when the audio is coming from a meeting app. Virtual audio lets JotMe hear the meeting and, when needed, send translated speech back into it.
| Mode | What happens |
|---|---|
| Listening only | You hear the other participant in your language |
| Speaking only | The other participant hears you in their language |
| Two-way | Both at once |
Complete the virtual-audio setup below before starting the meeting.
Online Meeting: set up virtual audio#
This section applies only when you selected Online Meeting. A virtual audio device acts like a cable between JotMe and the meeting app, allowing audio to move in one or both directions. Set it up before the meeting.
macOS. BlackHole#
| Mode | Needs |
|---|---|
| One-way | BlackHole 2ch |
| Two-way | BlackHole 2ch and BlackHole 16ch |
Install
JotMe can install BlackHole for you: Install BlackHole 2ch, and Install BlackHole 16ch as well if you want two-way.
Restart if it does not appear
"Restart your Mac if the device does not appear." JotMe confirms when it sees them: "BlackHole 2ch detected — ready for virtual routing."
Point your meeting app at it
"Set your meeting app speaker and microphone to the exact BlackHole devices shown below." JotMe names the exact device for the mode you chose.
Audio MIDI Setup is linked from the panel if you need to inspect devices yourself.
Windows. VB-CABLE and VAC#
| Mode | Needs |
|---|---|
| Uplink (speaking) | VB-CABLE: meeting microphone = CABLE Output (VB-Audio Virtual Cable) |
| Two-way | Virtual Audio Cable Line 1: meeting microphone and speaker = Line 1 |
"VB-CABLE handles uplink. VAC Line 1 handles two-way meeting audio capture." If Line 1 does not show up after installing, restart Windows.
Online Meeting: choose a routing recipe#
JotMe spells out the three combinations. Use the one that matches what you want:
| Goal | Settings |
|---|---|
| Both at once | Meeting capture → BlackHole 2ch. Outgoing → BlackHole 16ch. Incoming → your headphones. Turn on TTS my own voice. |
| Only uplink (they hear you translated) | Incoming stays same as outgoing, outgoing → BlackHole 2ch. Meeting mic = BlackHole 2ch, meeting speaker = headphones. |
| Only hear others translated | Meeting app speaker → BlackHole 2ch. JotMe captures others from BlackHole 2ch and sends the translation to your headphones. |
The signal path for the uplink case, in JotMe's own words: "Real mic → JotMe → translated TTS → BlackHole 2ch. Meeting mic = BlackHole 2ch."
Tip
Use headphones. With speakers, the translated voice comes out of your laptop, is picked up again by the microphone, and gets translated a second time.
Online Meeting: TTS my own voice#
Turn this on "when JotMe sends translated speech to the meeting app through BlackHole": it is what makes your translated voice go to the meeting rather than to your own ears.
Confirm the meeting app is using the virtual devices shown by JotMe. Turn on the voice-to-voice section, select Start live translation in the top bar, and speak.
Adjust translated voice volume#
Open the voice-to-voice menu and use Translated AI voice volume. Adjust it while translation is running; the setting is saved on this computer for your next session.
- Incoming — what you hear changes the translated voice in your headphones or speakers.
- Outgoing — what others hear changes your translated voice sent to the meeting app.
- Two-way shows both controls so you can adjust each direction separately.
- Other has a single Translated voice control for the selected audio source.
The default is 300% when you have not saved a volume preference. Your existing saved level is preserved. Lower it to 0% to mute translated speech, or choose any level up to 300%. Incoming playback carries a small built-in boost (about 30%) because meeting apps amplify the copy others hear. Boosted audio uses peak limiting to reduce distortion. These controls affect translated speech only, including playback through virtual audio devices; they do not change your microphone or original meeting audio.
Translated AI voice volume
If translated speech is still quiet, check the selected JotMe output device, that device's system volume, and the meeting app's speaker or microphone level.
Balance faster and smoother voice playback#
Open Settings → Live translation → Voice translation quality. Choose a shorter wait for quicker responses or Smooth for more buffered speech. The default wait is 0.5 seconds; Smooth selects 2 seconds. You can adjust the maximum wait from 0 to 5 seconds. Playback can begin earlier when enough audio is ready.
This preference applies to incoming and outgoing voice in two-way online meetings and to shared web listeners. It changes playback buffering, separately from the caption translation-quality setting and language detection.
Supported output languages#
Fast voice and captions offer 13 output languages: Chinese (Mandarin, Simplified), English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese, Russian, Spanish, and Vietnamese. Each web participant can choose a different allowed output language. Fast detects the spoken source automatically.
Avoid repeating speech in the same language#
Leave Same-language output off when you do not want Japanese speech repeated to a Japanese listener, or English speech repeated to an English listener. JotMe checks source-language and source/output text evidence for each direction or participant language, suppresses matching translated voice and captions, and keeps the original captions. Later corrections can clear a translated caption and cancel audio that is still queued.
Short or ambiguous speech may still pass through before there is enough evidence. Audio already heard cannot be taken back. The playback-quality setting above does not guarantee language detection. Enable Same-language output only when you deliberately want output in the source language too.
Other#
Use Other when there is no meeting app to connect to. It uses audio that your computer can hear directly:
- Microphone audio for an in-person conversation, meeting, or event.
- System audio for a training video, YouTube, Netflix, or other media playing on this computer.
No virtual audio setup
Other does not use BlackHole, VB-CABLE, VAC, or any other virtual audio device. Choose an audio source in JotMe and start translating.
Note
The audio source is locked during a session. Stop recording before switching between Microphone audio and System audio.
For participants#
If the people who need translated audio are guests rather than you, they can listen in their browser and you do not need to route translated audio into a meeting app. Turn on voice translation for participants, then see Captions or voice for participants.