Translation Guide
Convert speech from any supported language to English text output.
What is Translation Mode?
MurmurTone's translation feature uses Whisper's built-in translation capabilities to convert spoken audio from any of 17 supported languages directly into English text. This happens in real-time as you speak.
Unlike separate transcription and translation services, Whisper performs speech-to-English-text in a single step, which often produces more natural results than transcribing first and then translating.
Translation vs Transcription
Choose the right mode for your use case:
Transcription (Default)
Use when you want text in the same language you're speaking.
- Speak Spanish, get Spanish text
- Speak Japanese, get Japanese text
- Preserves original language
Translation
Use when you want English text regardless of what language you speak.
- Speak Spanish, get English text
- Speak Japanese, get English text
- Always outputs English
Common Use Cases for Translation
- Bilingual communication: Think in your native language, produce English documents
- Note-taking: Take notes in your preferred language, get English output for sharing
- Learning English: Practice speaking and see how your words translate
- International teams: Contribute in your language to English-language projects
Supported Languages
MurmurTone can translate from any of these languages to English:
| Language | Code | Notes |
|---|---|---|
| Auto-detect | auto | Recommended - automatically detects your language |
| English | en | Useful for heavy accents or non-native speakers |
| Spanish | es | All Spanish variants supported |
| French | fr | |
| German | de | |
| Italian | it | |
| Portuguese | pt | Brazilian and European Portuguese |
| Dutch | nl | |
| Russian | ru | |
| Chinese | zh | Mandarin; Simplified and Traditional |
| Japanese | ja | |
| Korean | ko | |
| Arabic | ar | Modern Standard Arabic |
| Hindi | hi | |
| Polish | pl | |
| Turkish | tr | |
| Vietnamese | vi | |
| Thai | th |
Model Recommendations
Translation quality varies significantly by model size. Larger models produce more accurate and natural-sounding translations.
| Model | Translation Quality | Best For |
|---|---|---|
| Quick (tiny) | Basic | Simple phrases only; may miss nuances |
| Standard (base) | Good | Casual use; simple sentences |
| Recommended (small) | Very Good | Most users; good balance of speed and quality |
| Professional (medium) | Excellent | Complex sentences; technical content |
| Studio (large-v3) | Best | Maximum accuracy; professional translation |
How to Enable Translation
- Right-click the MurmurTone icon in your system tray
- Select Settings
- Go to the Recognition tab
- Scroll down to the Translation section
- Toggle Enable translation on
- Optionally set a specific source language (or leave on Auto-detect)
- Click Save
Once enabled, all your voice recordings will be translated to English instead of transcribed in the original language.
Troubleshooting
Wrong language detected
If auto-detection keeps picking the wrong language:
- Set the specific source language manually in settings
- Speak a few words clearly at the start of each recording
- Try a larger model for better detection accuracy
Poor translation quality
If translations seem inaccurate or unnatural:
- Upgrade to a larger model (Professional or Studio)
- Speak in complete sentences rather than fragments
- Reduce background noise for clearer audio input
- Speak at a moderate pace
Translation not working
If you're still getting text in your spoken language:
- Verify the "Enable translation" toggle is on in Settings
- Save your settings and restart MurmurTone
- Check that you're not in transcription-only mode
Mixed language output
If you're getting a mix of English and your source language:
- This can happen with code-switching (mixing languages while speaking)
- Try to speak in one language at a time
- Set the source language explicitly instead of using auto-detect