[Speech] VOICEVOX integration

You can have OneComme's Speech Voice generated by VOICEVOX's speech synthesis engine.

You can download VOICEVOX from the site below.

https://voicevox.hiroshiba.jp/

Caution

Version 0.20.0 or later is required to integrate with OneComme
Each speech synthesis engine has its own license, so make sure you fully check VOICEVOX's terms of use before using it

VOICEVOX settings

To integrate, the VOICEVOX app or VOICEVOX Engine must be running (engines other than VOICEVOX Engine are not supported)
Start the engine using the VOICEVOX app, or vv-engine/run in the VOICEVOX folder

① Connect via Integration

Open Integration from the menu at the top right of OneComme

CommentersStop speechClear commentsSpeechPersistSettingsTemplatesDummy CommenterIntegration

Open the "VOICEVOX" tab

IntegrationYUKARINETTE CONNECTOR NEO / TRANS-THROUGHBouyomiChan

To integrate, check "Enable"

If the VOICEVOX integration succeeds, the Speech Voice list will be displayed
If it isn't, check that VOICEVOX is running, then click the reload button next to the Speech Voice selector to fetch it again

Select a voice and click "Test Connection" — if it reads text aloud, the connection is complete

② Select it as the Text-To-Speech Engine

If the Default Text-To-Speech Engine is still set to "Built-In Voice Engine", VOICEVOX will automatically be selected as the Default Text-To-Speech Engine as soon as you enable the integration

If another engine is already set as the default, it won't switch automatically, so select VOICEVOX using one of the methods below

Open "Speech Settings" and select VOICEVOX under "Default Text-To-Speech Engine" to use VOICEVOX for all streams
If you want to use different engines per stream, select VOICEVOX under "Text-To-Speech Engine" in "Individual options"

If VOICEVOX isn't selected as the engine anywhere, it won't be used for speech even if the integration is enabled

Also, if you disable an integration that was enabled, the Default Text-To-Speech Engine will automatically be reselected from the remaining engines

Settings

IntegrationYUKARINETTE CONNECTOR NEO /TRANS-THROUGHBouyomiChanEnabledExecutable File PathAuto RunHostnameOutput DeviceSpeech VoiceShikoku Metan (SweetSpeech VolumePort10 Speech SpeedSpeech Voice PitchIntonationAutomatic Speed Adjustment17 Maximum Queued ItemsTest ConnectionAccelerated Character Count

① EnableEnable speech using VOICEVOX
② Auto RunStarts VOICEVOX when OneComme starts
If you don't need the GUI, we recommend setting the engine's executable file
※If auto-started, it will automatically close when OneComme closes
③ Executable File PathSpecify VOICEVOX's executable file (VOICEVOX.exe, run.exe, etc.) (※1)
④ HostnameHostname to connect to (usually no need to change)
⑤ PortPort number of VOICEVOX's local HTTP server (usually no need to change)
⑥ Output DeviceSelect the device to play the Speech Voice on
If not selected, it will play on the default device
⑦ Speech VoiceVoice used for speech
⑧ RefetchRefetch the Speech Voice list from VOICEVOX
⑨ Speech VolumeAdjust speech volume
⑩ Speech SpeedAdjust speech speed
⑪ Speech Voice PitchAdjust speech pitch
⑫ IntonationAdjust speech intonation
⑬ Taking A BreathAdjust breathing during speech
⑭ Automatic Speed AdjustmentAutomatically adjusts speech speed according to length
⑮ Acceleration CoefficientIncreases speed by the specified coefficient for every ⑯ characters
⑯ Accelerated Character CountIncreases speed for every specified number of characters
⑰ Maximum Queued ItemsAutomatically skips older items once the specified number has built up, to prevent speech from backing up too much
⑱ Test ConnectionRun a speech connection test

※1 - On Mac, you can set this automatically using the steps below

  1. Right-click the VOICEVOX app to open the menu
  2. While the menu is open, hold down the option key
  3. Select "Copy '"VOICEVOX"' as Pathname"
  4. Paste the copied path into the Auto Run executable file path field

Specifications --

While VOICEVOX is selected as the Text-To-Speech Engine, the settings in the red box in the image below, under Settings > Speech (Built-In Voice Engine Settings), are not used

SettingsSpeechCommentsWebsite specificNotificationOthersSpeech VoiceSpeech VolumeSpeech SpeedSpeech Voice Pitch

Settings such as whether to include the name in the spoken text (Syntax for speech) and character limits continue to work regardless of which engine is selected

Use the preview to check that everything works as expected

Per-stream speech settings

If you select VOICEVOX under "Text-To-Speech Engine" in Individual options, you can override the Speech Voice and settings for each stream

For details, see "Individual options"