TTS-Voice-Wizard

• ElevenLabs updates
  • you can now select the Turbo v2.5 model
  • the option for optimize streaming latency goes up to 5 now
  • legacy voices now show up by default (elevenlabs updated their API and moved many voices to legacy (https://www.reddit.com/r/ElevenLabs/comments/1e00gbo/important_elevenlabs_default_voices_update_on/)
• Lyrics variable works with windows media integration now (more info on this integration here)
• HRPercent is now calculated properly

• Word Boundaries option for Word Replacement tab
  • Isolates whole words so that they can be replaced without replacing parts of other words.
  • enabled by default

• Azure explicit punctuation: https://github.com/VRCWizard/TTS-Voice-Wizard/issues/76
• TikTok errors will give a full stack trace in the log to better troubleshoot connection issues

• VRChat keyboard will no longer be displayed for whitespace/empty strings (when the option is enabled)
• Option added to disable speech to text translations from azure
• Upon an unsafe launch a permanent save is now created for the last good backup file when it is loaded
• Counters Tab updated
  • large quantities of OSC data will no longer crash TTS Voice Wizard when counters are enabled
  • you now need to click the apply changes button upon changing any of the text fields in the Counter Tab
  • more counter slots

Lyrics can now be sent from BigAtomikku/VRC-Spotify-OSC to TTS Voice Wizard

1. When setting up the config file set the sent port to 4026 (very important step)
   
2. In TTS Voice Wizard Activate the OSC Listener on the same port 4026
   
3. In TTS Voice Wizard use the {lyrics} variable to output lyrics
4. Note the chatbox can only update as quick as 1500ms so if a song has really fast lyrics (think Eminem) some lyrics will be skipped
   

• fixes for using chatgpt integration

v1.6.8.2

• fix for naming inconsistency (same as previous version)

v1.6.8.1

• minor fixes to chatgpt integration tab

v1.6.8

• OpenAI TTS
  • Available with Pro and with your own API key
• ChatGPT Integration (Text to ChatGPT to Speech)
  • Available with Pro and with your own API key. (with pro you can send a prompt and your current message. With your own key you can send full unlimited context)
• Code optimization to reduce response times, a couple ms have been shaved off of several TTS and STT options
  • Faster responses from pro tts options and deepgram speech to text
• VRChat chatbox typing indicator toggle
• Minor UI changes
  • Deepgram silence scale
  • tooltips added for some ui elements
• Bug fixes
  • fixed issue when switched stt methods when whisper was running

v1.6.7.3

• Deepgram Aura TTS (english voices only) available on the dev branch of Voice Wizard Pro
  • Listen to voices here: https://developers.deepgram.com/docs/tts-models
• Added button to add progressbar for media integration
• Added tooltips to media integration buttons and some general setting buttons
• Added Albanian language for TTS though Azure/Pro and Speech to Text via Azure.
• Added "/TTSVoiceWizard/TextToText" OSC endpoint
• hotkeys no longer trigger multiple times when held down
• Removed zh-CN-XiaoxuanNeural voice from azure voice list

v1.6.8.1

• minor fixes to chatgpt integration tab

v1.6.8

• OpenAI TTS
  • Available with Pro and with your own API key
• ChatGPT Integration (Text to ChatGPT to Speech)
  • Available with Pro and with your own API key. (with pro you can send a prompt and your current message. With your own key you can send full unlimited context)
• Code optimization to reduce response times, a couple ms have been shaved off of several TTS and STT options
  • Faster responses from pro tts options and deepgram speech to text
• VRChat chatbox typing indicator toggle
• Minor UI changes
  • Deepgram silence scale
  • tooltips added for some ui elements
• Bug fixes
  • fixed issue when switched stt methods when whisper was running

v1.6.7.3

• Deepgram Aura TTS (english voices only) available on the dev branch of Voice Wizard Pro
  • Listen to voices here: https://developers.deepgram.com/docs/tts-models
• Added button to add progressbar for media integration
• Added tooltips to media integration buttons and some general setting buttons
• Added Albanian language for TTS though Azure/Pro and Speech to Text via Azure.
• Added "/TTSVoiceWizard/TextToText" OSC endpoint
• hotkeys no longer trigger multiple times when held down
• Removed zh-CN-XiaoxuanNeural voice from azure voice list

v1.6.8 (current)

• OpenAI TTS
  • Available with Pro and with your own API key
• ChatGPT Integration (Text to ChatGPT to Speech)
  • Available with Pro and with your own API key. (with pro you can send a prompt and your current message. With your own key you can send full unlimited context)
• Code optimization to reduce response times, a couple ms have been shaved off of several TTS and STT options
  • Faster responses from pro tts options and deepgram speech to text
• VRChat chatbox typing indicator toggle
• Minor UI changes
  • Deepgram silence scale
  • tooltips added for some ui elements
• Bug fixes
  • fixed issue when switched stt methods when whisper was running

v1.6.7.3

• Deepgram Aura TTS (english voices only) available on the dev branch of Voice Wizard Pro
  • Listen to voices here: https://developers.deepgram.com/docs/tts-models
• Added button to add progressbar for media integration
• Added tooltips to media integration buttons and some general setting buttons
• Added Albanian language for TTS though Azure/Pro and Speech to Text via Azure.
• Added "/TTSVoiceWizard/TextToText" OSC endpoint
• hotkeys no longer trigger multiple times when held down
• Removed zh-CN-XiaoxuanNeural voice from azure voice list

• Deepgram Aura TTS (english voices only) available on the dev branch of Voice Wizard Pro
  • Listen to voices here: https://developers.deepgram.com/docs/tts-models
• Added button to add progressbar for media integration
• Added tooltips to media integration buttons and some general setting buttons
• Added Albanian language for TTS though Azure/Pro and Speech to Text via Azure.
• Added "/TTSVoiceWizard/TextToText" OSC endpoint
• hotkeys no longer trigger multiple times when held down
• Removed zh-CN-XiaoxuanNeural voice from azure voice list

• Fix for playing local files on spotify that don't have an album artist

v1.6.7 (current)

• Updated the CoreOSC package to allow for receiving utf8 characters via OSC
• Changed the elements of the progress bar variable {progressBar E:◯ L:13} to use more fitting utf characters. ┣━━━━┫ vs |-----------|
  • You will likely need to lower the length (L) variable for the change in characters. A length of 13 will fill a line of the VRChat chatbox.
    

v1.6.6.8

• 2 optional override bools added for OSC to TTS endpoint /TTSVoiceWizard/TextToSpeech
  • bool useChatbox
  • bool useKillFrenzyAvatarText

The bools allow you to optionally forward text the VRC Chatbox or Killfrenzy Avatar Text without adjusting the in app settings. When sending an OSC message to this endpoint you may still send just a string or both bools (sending only one of the bools will not override the settings).

Example of possible combinations in TouchOSC:

v1.6.6.7

• fixed vrchat counter causing timeout if different counters were activated in quick succession.
• spotify exceptions should always give a message.

v1.6.6.5

• minor bug fixes
• ElevenLabs turbo model added
• removed voices for TikTok API that no longer functioned (optimus prime voices)
• New Azure Voices:

es-ES | Spanish (Spain) | XimenaNeural (Female)
fr-CA | French (Canada) | ThierryNeural (Male)
fr-FR | French (France) | VivienneNeural (Female)
it-IT | Italian (Italy) | GiuseppeNeural (Male)
ko-KR | Korean (Korea) | HyunsuNeural (Male)
pt-BR | Portuguese (Brazil) | ThalitaNeural (Female)
zh-CN-XiaoxiaoDialectsNeural
de-DE | German (Germany) | de-DE-FlorianMultilingualNeural (Male)
de-DE | German (Germany) | de-DE-SeraphinaMultilingualNeural (Female)
en-US | English (United States) | en-US-AvaMultilingualNeural (Female)
en-US | English (United States) | en-US-EmmaMultilingualNeural (Female)
fr-FR | French (France) | fr-FR-RemyMultilingualNeural (Male)
en-US | English (United States) | en-US-BrianMultilingualNeural (Male)
en-US | English (United States) | en-US-AndrewMultilingualNeural (Male)
fr-FR | French (France) | fr-FR-VivienneMultilingualNeural (Female)
zh-CN | Chinese (Mandarin, Simplified) | zh-CN-XiaoxiaoMultilingualNeural (Female)
zh-CN | Chinese (Mandarin, Simplified) | zh-CN-XiaochenMultilingualNeural (Female)
zh-CN | Chinese (Mandarin, Simplified) | zh-CN-YunyiMultilingualNeural (Male)

v1.6.5.2

• uberduck voice selection fix (although some voices in the list no longer function)
• now using latest version of windows media controller, (report any crashing and include event viewer details)
• {albumArtist} added for windows media and Spotify media integration

v1.6.5

• vrchat "/chatbox/typing" indicator added for whisper and deepgram (if VAD is active)
• fix for whisper timestamps being offset for VAD so super short activations wouldn't be recognized
• possible fix for group in word replacement tab breaking on some computers
• hotkeys backend refactored, new hotkeys added for changing voice presets quickly

• 2 optional override bools added for OSC to TTS endpoint /TTSVoiceWizard/TextToSpeech
  • bool useChatbox
  • bool useKillFrenzyAvatarText

The bools allow you to optionally forward text the VRC Chatbox or Killfrenzy Avatar Text without adjusting the in app settings. When sending an OSC message to this endpoint you may still send just a string or both bools (sending only one of the bools will not override the settings).

Example of possible combinations in TouchOSC:

• fixed vrchat counter causing timeout if different counters were activated in quick succession.
• spotify exceptions should always give a message.

v1.6.6.5 (current)

• minor bug fixes
• ElevenLabs turbo model added
• removed voices for TikTok API that no longer functioned (optimus prime voices)
• New Azure Voices:

es-ES | Spanish (Spain) | XimenaNeural (Female)
fr-CA | French (Canada) | ThierryNeural (Male)
fr-FR | French (France) | VivienneNeural (Female)
it-IT | Italian (Italy) | GiuseppeNeural (Male)
ko-KR | Korean (Korea) | HyunsuNeural (Male)
pt-BR | Portuguese (Brazil) | ThalitaNeural (Female)
zh-CN-XiaoxiaoDialectsNeural
de-DE | German (Germany) | de-DE-FlorianMultilingualNeural (Male)
de-DE | German (Germany) | de-DE-SeraphinaMultilingualNeural (Female)
en-US | English (United States) | en-US-AvaMultilingualNeural (Female)
en-US | English (United States) | en-US-EmmaMultilingualNeural (Female)
fr-FR | French (France) | fr-FR-RemyMultilingualNeural (Male)
en-US | English (United States) | en-US-BrianMultilingualNeural (Male)
en-US | English (United States) | en-US-AndrewMultilingualNeural (Male)
fr-FR | French (France) | fr-FR-VivienneMultilingualNeural (Female)
zh-CN | Chinese (Mandarin, Simplified) | zh-CN-XiaoxiaoMultilingualNeural (Female)
zh-CN | Chinese (Mandarin, Simplified) | zh-CN-XiaochenMultilingualNeural (Female)
zh-CN | Chinese (Mandarin, Simplified) | zh-CN-YunyiMultilingualNeural (Male)

v1.6.5.2

• uberduck voice selection fix (although some voices in the list no longer function)
• now using latest verson of windows media controller, report any crashing and include event viewer details
• {albumArtist} added for windows media and spotify media integration

v1.6.5

• vrchat "/chatbox/typing" indicator added for whisper and deepgram (if VAD is active)
• fix for whisper timestamps being offset for VAD so super short activations wouldn't be recognized
• possible fix for group in word replacement tab breaking on some computers
• hotkeys backend refactored, new hotkeys added for changing voice presets quickly

This is a demo of a different whisper library, it has not been thoroughly tested so there are likely bugs

v1.6.6.1 (current)

• more robust output purification and removal
• metrics on by default for this version
• metrics changed to be more visible and distinguishable
• fixed model download loop for models that were considered too small to be real

v1.6.6

• Fixed memory leak in previous version (which caused crashing, oopsie)
• Min VAD Duration (s) and Min Avg Confidence have been added to help fight against hallucinated outputs, you can enable debug text to see everything that gets filtered out

v1.6.5.9

• Whisper Library Changed from https://github.com/Const-me/Whisper to https://github.com/sandrohanea/whisper.net (so complete rework behind the scenes)
• Whisper can be "Recognize Once" or continuous
• Whisper model selection now uses absolute path
• Whisper can be GPU (and CPU) or just CPU
• Whisper now supports most if not all models here https://huggingface.co/ggerganov/whisper.cpp/tree/main (including quantized models)
• Whisper now supports distil models: https://huggingface.co/collections/distil-whisper/distil-whisper-models-65411987e6727569748d2eb6

This is a demo of a different whisper library, it has not been thoroughly tested so there are likely bugs

v1.6.6 (current)

• Fixed memory leak in previous version (which caused crashing, oopsie)
• Min VAD Duration (s) and Min Avg Confidence have been added to help fight against hallucinated outputs, you can enable debug text to see everything that gets filtered out

v1.6.5.9

• Whisper Library Changed from https://github.com/Const-me/Whisper to https://github.com/sandrohanea/whisper.net (so complete rework behind the scenes)
• Whisper can be "Recognize Once" or continuous
• Whisper model selection now uses absolute path
• Whisper can be GPU (and CPU) or just CPU
• Whisper now supports most if not all models here https://huggingface.co/ggerganov/whisper.cpp/tree/main (including quantized models)
• Whisper now supports distil models: https://huggingface.co/collections/distil-whisper/distil-whisper-models-65411987e6727569748d2eb6

This is a demo of a different whisper library, it has not been thoroughly tested so there are likely bugs

• Whisper Library Changed from https://github.com/Const-me/Whisper to https://github.com/sandrohanea/whisper.net (so complete rework behind the scenes)
• Whisper can be "Recognize Once" or continuous
• Whisper model selection now uses absolute path
• Whisper can be GPU (and CPU) or just CPU
• Whisper now supports most if not all models here https://huggingface.co/ggerganov/whisper.cpp/tree/main (including quantized models)
• Whisper now supports distil models: https://huggingface.co/collections/distil-whisper/distil-whisper-models-65411987e6727569748d2eb6

Warning: With this library/implementation for whisper running larger models will now cause a much more noticeable delay before output

• uberduck voice selection fix (although some voices in the list no longer function)
• now using latest verson of windows media controller, report any crashing and include event viewer details
• {albumArtist} added for windows media and spotify media integration

• vrchat "/chatbox/typing" indicator added for whisper and deepgram (if VAD is active)
• fix for whisper timestamps being offset for VAD so super short activations wouldn't be recognized
• possible fix for group in word replacement tab breaking on some computers
• hotkeys backend refactored, new hotkeys added for changing voice presets quickly