Plugins SpeechRecognition

[MacGruber]

MacGruber updated SpeechRecognition with a new update entry:

Version 3 (free)

Changelog

• Fixed null reference exception that was apparently introduced due to an API change with VaM 1.21 or 1.22. This actually made the GrammarRecognizer unusable when the plugin was added to a new scene. Thx "Will" for reporting
• Fixed another silent null reference exception.
• Updated demo scene to match new plugin version.

Read the rest of this update entry...

 

[bill_prime]

@MacGruber Quick question:

Do you foresee any issue with having multiple instances of this plugin in a scene? I am trying to manage different states and it would be easier to manage if I have multiple instances.

 

[MacGruber]

Do you foresee any issue with having multiple instances of this plugin in a scene?

I have not tried, at least I don't remember. Quick glance through the code says it might just work. Don't be surprised, if you run into trouble, though. However, as far as I see by looking at this for 2 minutes is that you would receive error messages (e.g. bad XML input) multiple times. That's because each plugin registers itself to that error callback.

 

[bill_prime]

I have not tried, at least I don't remember. Quick glance through the code says it might just work. Don't be surprised, if you run into trouble, though. However, as far as I see by looking at this for 2 minutes is that you would receive error messages (e.g. bad XML input) multiple times. That's because each plugin registers itself to that error callback.

Okay cool, thanks for checking! Seems like it might just be working haha. Will do some more testing.

 

[Saint66]

I just cannot get this to work on Windows11. Correct Microfone is set as default, Windows, Whisper and other tools recognize my speech, language is set to English US.
I have created a simple scene with one person atom, added the plugin to either session or scene plugins, set up a simple phrase with keyword "hello" and applied a simple undress trigger.

Also demo scene is not reacting. I am out of ideas, maybe it's just the insider version of Win11? But what makes me curious, I can use speech input in every other tool I tried.

 

[MacGruber]

Well, all those VaM speech plugins are based on Unity's Win10 integration. The Unity version VaM uses is from 2018, which might as well be a century old. It likely simply doesn't support Win11.

 

[Cervihel]

I installed it on the newest version of VaM. Started your included testscene and my voice wasnt recognized. in SpeechRecognition plugin it says "Apparently speech recognition is not supported on your machine". I have Windows 10 Pro 64bit. Cortana and speechrecognition in windows work and are functional. What do i miss?

Also the default mic in windows settings is set correctly

 

[MacGruber]

I installed it on the newest version of VaM. Started your included testscene and my voice wasnt recognized. in SpeechRecognition plugin it says "Apparently speech recognition is not supported on your machine". I have Windows 10 Pro 64bit. Cortana and speechrecognition in windows work and are functional. What do i miss?

Also the default mic in windows settings is set correctly

Check the Notes/FAQ section on the resource page. Possibly missing language packs.

 

[Cervihel]

Check the Notes/FAQ section on the resource page. Possibly missing language packs.

Just checked because i have tested alot of settings, English (United States) ist the installed and selected option.
What else could it be?

 

[huiyu914]

I encountered a problem where my voice commands actually take effect randomly when I don't have the plugin enabled in the game.

For example, while playing other games, I opened the VAM scene in the background, but I did not check the plugin to enable it. However, my voice will randomly take effect in the VAM. For example, when I say clothing, it will automatically change the character's clothing

love your plug respect

 

[MacGruber]

I encountered a problem where my voice commands actually take effect randomly when I don't have the plugin enabled in the game.

For example, while playing other games, I opened the VAM scene in the background, but I did not check the plugin to enable it. However, my voice will randomly take effect in the VAM. For example, when I say clothing, it will automatically change the character's clothing

love your plug respect

Why on earth do you have VaM running in the background while playing other games? It sounds like expected behavior....VaM is running, so it does stuff?

 

[huiyu914]

Why on earth do you have VaM running in the background while playing other games? It sounds like expected behavior....VaM is running, so it does stuff?

I am a scene author and often use VAM to create scenes. However, during breaks, I go out to watch videos, browse web pages, and research plugins. However, I do not exit VAM because there is no need to reopen it every time, which is time-consuming. But when VAM is in the background, it will also randomly and uncontrollably respond to voice commands. For example, if I chat with someone and say a sentence about clothes, the game will suddenly change clothes, even if I remove the hook of the plugin startup in the game

 

[MacGruber]

even if I remove the hook of the plugin startup in the game

If the plugin is already started, that does not do anything. However, you can disable the plugin, which internally deactivates all phrases. This can also done by trigger.

If you keep VaM running in the background, I strongly recommend you use my FrameRateControl plugin (part of Essentials), which has the "Pause VaM when not focused" setting. That prevents wasting CPU/GPU resources on VaM when you don't actually use it....so you got more resources for whatever game you are playing....or use less power, meaning less fan noise and heat. However, this likely does not affect SpeechRecognition, since the actual recognition happens in Windows and is not affected by this (but have not tried).

 

[Flgamer]

Speech recognition periodically refuses to see my Oculus microphone even though I have it set as Windows default. In fact, I've had the plugin fail in the middle of a session after working for over an hour. It just arbitrarily stops hearing my commands. When I do the test response, it works but not with voice commands.
I cannot find any legitimate reason for this behavior.

I depend a lot on this plugin because I create interactive scenes where I have conversations with my models and use the plugin to issue pose and animation commands. I haven't shared any of these with the hub as yet.

I've found that with the proper triggers you can do anything with speech recognition that you can do with UI buttons. It can create a very immersive environment in VR.