Question Unity Windows Speech Recognition is more accurate than Using System.Speech.Recognition in CSharp?!

Hello All. Im getting out of ideas. I need to build a Windows-Application within CSharp for Speech Recognition. SpeechToText. But if i use this (Built in in Windows) it's very inaccurate. Even i train the whole time with it. It's also inacurrate using it directly in windows to input some dictation. Now my question. If i use Unity.Windows.Speech in Unity. Its very nice and accurate. But i know it's based on the above System.Speech i mention. What i try to find out is what is the difference. Is there a special grammar used in Unity? And when? Can i access that to build my app outside of unity? I need a solution to speech recognize in background. Speech Reco in Unity only works when it have focus. When it runs its super accurate. So how can i use this accuracy for my background project. I know its a special question. But i already investet hours for this. Thanks for any advice.

 

You can try the Recognissimo plugin on the Asset Store or use the VOSK library. I found windows speech to be obtuse unless using an Azure server and I do not want to give any of the oligarchs any of my business.