Hi all,
During the presentation video for the Tab S10, where once again the only new feature seems to be “Galaxy AI,” I started questioning Samsung’s voice transcription. Is it just me, or is it seriously flawed? It barely works when I use voice typing in Portuguese, and when transcribing audio from meetings, it produces such nonsensical words that it’s hard to believe it’s AI. I’m not sure if Google’s voice input would be any better, though I worry about my data being used by Google. Is there any way to “calibrate” this voice input?
I think it’s possible to download different versions of English. For instance, my default transcription language is “English UK,” but I’ve found “English US” to be much more accurate for transcribing non-native English spoken here in the Middle East.
It also takes some time for the AI to adjust to your specific use case. I mainly transcribe Teams conversations with audio from my laptop speaker. If I switch to a live conversation in a large meeting room, the AI takes a few minutes to adapt because the voices sound different.
Samsung voice input is best offline on device input service period , google voice input is completely online(only on pixels it’s on device), so it’s more accurate, but it’s not worth it coz of the privacy tradeoff that comes with gboard and google voice input.