Multi language/models support on trigger

[noirbizarre]

In addition to #6 which covers non-english or alternative models retrieval, I would be great if we could have 2 models loaded in parallel and the possibility to trigger them with 2 distinct keyboard shortcuts.

My use case is the following: I am a native French speaker, but I work in English. So depending on the context and the interlocutor, I have to switch between those 2 languages. So it would be very practical to have a shortcut to trigger speech-to-text with an English recognition model and another to trigger it with a French one.

Note

I say 2 because this is my use case, but there is a high probability that some people may need more than 2 so maybe the feature could be just the ability to specify a model with an optional --model parameter to the command voxd --trigger-record

[jakovius]

Oh, that is an interesting feature suggestion. there is likely significant amount of people that could use at least two different languages, readily available to rapidly switch.
I will look into it.
In the meanwhile, language support is now available: try the new build v1.7.0 (introduced multiligual support)

[alcuinus]

Great programme! It works very well with English, even using the small modules. I frequently switch from English to other languages, particularly Italian and French; I notice that with these languages recognition is extremely slow. i was not able to find a language specific model, e.g. ggml-base.it.bin or the like. Am I missing something? At any event, thanks again, because whenever I write English this programme helps a lot! Incidentally: with some modules the programme hangs (e.g. the "medium" generic module). I am on Linux Mint 22.1, Xfce v: 4.18.1.