cross-posted from: https://lemmy.blahaj.zone/post/21928256
Hello girlies and other folk,
my boyfriend is programming a desktop app that displays your current voice input’s frequency in Hz over time with a real time graph, similar to the app “voice tools” many of us use for voice training.
I’m trying to garner interest for such a desktop app and would appreciate input about it so I can show him that it’s not something only I would want.
I would also be interested in the OS you would be using, since currently it’s only on Linux (as we use arch btw).
The image shows what it’s currently looking like and the settings window. The entire point is for it to be always on top of everything else so you can always see how you’re doing.
And for the other nerds: it’s written in Python (making it quite large, about 2GB, he’s trying to port it to Rust (based) and make it smaller)
Will Rust make voice training blazingly fast? 🥺
I doubt it, sadly 😭
The current performance is already good, it’s just a huge file with python
Awesome project! Python can be like that sometimes and it’s frustrating. Easy as anything to work with though.
Looking forward to seeing a live demo eventually, perhaps using a tone generator 😊🏳️⚧️
Pretty cool! How is the frequency detection performed, though? While Python typically has larger files sizes, 2GB seems like an extreme amount.
Will it be open source? Curious to learn more about it. 😇
Of course it will be, the size is due to tensorflow which handles pitch detection
Could you elaborate on how Rust will decrease the binary size? You cannot meaningfully decrease the neural network, right?
Not trying to hate, just curious.
I think its because of the python installer and it needing more dependencies to be fully present? I’m working on getting the current build out of him, I hope that will allow others to have a more educated look than me.
This is awesome. Thanks for sharing.
Would totally use the heck out of this linux app after I start voice training
