NyVox 2.4

Some big changes, mostly to the UI, but also some model adjustments and added features.

-New UI. Darker, sleeker, probably going to change for the next release. Red is a very aggressive color. Once I find a color scheme I like then I'll swap over. You can change the UI color in the settings yourself if you want to try different colors.

-New models A and B. B is chosen by default and the pros and cons are as follows.
B: slightly more 'robust' and better at handling unseen words, more 'diverse'.
A: higher 'fidelity' than B but at the cost of 'pronunciation' and 'similarity'.

You can swap the models while NyVox is running to compare them for yourself.

-In addition to the old 100+ voices, you can now try:

-Instant Voice Cloning. Still in the proof of concept stage, it only works for certain voices, particularly ones seen during training. But this is how it will look when it DOES work! Load in an 8 second .wav file of the person you want to sound like!

-Voice Designer. Still under construction, but in the meantime, there is a neat little color palette you can peruse to mix and match voices!

-Record To File. While NyVox is running live, press record. It'll record till you stop then you can save it to a file! Woo!

-Convert File. Load in a .wav file to convert using the voice you have selected or designed. It'll save it to the same directory with _convrted.wav appended to the filename.

-More settings to adjust to get the voice just how you want it.

--Microphone Preamp: for if you need it to better pick up your voice.

--Normalize: Attempts to normalize the input volume to better help the models.

--Change input and output devices while running!

-- Toggle listening to yourself while running!

--Static, Limit Energy: may help with certain voices.

--Deplosive: limit how loud plosives are.

--Highpass Filter: Removes lower frequencies to give a crisper feel to the audio.

--Lowpass Filter: Can be used as a 'bad mic filter'

--CPU: does not work yet on most current hardware, but you're welcome to try.

-Login method has been updated. Just link your patreon to the web portal then click "open NyVox". If the button is not prompting to open NyVox, you may need to run it as administrator at least once so it can set the proper registry keys. 

-Details panel with additional information to view. Input/output wave. Pitch, Energy, Speech data, and process duration.

---------------------------------------------------------------------------

Thank you to all my Patrons, you allow me to work on this and also afford fancy cloud GPUs!

If you have any issues, please feel free to message me on Discord!

Once tested by Patrons, and any needed fixes and updates applied, then I'll add an installer and open this version to the public, but please keep this private to Patrons only for now, probably for a few days or so, thank you!

Research and development is also still going strong. New techniques and technologies are constantly coming out, some of which I intend on using. My end goal is still zero-shot any-to-any, so users can load a short clip of a voice and sound like that person without needing training. Doing that while keeping the model small, fast, and of high quality continues to be a significant challenge, but I do have several pathways to get it done and I intend on trying all of them!

Thanks again!