Samuel Proulx

For anyone else struggling, the best thing to do is use Ollama for vectorization. Works faster even with CPU only, and never gets these errors.

NVDA was going to create a way to load old synth drivers into the 64-bit NVDA. But I'm not seeing any information about this in the developer's guide.

The solution is creating a 32-bit process to load the DLL, then using RPC to shuttle data between the processes. It will, however, introduce a lot of lag.

I took a stab at this, for threshold eloquence. However, say all doesn't work, and I don't know why. https://github.com/fastfinge/eloquence_64/

Fixed. Forgot a subdirectory in the build script. Sorry! Classic "It works on my machine".

What does eloquence-host.log in the NVDA addon folder show? Also, we should probably move into the repo I created for this work so as not to spam this issue.

Enabled issues on the repo. And made a new PR that fixes it? Maybe? I'm baffled; it's working on both my machines.

Correct. The project is laid out like this: * host_eloquence32.py: runs in its own 32-bit python process, loads eci.dll, and receives commands from NVDA and sends back audio. * _eloquence.py:...

Thanks, excellent work! I have merged your code, and also fixed say all. However, when using say all on the web, the audio can become choppy. But this otherwise works....

I am having this exact issue. I have the following network: * interfree, a server hosted in a data center that can act as a lighthouse * Giant2: my home...