With a highlighted piece of audio, I always get a error message, this one was for the separation tool
23:24:54: Audacity 3.4.2
Log Cleared.
23:25:05: Error: Directory ‘C:\Program Files\Audacity\openvino-models’ couldn’t be created (error 5: Access is denied.)
23:25:05: Operation ‘Pre-processing’ took 0.000000 seconds. Poll was called 0 times and took 0.000000 seconds. Yield was called 0 times and took 0.000000 seconds.
23:25:05: Error: In Music Separation, exception: Exception from src\inference\src\core.cpp:134:
[ GENERAL_ERROR ] Model file C:\Program Files\Audacity\openvino-models\htdemucs_v4.onnx cannot be opened!
23:25:07: Operation ‘OpenVINO Music Separation’ took 0.028000 seconds. Poll was called 2 times and took 0.000001 seconds. Yield was called 0 times and took 0.000000 seconds.
Sometimes the beginning of it working, starts, but immediately an error comes up.
A side issue, I don’t use this forum much, seems everytime I come back here there’s issues logging in, like you’re changing how its done, if so, this is highly annoying.
From your log, it looks like the content of ‘openvino-models.zip’ wasn’t copied correctly. Double check that this folder, C:\Program Files\Audacity\openvino-models exists and contains the file htdemucs_v4.onnx.
It seems to me like the ‘openvino-models’ folder isn’t in the right place, based on the log of the plugin and the output you got from cmd.exe. Double check that this folder exists, and is the one that contains the models (C:\Program Files\Audacity\openvino-models).
A common issue I’ve seen here is that users sometimes extract the zip file in a way that creates a ‘double’ openvino-models folder structure (i.e. C:\Program Files\Audacity\openvino-models\openvino-models).
Honestly, this is a joke. the module folder is in the audacity folder, the openvino is all in there, the preferences all switched to enabled, now theyre all as failed with opups.
Yes that was my problem, a folder within a folder. Changed that and it now all works. Transcription is amazing. How does it translate though, there’s a choice for source language but not the language to translate to?
Glad you were able to get it working! Transcription uses the ‘whisper’ AI model which is sort of biased toward english when it comes to translation – as it can only translate to english.
Sure thing, and just a heads up – we have an upcoming release of these plugins (that will coincide with Audacity’s 3.5.0 release). As part of this update, we’ll support larger whisper models .The version you’re using only supports the very small ‘base’ model, which in general works fine for english and ‘rough’ translations, but the larger whisper models support much more accurate translations from other languages to english. Cheers!
That’s great. Just tried it from Swedish and I agree the translation is quite rough so looking forward to the update. How can I get an alert when it’s ready? Many thanks.
There’s a ‘watch’ button at the top where you can click and choose what kind of updates to subscribe to. One of them is ‘releases’. This is what you want.