Sorry, didn't check that. Try here:
Excuse me if the questions below are silly, but sometimes the translation into Portuguese isn't clear and I have doubts.
- I understand that you said you added yourself as a user to the Render group. Can you tell me how you did that?
- How did you download the necessary models?
I installed Audacity 3.7.5 using the commandsudo snap install audacity --edgewhich I got from the website: Install Audacity on Linux | Snap Store. Audacity already came with OpenVINO in the effects menu, but the separation test I did resulted in an error. The log is below:
09:31:06: File name is /home/dunham-zorin/Downloads/O Grande Momento CHEGOU! Mudou a Trajetoria? [lCI5eHWaw7w].mp4
09:31:06: Mime type is *
09:31:06: Opening with libav
09:31:06: [mov,mp4,m4a,3gp,3g2,mj2 @ 0x63bfa093eec0] stream 0, timescale not set
09:31:06: Open(/home/dunham-zorin/Downloads/O Grande Momento CHEGOU! Mudou a Trajetoria? [lCI5eHWaw7w].mp4) succeeded
09:31:06: [opus @ 0x63bfa08c1fc0] Could not update timestamps for skipped samples.
09:31:07: Operation 'Importing FFmpeg compatible files' took 1.813000 seconds. Poll was called 26013 times and took 0.214919 seconds. Yield was called 39 times and took 0.043902 seconds.
09:31:08: Operation 'Music Information Retrieval' took 0.000000 seconds. Poll was called 0 times and took 0.000000 seconds. Yield was called 0 times and took 0.000000 seconds.
09:31:26: Operation 'Preprocessing' took 0.000000 seconds. Poll was called 0 times and took 0.000000 seconds. Yield was called 0 times and took 0.000000 seconds.
09:31:26: Error: In Music Separation, exception: Exception from inference/src/cpp/core.cpp:121:
Check 'util::directory_exists(path) || util::file_exists(path)' failed at frontends/common/src/frontend.cpp:113:
FrontEnd API failed with GeneralFailure:
ir: Could not open the file: "/snap/audacity/1214/usr/lib/openvino-models/htdemucs_v4.xml"
09:32:04: Operation 'OpenVINO Music Separation' took 0.025000 seconds. Poll was called 2 times and took 0.000006 seconds. Yield was called 0 times and took 0.000000 seconds.
Hi Dunham,
I'm not sure what the --edge in there does? I'm not seeing that on the page you linked. Probably me!
I am not experienced enough to read what the log is telling you - perhaps somebody else can help with that?
I followed the sequence found here:
GitHub - snapcrafters/audacity: Snap of Audacity: Audio software for multi-track recording and editing
sudo snap install audacity (which you have done)
sudo snap install intel-npu-driver
Then install the models using:
sudo audacity.fetch-models
This gives you a prompt in the terminal with instructions - enter the number of the model you wish to download. Then give it time until it tells you its done. I don't have great internet so this took some time.
Now to add yourself to the render group.
To check your current groups, please run the following from a terminal:
groups
If you do not see render listed in the output, you may add your current user with the following:
sudo usermod -a -G render $USER
Replace $USER with your username.
Now log out and back in to ensure your active session has the render group included. I actually had to do a restart at this point. Then check by typing in the terminal again -
groups
you should see 'render' listed in the output. That should be everything complete. Run audacity and try one of the models.
Hello,
Thank you for the information, but unfortunately it didn't work here, because my hardware is not Intel, but AMD.
When I applied the command: sudo audacity.fetch-models
It returned this message:
The OpenVINOā¢ AI models are only supported on Intel hardware. Exiting.
But that's okay, I use other audio applications for other needs, as I didn't use the separation feature very much.
Thank you very much for your help.
Yes, it is a specifically Intel technology. I was previously unaware of it.
For music separation, there are various online tools, and a lot of people on forums I just researched recommend demucs:
GitHub - facebookresearch/demucs: Code for the paper Hybrid Spectrogram and Waveform Source Separation
Online tools eg
Split Vocals and Instruments | RiffGen
Note: I have not tried using any of these tools, just highlighting possible avenues of exporation for your use-case.
Best wishes
Thank you very much. I will test it.
I have been using Photoshop CS2 and CS6 on linux for a long time. They are easy to install via Playonlinux.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.