Hi there! I recently purchased the Mod Dwarf, and I’m really excited about it! However, I’m experiencing a minor issue: when I load the Neural Amp Modeler plugin and select a .nam file, the CPU immediately jumps to 100%. The Mod disconnects and restarts, automatically removing the plugin. During those few seconds before the device restarts, the sound is completely glitched.
I’ve tried inserting the NAM in an entirely empty pedalboard, and I’ve attempted setting the Buffer Size to 256, but the issue persists.
Am I doing something wrong? Could there be an issue with my Dwarf?
It’s worth noting that both the Firmware and the Plugin are properly updated to the latest version.
The architecture of the NAM is very heavy and not really suited to run on devices other than a powerful PC, unless you use one of the nano models which are special NAM models of much lighter weight.
NAM is not especially optimized as it’s normally meant to run on very powerful CPUs.
For running on embedded devices, you are far better off using the AIDA-X neural modeler which is using a way better optimized architecture and can offer an equivalent performance with a much smaller footprint.
I am trying out 64 bit Patchbox on my RPi5 and so far it works great however no Neural Amp Modeler folder has been created after installing the plugin. By SSH, can I create a folder somewhere to make this work?
I will try this out more in the coming weeks and if it works out well I will make a donation for sure.
So as I correctly understand, with a Mod Duo (1.13.1.3122), there is no way to upload Nam files to the File Manager, as there is no dedicated folder for it?
Sames goes for the Aidax DSP where I can’t upload .aidax files.
Just tested with .json files, this works.
Hi, I build locally with MPB the latest neural-amp-modeler from @MikeOliphant and the results are great!
CPU for nano model around 42% in Dwarf. And feather models now works! but with 76% CPU. Also changing models with snapshots works much better, no xruns or clicks or pops.
The downside is that the buffer option does not work, but you can use the portal anyway.
Hope the update reaches the Plugin shop.
For the adventurers, I just changed the .mk file to point the latest release and the only error was in the submodule RTNeural-NAM, something about C++ compatibility with the span include. I was surprise it was just working adding some file from GitHub - tcbrindle/span: Implementation of C++20's std::span for older compilers…
Keep it rocking! Maybe with a little more optimization we will reach the next size nam models
I couldn’t test Aida-X models in your plugin, as it is not showing any Aida-X model in the list. Probably mod-ui only expects one type of file for each plugin, but I had not look in detail… at the moment I’m just happy with nam models.
@fer You might want to check out the latest version of the lv2 repo - you may see significant performance gains for both NAM and Aida models on the Dwarf.
I tried in my Dwarf and there is some performance improvement, but not much since your last version using RTNeural. Now Nano NAM models shows around 39% CPU and Feather around 69%. But still not possible to load Lite or Standard models.
It is all very architecture-specific. For example, the RTNeural implementation worked well on x64, but did not do well at all on my Raspberry Pi 4. I would have thought the Dwarf would be more similar to the Pi, but
You may be able to get a bit more performance if you grab the latest version of the repo and add “-DWAVENET_FRAMES=128” to your CMake commandline. Or it could be worse…
How can we find the nano captures? in the link in this thread its sending me to the tonehunt website. I assumed they changed the website and now it just show the home. I could not find the option #nano-mdl in the search filters.
… now nano 36% and Feather 66% with my Dwarf at 128 frames.