Found this today and will compile it immediately!

[38github]

I will report back soon...

[38github]

Using the guide I compiled it in Ubuntu 23.10 (arm64/aarch64). There are a few things that I am missing and would personally change.

Critical:

• Resampling doesn't seem to be implemented but that might be, if I am not wrong, implemented from the iPlug2 library in the official plugin which you obviously is not using.

• CPU Usage is very high. It is almost thrice compared to neural-amp-modeler-lv2 and official plugin. Cmake doesn't seem to care about exported CFLAGS CXXFLAGS etc like other Cmake builds I have done.

• Low and hi cut filters do not cut at 20 and 20000 when using higher sample rates which I like BUT can they be set to showing OFF when they are in their default positions? I use quite a few high sample rate models and like unfiltered signals and be sure about it.

Less important but still nice to have:

• CLAP and LV2. As far as I know JUCE supports these and would probably be appreciated by quite a few users.
• JUCE submodule. This would make compiling easier and take less time for end user.
• Buttons' font size increase. These are difficult to read .
• Brown and orange is beautiful I think but it, too me, makes reading difficult.

If these things are changed then a lot of people who have been hoping for an iPlug2 Linux port will be very happy I think. I really look forward to what you will achieve and I will gladly test new commits and report back my findings.

Regards

[Tr3m]

Hey there, thanks for taking an intererest in my project. Regarding your findings:

Resampling doesn't seem to be implemented but that might be, if I am not wrong, implemented from the iPlug2 library in the official plugin which you obviously is not using.

That's correct, resampling isnt't being implemented in this version. I'm currently working on an update to accomodate some of the changes the official NeuralAmpModelerCore repo has gone through since I started the project, with resampling being top priority. I've been experimenting with the resampler Steven has included in later versions of that repo as well as with the one included in Jatin's chowdsp_utils library.

CPU Usage is very high. It is almost thrice compared to neural-amp-modeler-lv2 and official plugin. Cmake doesn't seem to care about exported CFLAGS CXXFLAGS etc like other Cmake builds I have done.

I haven't had the chance to test performance against other forks but that is certainly concerning and something i'll have to look into. Have you tried building using the Makefile exporter from the Projucer instead of cmake to check if cpu usage is equally high on that build?

Low and hi cut filters do not cut at 20 and 20000 when using higher sample rates which I like BUT can they be set to showing OFF when they are in their default positions? I use quite a few high sample rate models and like unfiltered signals and be sure about it.

That's on oversight on my part. The behaviour your're describing sounds more intuitive and i'll be sure to include it in the next update.

Regadring the less important stuff, most of the things you mention I have thought about as well. More plugin formats are planned for when I tag a release. I'm not 100% happy with the UI either. An overhaul is planned after the more urgent stuff is out of the way.

Thanks a lot for your feedback. I can't promise a time frame when these updates will be delivered at the moment, but it should be sooner rather than later.

[38github]

Thank you for taking the time to respond. I will look into the compiling/Projucer at some point. please tag me when you have updates and maybe would like some feedback.

[Tr3m]

Hey @38github, I've been pushing some updates these past few days on the dev branch addressing some of the issues you've brought up here.

Most notably:

• Resampling has been implemented.
• CPU usage should now be on par with the original plug-in and other forks.
• The output filters are now bypassed when the knobs are on their default positions, with the change being reflected on the GUI side as well.

I would love an extra pair of eyes/ears on the build if you're still down for doing some testing. Check out the README for instructions on how to build, the process should be more straightforward this time around.