new version GuitarMidi-lv2 v2.1
gm-linuxaudio.privatize920-OGrP24aSwZHk1uMJSBkQmQ@public.gmane.org Thu, 28 May 2026 16:31:57 +0200
| Newsgroups | gmane.linux.audio.users |
|---|---|
| Message-ID | <[email protected]> |
Hi, version 2.1 is out with some improvements. ## CPU performance - implemented performant filterbank - removed dspfilters submodule ## Detection - added note detection volume to reduce false positives ## Other improvments and new contributors By Tomtiger66: Fix hardcoded tflite path: use LV2 bundle_path instead of /usr/lib/lv2=20 Get it at https://github.com/geraldmwangi/GuitarMidi-LV2/releases/tag/v2.1 Regarding the mod dwarf, I=E2=80=99ve improved the plugins runtime performa= nce on the cpu so that it runs on the mod without xruns However latency is still very high ~50ms. This is due to the neural net having ~800k parameters. I cant release it this way. I have thought about the implementation of a monophonic but that is an entirely different model that needs to be designed, trained and tested. taking time away from my task to provide rock solid polyphony for linux and linux-powered devices. And in some way this move feels backward since the core tenet of the plugin is that you should be able to seemlessly switch from solos to chords and this I have achieved. There is hope though. I have ordered the coral usb accelerator https://botland.store/google-coral/16911-google-coral-usb-accelerator-edge-= tpu-ml-arm-cortex-m0-193575021935.html for which tensorflow lite has support, I think this is the way forward, my plan is anyway to support new guitar gear that have tensor processing units. Thanks for your understanding=20 Regards,=20 Gerald