Release v0.1.6

• Added input calibration for models that support it
• CMake options for using RTNeural for NAM models (defaults to off for WaveNet and on for LSTM)
• Improved RTNeural WaveNet performance
• Added ".aidax" file extention

If you are downloading binaries, make sure you get the correct version. The archives ending in "v3" have optimizations for (relatively) recent x64 CPU architectures - they will not run on older machines.

Release v0.1.5

• Use RTNeural for NAM LSTM models (approx. 2x performance increase)
• Added support for RTNeural keras json models (ie: Aida-X models)
• Updated to latest NAM core
• Removed dc blocker

Release 0.1.4

• sync to latest NeuralAmpModelerCore (fixes WaveNet head check on some systems)
• make dc blocker sample rate dependent
• x64 optimizations are now off by default

0.1.3

• Simplified NAM processing loop
• Handle model loudness normalization in plugin

0.1.2

Added 5Hz high pass filter to clean up any DC offset created by the model.

0.1.1

Fixed model display after preset loading.

0.1.0

What's Changed

• Create LICENSE by @mikeoliphant in #2
• Use fabsf to ensure no double promotion by @falkTX in #7
• Add LV2 save/restore of model path by @mikeoliphant in #12
• Added support to save/restore model file with presets/state by @sinai007 in #11
• Create LICENCE.md by @mikeoliphant in #13
• Update build.yml by @mikeoliphant in #16
• Output fixes. patch get/state changed by @rerdavies in #17
• Add file handling for Mod by @micahvdm in #20
• Fix install target by @runiq in #30
• Rework worker and cleanup by @falkTX in #34

New Contributors

• @mikeoliphant made their first contribution in #2
• @falkTX made their first contribution in #7
• @sinai007 made their first contribution in #11
• @rerdavies made their first contribution in #17
• @micahvdm made their first contribution in #20
• @runiq made their first contribution in #30

Full Changelog: https://github.com/mikeoliphant/neural-amp-modeler-lv2/commits/0.1.0