NeuralNote comes as a simple AudioFX plugin (VST3/AU/Standalone app) to be applied on the track to transcribe.
The workflow is very simple:
Gather some audio
Click record. Works when recording for real or when playing the track in a DAW.
Or drop an audio file on the plugin. (.wav, .aiff, .flac, .mp3 and .ogg (vorbis) supported)
The MIDI transcription instantly appears in the piano roll section.
Listen to the result by clicking the play button.
Play with the different settings to adjust the transcription, even while listening to it
Individually adjust the level of the source audio and of the synthesized transcription
Once you're satisfied, export the MIDI transcription with a simple drag and drop from the plugin to a MIDI track.
NeuralNote uses internally the model from Spotify’s basic-pitch. See their blogpost and paper for more information. In NeuralNote, basic-pitch is run using RTNeural for the CNN part and ONNXRuntime for the feature part (Constant-Q transform calculation + Harmonic Stacking). As part of this project, we contributed to RTNeural to add 2D convolution support.
Watch the presentation video for the Neural Audio Plugin competition here.
Available for Windows, Linux and Mac as a VST3/AU/Standalone app and the VST3 plugin is compatible with Ardour!
Hello,
I installed the standalone version on linux.
Then executed the program.
I can convert to midi, but when i drag the midi transcription onto a folder, nothing happens.
When i drag it the mouse looks like an empty folder.
What is it that i am doing wrong?
You need to drag and drop to a DAW’s Midi/Instrument track, You can see how to use it in this video https://www.youtube.com/watch?v=6_MC0_aG_DQ for me, I prefer to use the VST3 because I use Ardour.
Ok, i can do that, i have Ardour and the plugin, and it works, but what is the use of the stand alone application if it is not possible to get the midi file that is generated?
Have a nice day.
Jean
The script that comes just does polyphonic transcription
I have a modified version of said script that allows for monophonic and polyphonic transcription, I added monophonic transcription because I wanted to do voice to midi with some more accuracy.
Great work for the monophonic audio to midi LUA script! I had been looking for that for a long time, I think this script is a better option because it integrates well with Ardour!
This looks nifty (as does NeuralNote, of course), thanks for sharing
Am I right in thinking you could use your script, for example, to output a midi file from a vox performance → tidy it up as needed → then use that midi file as the basis for pitch correction using a plugin like GSnap or x42 Auto Tune…?
Just the ticket Thank you jmantra. Unfa did some great work demonstrating Ardour’s functionality and stuff like that, it’s a shame he’s not doing it anymore.
Anyway thanks again for the scripts and info, I look forward to trying it out.
Fascinating stuff especially for a ‘Midiot’ like myself. I’m wondering what the fine-tuning differences are between the 2 methods? I see the lua script doesn’t estimate velocity, does NeuralNote? I see NeuralNote also has split and quantize functions etc… are these automated in the lua script or expected to be carried out in the rendered file?
Could anyone from the Win/Mac side comment on how this would compare to something like the Jam Origin Plugin?
Could this be used to help tune the vocal track? First convert vocal to midi and correct it manually and together with X42 Auto tune tune the vocal like this?
JamOrigin’s MidiGuitar is more like a live performance plugin, in the other hand, NeuralNote is more focused in convert audio(mono or stereo) into a midi file, like Celemony Melodyne but without an audio tuning effect, that’s why it has many midi adjustments so you can tweak the midi file before export or drag to the track, although it also has a midi output port you can route to a midi track too.
