damrsn / neuralnote Goto Github PK

say two singers are singing, id like if the midi files werent stuck together (especially cause rhey could be sjngjng jn tje same range / on the same note)

Nothing wrong found so far, just wanted to say thank you for this amazing piece of software!

Thank you for making NeuralNote

It would be really great to have access to polyphonic pitch bend expressions on the transcribed notes.

CLAP can do it with CLAP_NOTE_EXPRESSION_TUNING

Edit: This would require live output, so exporting MIDI with MPE note expressions would be needed instead

(this is just a discussion issue)

I'm trying to try NeuralNote on Linux and wasn't successful when trying to build it (seems to require some dependencies to become Linux-compatible) nor when using yabridge+Wine, error is

22:30:29 [NeuralNote-ACDFPWZr] Initializing yabridge version 5.0.5
22:30:29 [NeuralNote-ACDFPWZr] library:       '/lib64/libyabridge-vst3.so'
22:30:29 [NeuralNote-ACDFPWZr] host:          '/usr/bin/yabridge-host.exe'
22:30:29 [NeuralNote-ACDFPWZr] plugin:        '/home/mistria/.wine/drive_c/Program Files/Common Files/VST3/NeuralNote.vst3'
22:30:29 [NeuralNote-ACDFPWZr] plugin type:   'VST3'
22:30:29 [NeuralNote-ACDFPWZr] realtime:      'no'
22:30:29 [NeuralNote-ACDFPWZr] memlock limit: '8388608 bytes, see below'
22:30:29 [NeuralNote-ACDFPWZr] 
22:30:29 [NeuralNote-ACDFPWZr]    With a low memory locking limit, yabridge may not be
22:30:29 [NeuralNote-ACDFPWZr]    be able to lock its shared memory audio buffers into
22:30:29 [NeuralNote-ACDFPWZr]    main memory. Performance may be degraded until you fix 
22:30:29 [NeuralNote-ACDFPWZr]    this. Check the readme for instructions on how to do that.
22:30:29 [NeuralNote-ACDFPWZr] 
22:30:29 [NeuralNote-ACDFPWZr] sockets:       '/run/user/1000/yabridge-NeuralNote-ACDFPWZr'
22:30:29 [NeuralNote-ACDFPWZr] wine prefix:   '/home/mistria/.wine'
22:30:29 [NeuralNote-ACDFPWZr] wine version:  '8.11 (Staging)'
22:30:29 [NeuralNote-ACDFPWZr] 
22:30:29 [NeuralNote-ACDFPWZr] config from:   '<defaults>'
22:30:29 [NeuralNote-ACDFPWZr] hosting mode:  'individually, 64-bit'
22:30:29 [NeuralNote-ACDFPWZr] other options: '<none>'
22:30:29 [NeuralNote-ACDFPWZr] 
22:30:29 [NeuralNote-ACDFPWZr] Enabled features:
22:30:29 [NeuralNote-ACDFPWZr] - bitbridge support
22:30:29 [NeuralNote-ACDFPWZr] - CLAP support
22:30:29 [NeuralNote-ACDFPWZr] - VST3 support
22:30:29 [NeuralNote-ACDFPWZr] 
22:30:29 [NeuralNote-ACDFPWZr] [Wine STDERR] Initializing yabridge host version 5.0.5
22:30:29 [NeuralNote-ACDFPWZr] [Wine STDERR] Preparing to load VST3 plugin at '/home/mistria/.wine/drive_c/Program Files/Common Files/VST3/NeuralNote.vst3'
22:30:30 [NeuralNote-ACDFPWZr] [Wine STDERR] Finished initializing '/home/mistria/.wine/drive_c/Program Files/Common Files/VST3/NeuralNote.vst3'
22:30:30 [NeuralNote-ACDFPWZr] [Wine STDERR] 0228:fixme:ntdll:EtwEventSetInformation (deadbeef, 2, 0000000180808179, 46) stub
22:30:30 [NeuralNote-ACDFPWZr] [Wine STDERR] 0238:fixme:kernelbase:AppPolicyGetThreadInitializationType FFFFFFFFFFFFFFFA, 0000000001AAFE10

Can you please provide hints about the best direction to make progress with using NeuralNote on Linux?

Duh.

right now I need to separately use a KeyFinder or experiment a bit to know which key my sample audio is in and restrict the midi output with this info. it would be very convenient if this could be done in the plugin with the press of a button.

Really keen to try it out.

For some reason I am unable to have FlStudio 21 (latest release) recognise the plugin.

I exported the vst to a directory and setup FlStudio to be aware of it.

Here is the vst in the directory...

Here is the proof the setup is correct and its unable to find it after a scan...

I've done vst adding many times before in FlStudio, so the process is well understood by me.

Any suggestions please?

dont have neuralnote be fixed to one size. weirdly enough updating the window size is the only way to refresh what the gui is displaying on linux with wine

Thanks @DamRsn! Really appreciate this plugin and already used it a couple of times in my DAW ;)

Hi
I can't open Neural Note standalone app on a MacBook Pro (OS Ventura 13.6). I don't under stand why
best regards
Christophe

or allow the user to input hexs to decide the colors of components.

the rainbow hurts my eyes man

I built those output, any ideas how to use it in Xcode? Thanks indeed~

Hello,
I was looking for a plugin that would be able to convert my acoustic recordings to MIDI without the extra trouble of having to replay it on a digital instrument just to get the MIDI file I need. So far Neuralnote works great, it’s incredibly accurate and fast once it’s picked up on the source file. There is a big issue though, I’m a blind user on windows with the NVDA screen reader and Reaper as my DAW. When I insert Neuralnote, only the record button is labeled and then the playback button after it has picked up on the audio I play from a different track. There seems to be no way for me to insert audio without recording playback which can be time consuming, and once that audio is recorded I don’t have access to the three knobs: “note sensability”, “split”, “min note duration”. I needed sighted help to adjust those and there doesn’t seem to be a way to save the MIDI or export it straight from the plugin window once I’m satisfied with it. I know that normally you can drag and drop items in and out of the FX window, but that is not possible for me with a screen reader and strangely the tracks seem to disappear once NeuralNote is added, although the latter may be a Reaper exclusive issue.

I'm working on a port of NeuralNote to visionOS and hitting some trouble getting onnxruntime to compile for visionOS.

Comment here says that Features is doing a CQT (what does "harmonically stack" mean?). I was wondering if other CQT code could be used instead of the onnx model? For example: https://github.com/cannam/constant-q-cpp

On Windows, I get an error when linking with provided onnxruntime.lib:

'onnxruntime.lib' was created by a different version of the compiler than other objects like NeuralNote_Standalone ...

My MSVC version is 19.29.30148.0

theres plenty of foss tuners. it would be useful if one was built-in so people doing vocals can see how off they are in real-time and correct themselves

I read in the doc
"""
Very few DAWs support audio input/MIDI output plugins as far as I know. This is partially why NeuralNote is an Audio FX plugin (audio-to-audio) and that MIDI is exported via drag and drop.
"""
I use a DAW that seems to allow it (Carla), and I also tried Widi VST which does allow to create a new midi source that Carla detect (as a separate input, detached from the plugin, but still present and still allowing to directly map the midi transcription).
If this pattern of creating a new virtual midi input for the plugin output can work, I would recommend to implement it.

was wondering if you knew how / wanted to set up this midi file github page resource
Fannon/midi-chords-and-scales-pack#2

heres a list of free midi resources, i want there to be a singular github for everyone to access them

currently doesnt detect the volume of a note and write that info to the midi

First of all thank You for this app. It does a very good job of audio 2 midi. I use it on guitar tracks with good results in Cubase when I connect a guitar instrument (Ample guitar).

But I would like to get a human playable transcription out of it, too. So how can I separate the MIDI into different tracks for various instruments?

I use stemming to get voice and guitar separated.
But in recordings there are at least two guitars (lead and rythm) to separate.

How can I go on with NeuarlNotes results?

Hi! Awesome work. If youre taking any feature requests:
-being able to trigger external hardware in real time by looping would be amazing.
-Also having a rearrange/shuffe option for the output midi (like the idea from Harvest Plugins Harvest)
Mostly to trigger a drum machine in real time.

It's a project to improve the design of open source apps. So it provides resources, connects designers (like OP) with engineers/projects, spreads awareness etc.

https://opensourcedesign.net/jobs/ apply the project here!!

Thanks for porting this. Much appreciated.

Is there any way to get around the 3 minute maximum for inputs?

Can there be too many extra notes improved?

if so, add to f-droid

I would really like to use this VST but am unable to locate it. Does your team have any recommendations on where to put the install file?

Thank you for this amazing application!

When using logic, I could not drag and drop files into the plugin (.component version) , which I believe is the main point of it being a "plugin" as opposed to an individual standalone application. I saw this functionality in other DAWs in youtube videos, so I was hoping it could also work in logic, one of the main DAWs used on Macs. Sorry if I'm missing something!

Appreciate all of your hard work

I'm really eager and excited to try this tool but I'm using windows and pro tools, I do not understand how to get this working. I put the VST3 file into its appropriate folder but this is not recognized in my DAW. Someone please helppp!!!

Is it possible to get a web/pip install version of this so it's not constrained to a DAW?

Support dropping mp3 files

https://www.youtube.com/watch?v=ajN_fWja_84&list=PLBbMkfXvxdzpj_tizshI6z07nPgqeHZLp&index=1

In the UI, the NeuralNote text in the background is a bit pixelated. It's less sharp than the text in the Record button

Hi,
I've just installed this, thought i'd be able to spit out a MIDI file from the standalone app, now trying and failing to see how to connect to DAW... BTW my file is drums only... could this be why nothing is showing up?
Are there any instructions available?
Tx
Mark

Need help on how to install on windows.
kind of confusing with codes and etc

Either implement a simple synth inside NeuralNote, or find an elegant way to route the transcribed MIDI to another synth.

Add Play and Stop buttons, a clickable playhead and a mixer for levels of the source audio and the synthesized transcription.

This would help to adjust the parameters to get the best transcription possible.

i tried to upload files as .wav and .mp3 both in standalone and daw shows this, please help~~

Add Zoom functionality so that the audio and the transcription is easier to navigate.

A command like Ctrl + Scroll seems appropriate.

First of all, thanks for this cool project!
In Could NeuralNote transcribe audio in real-time?, you mention that:

The CQT requires really long audio chunks (> 1s) to get amplitudes for the lowest frequency bins.

In my experience, the lowest piano key (27.5Hz) requires about 0.62s of samples to be detected. Still not quite real-time, but technically it is possible to detect correlation even earlier, when using a streaming variant of the transform.

Out of curiosity, how difficult do you think it would be to use such streaming transform with your neural network?