spotify / basic-pitch-ts Goto Github PK

As a caption

I think it would better to have some show case / references implementation about how to use this module .

for example, if would be great if the official one can be open sourced

Hi is there any documentation to start contributing to this interesting project ?
My goal is to try adapt it to real time audio2midi.
I would like to understand better the application flow with some description of the functions and input and output parameters without interpreting all the code.
For example :

1. How do the OUTPUT_TO_TENSOR_NAME.frames, OUTPUT_TO_TENSOR_NAME.onsets, OUTPUT_TO_TENSOR_NAME.contours are related to note on , note off , pitchbend messages ?
2. What is melodiatrick ?
3. What is energy ?
4. What outputToNotesPoly function does ?

Do you have some performance numbers for the model.execute call in ms (single batch) ?
Thx a lot

This library is amazing! However, when it's processing it locks up the UI so animations stop and buttons are not clickable.

I was wondering if I could put this in a service worker and am not sure if that has been done before.

The first problem is that a service worker can't import files and I can't figure out how to have my service worker use files in node_modules. I'm guessing I'll need a build step.

I was wondering if there were already a build option in this library for being able to load it the using a single <script> tag, because I think I'd be able to use that in my service worker. I didn't see that.

So I think I might need to write that, which I'll try to do but I'm not sure if I can.

Also I'm not sure if that will even work since there are two other files that get loaded in (the bin and the json files) and I'm not sure if that will work inside the service worker.

Any pointers you can give while I flail away at this would be appreciated.

Thanks!

In what units are the values given in the pitchBend array on a note event?

In my experiments, I'm seeing values like 0, -1, 1, 2, 4, and 5. I see from the other basic pitch repo an issue that seems to suggest these values are semitones, but numbers like 4 and 5 make that seem odd. Further, I'm not seeing any non-integer values, meaning that if we're in semitones, the resolution of this library doesn't consider cents, which also seems odd.

Hoping for some clarification, thank you!

Hi everyone 👋
Documentation say: "While you may use stereo audio as an input to our model, at prediction time, the channels of the input will be down-mixed to mono, and then analyzed and transcribed."

I'm trying to get stereo audio notes using basic-pitch and I got this error:
Error: Input audio buffer is not mono! Number of channels is 2. Should be 1

Why this happens?

I'm trying to use Basic Pitch in my Next.js application and struggling to read the model.json file required to initialize the class. Why is the model not baked into the initializer? Why is the model not importable from the Basic Pitch module? Maybe I'm missing something obvious.

Thanks!

Hey all, amazing library. As I was setting it up I ran into:

It looks like the onsets, frames, and contours, get defined. Pct is 1 as well after await basicPitch.evaluateModel. The problem i'm running into is after the model runs noteFramesToTime(addPitchBendsToNoteEvents())) keep throwing errors. I was wondering if ya'll had run into this by any chance? It seems to be happening around toMidi && globalMax.

I'm using NextJs and typescript as well. I verified that my audioRecording was a mono recording and set to the correct sampleRate.