yamadapc / augmented-audio Goto Github PK

Implement looper with 3 tracks that works a bit smoother than existing options.

Features:

• 3 stereo tracks
• Relatively long max recording time
• Still try to integrate into ableton as much as possible: first loop can set whole transport tempo and so on
• Quantize & allow both start/end after first beat, fixing the performance issues by always recording & fading from pre-button press material
• Avoid pops with the same approach above

Implement offline render functionality into the test host.

• 200MB memory usage at the moment for the GUI
• Limit the nº of tokio workers & threads in general
• Look into iced

• It'd be great to have a metric for coverage - test coverage

Similar to dsp graph, but using audio-processor-traits

Buttons for

• Start
• Stop
• Pause

Add a MIDI trait

This is to be fixed in plugin-host-gui2.

This is useful for:

• Syncing UI for DAW parameter changes
• Running certain pre-processing on parameter-set (rather than in the process callback)

• Better documentation
• Better examples

• Port webview plugin-host part to use iced
• Write a document explaining why it doesn't make sense to use tauri in this type of application

Support rendering plugin output into a file.

• Attach delegate to window to update state when it's closed
• Allow the window to be re-opened/closed
• Change the style-mask so it has a close button

This is only on the CLI.

Try making the build fail if clippy warns.

Centralise at least style state in side of a single shared context reference.

Explore centralising all UI state behind a context struct which provides dynamic binding into it.

• Audio processor options are built based on CLI input & the default audio thread options
• If the user uses a non-default host/device, the stream options will be different (for example it's multi-channel)
• In that case, we should respect the stream options on the audio processor options
• That might be configurable (output to channels X-Y)

• Set-up blog/documentation web-site with gatsby

Currently darwin-webkit message-passing is broken and a fixed version is in this repository ; we should fix it there.

It'll create threads & leak them at the moment. Subscriptions should be cancelled.

Add a way to control volume on the host GUI

• Stereo
• Ping-pong
• Tempo sync with DAW host
• Multiple taps?
• Built-in modulation?

When re-opening the plugin-host GUI I want to see the same thing I had before.

plugin-host will leave copies of dylibs it loads to work-around reloading of dylibs in macOS.

• Group all solutions under one namespace

https://github.com/RustAudio/wmidi

Currently the web plug-in will work work against the react dev server. The HTML assets should be bundled inside the plug-in along with an HTTP server.

Based on example, relay MIDI messages into plug-ins

& optimise it when buffer types are matching

I believe there's a memory leak around windows not being cleaned-up

It might be that WKScriptMessageHandler is actually slower.

Idea on how to benchmark:

• Send a message to increment a counter into JavaScript, send the response back
• Send the next increment ; do this until a count is reached
• Measure average round-trip time

Implement framework to do MIDI mapping from react.

https://github.com/yamadapc/augmented-audio/runs/3156489201?check_suite_focus=true

For webview:

• Either pick or write a webview implementation

For bundling:

• Figure out bundling for diff. platforms

For CI:

• Enable building on multiple platforms