Hi. I just wanted to ask if there is a planned support for v4 of AudioKit in Xcode 12? Or if there is a documentation for the v5 beta version which works in Xcode 12.

I cannot compile my app in Xcode 12 and hence cannot prepare it for the iOS 14 which is now already released. I have went into detail in this stackoverflow post. https://stackoverflow.com/questions/63860545/implementing-microphone-analysis-with-audiokit-v5

Thank you.

We have full MIDI support, but no OSC support yet. There seems to be a few options listed at Charles Martin's Blog:

http://charlesmartin.com.au/blog/2013/3/26/finding-a-good-osc-library-for-ios

but so far it looks like his Metatone might be best

https://github.com/cpmpercussion/MetatoneOSC

and it comes with this example

https://github.com/cpmpercussion/ExampleOSC

If you are on OSC wiz and want to knock this out, that would be great, or else you can also just comment that you would also like to have this feature, and that will motivate us to prioritize this higher.

AudioKit and Ableton Link

Just writing down a quick list of methods that get called by Ableton Links 'LinkHut' example app to execute a tempo update. Since Ableton Link is meant to keep different music apps in tempo / sync then updating the AKSequencer / tempo backend would probably be enough to integrate Link with AudioKit.

Integration

Note:

• Need to download Ableton Link / LinkHut from the [LinkKit release tab] (https://github.com/Ableton/LinkKit/releases)
• LinkHut is in the examples folder inside the release zip

Here are the steps taken in LinkHut from Ableton:

Below are brief breakdown of steps taken to setup Ableton Link with an audio engine to update tempo

All steps described below are based of the Ableton Link LinkHut example app.

The process of updating the tempo (when another user set's new tempo):

• First audio engine and Ableton Link objects are setup in ViewController -> viewDidLoad

  • _audioEngine = [[AudioEngine alloc] initWithTempo:_bpm];
  • This is where AVAudioSession shared instance and callbacks get setup
  • ABLLinkSetSessionTempoCallback(...)
    • This is where Ableton Link tempo change gets passed along to user callback to update UI
  • _audioEngine.linkRef, onSessionTempoChanged, (__bridge void *)self);
  • _linkSettings = [ABLLinkSettingsViewController instance:_audioEngine.linkRef];
  • [_audioEngine setQuantum:_quanta];

• Ableton Link will receive the tempo change from the other user (over the network for ex)

• Ableton Link will send an event to the AVAudioSession audio engine route

• User code has a callback that gets called when route event happens

• See 'AudioEngine.m' in the Ableton LinkHut app example:

  • callback in AudioEngine.m -> audioCallback
  • This is setup in 'setupAudioEngine' method that gets called during init:
  • This is where _linkData object is setup to connect to shared instance of AVAudioSession properties
  • 'AURenderCallbackStruct ioRemoteInput;'
    • gets called to setup i/o callback using 'audioCallback' function
    • audioCallback is defined to checks for a tempo update and saves the timestamp and new tempo
    • See 'static OSStatus audioCallback' method in AudioEngine.m
    • This is where the tempo actually gets updated when we receive an tempo event from Ableton Link
    • If new tempo is valid it passes along the new tempo event to another callback that informs the UI to display the new tempo to the user
    • This happens in 'ABLLinkSetSessionTempoCallback' defined in the ViewController.swift
    • This callback is setup in ViewController.swift -> viewDidLoad -> ABLLinkSetSessionTempoCallback

Something strange has started happening with the Sequencer. I'm not sure exactly when, but recently.

The "stop playing notes" functionality has essentially stopped working. I have confirmed that my code is calling the stop method on each track, but the notes continue sustaining, which is a real problem in my app.

Here's what the code looks like that I execute in direct response to the user pressing the "stop" button:

    public func stop() {
        if let sequencer = sequencer {
            internalPlaybackState = .stopped
            sequencer.stop()
            for track in sequencer.tracks {
                track.stopPlayingNotes()
            }
            sequencer.rewind()
        }
    }

In that code, "sequencer" is just an AudioKit Sequencer. The "internalPlaybackState" is something related to my app that should have no bearing here.

I'd be happy to help you reproduce it. All the source code that I'm using for playback is in the SongSprout Swift Package located here: https://github.com/dunesailer/SongSprout

The main playback class is Orchestrion.

I am trying to use AKMicrophone in objective C sample app like this and it crashes for all the pod versions above 4.5.1 on the init step. iOS version is 12.1.2, iphone 7plus, Xcode version 10.1. It works fine in swift sample app MicrophoneAnalysis. So I suspect maybe missing @objc somewhere? @property (strong, nonatomic) AKMicrophone *mic; Then in ViewDidLoad: _mic = [[AKMicrophone alloc] init]; // Crashes here.
I am also getting following error in the console: 2019-01-15 12:56:30.809227-0500 VoiceWavePoc[2856:1473174] [avas] AVAudioSessionPortImpl.mm:56:ValidateRequiredFields: Unknown selected data source for Port Receiver (type: Receiver)

While connecting an external I/O device , It is batter has a setting for user can select which channels are prefer . ig. User apogee quartet , quartet has 4 analog input , and 6 analog output , you can select specific I/O channels for recording and monitor . you can select input 1 & 2 for stereo recording or just 1 for mono recording.

For current version I test with quartet , if will using default input 1 , still works , but if your audio is via other channels , nothing can be recorded

A setting view with real time chart UI will be great for measuring the input/output volume

AudioKit version 4.5.5: crash in code AKSettings.audioInputEnabled = true mic = AKMicrophone() tracker = AKFrequencyTracker(mic) silence = AKBooster(tracker, gain: 0)

_AVAE_Check: required condition is false: [AVAudioIONodeImpl.mm:911:SetOutputFormat: (format.sampleRate == hwFormat.sampleRate)] *** Terminating app due to uncaught exception 'com.apple.coreaudio.avfaudio', reason: 'required condition is false: format.sampleRate == hwFormat.sampleRate'

Test device - iPhone 8 and simulators AudioKit.engine.inputNode.inputFormat(forBus: 0).sampleRate periodically returns in the application 44100.0 or 48000.0

On the iPhone se AKMicrophone() works stably, no crash

I utilized AKOfflineRenderNode to process recorded voice. The idea is to record voice and save it into the file. Next apply some effects to this file and save combined voice+effects into file.

I have successful result on my iPhone 6s running on iOS 11.0.3 but on other devices running on iOS 11+(for example last generation of iPad on 11.0.3) I always have "silent" file with same size about 60 Kb. I attached one for reference(rec.zip). Also the issue 100% reproducible on iOS 11 simulator.

Here is my setup(I removed effects setup since offline render doesn't work even on clean recording):

    fileprivate func setupAudioKit(){
        AKSettings.enableLogging = true
        AKAudioFile.cleanTempDirectory()
        
        AKSettings.bufferLength = .medium
        AKSettings.numberOfChannels = 2
        AKSettings.sampleRate = 44100
        AKSettings.defaultToSpeaker = true
        
        do {
            try AKSettings.setSession(category: .playAndRecord, with: .allowBluetoothA2DP)
        } catch {
            AKLog("Could not set session category")
        }
        
        mic = AKMicrophone()
        micMixer = AKMixer(mic)
        
        recorder = try? AKNodeRecorder(node: micMixer)
        
        if let file = recorder.audioFile {
            player = try? AKAudioPlayer(file: file)
            player.looping = true
        }
        playerMixer = AKMixer(player)
// Effects setup
        offlineRenderer = AKOfflineRenderNode(playerMixer)
        AudioKit.output = offlineRenderer

        AudioKit.start()
}

Here is export method:

     fileprivate func render() {
        offlineRenderer.internalRenderEnabled = false
        player.schedule(from: 0, to: player.duration, avTime: nil)
        
        let renderURL = URL(fileURLWithPath: FileHelper.documentsDirectory() + "rec.m4a")
        
        let sampleTimeZero = AVAudioTime(sampleTime: 0, atRate: AudioKit.format.sampleRate)
        player.play(at: sampleTimeZero)
        do {
            try offlineRenderer.renderToURL(renderURL, seconds: player.duration)
        } catch {
            print(error)
        }
        player.stop()
        offlineRenderer.internalRenderEnabled = true
    }

I'm sure you're all working on this, just making the issue to add some visibility for others who are wondering as well.

Xcode 10.2 with Swift 5 support was just released today. Hoping Swift 5 support is coming soon to AudioKit!

I'm integrating Audiobus into my AudioKit-based app, and have followed the instructions here, but when I call Audiobus.start(), the app will continue running for about a second, then I get some error output, followed by a crash:

2017-03-22 23:25:45.620918 Vulse[20369:7115604] [central] 54: ERROR: [0x16e167000] >avae> AVAudioIONodeImpl.mm:365: _GetHWFormat: required condition is false: hwFormat framerate: 23 2017-03-22 23:26:03.917219 Vulse[20369:7115604] *** Terminating app due to uncaught exception 'com.apple.coreaudio.avfaudio', reason: 'required condition is false: hwFormat'

So, I got the FilterEffects example up and running to see if the issue was just in my app. (First I had to change the input variable in viewDidLoad to a property, because it was deiniting too early and causing a crash.) But after fixing that, I unfortunately got the same error above.

Interestingly, for the second or two before the app crashes, everything seems to work properly: the app takes audio input from audiobus and I can hear it output with the reverb applied. Also (at least in my app) if I remove the Audiobus.start(), but still try to instantiate an AKStereoInput, I get a similar message: SetOutputFormat: required condition is false: format.sampleRate == hwFormat.sampleRate

Do you get the same crash when trying this yourself, with the latest AudioKit release? I'm using Audiobus 2.3.3, on an iPhone 6s running iOS 10.2.1.

AudioKit.AKFaderAudioUnit input format must match output format

2020-06-02 12:15:08.494929+0530 [avae] AVAEInternal.h:109 [AVAudioEngineGraph.mm:4170:UpdateGraphAfterReconfig: (AUGraphParser::InitializeActiveNodesInOutputChain(ThisGraph, kOutputChainFullTraversal, *conn.srcNode, isChainActive)): error -10868 2020-06-02 12:15:08.497253+0530 *** Terminating app due to uncaught exception 'com.apple.coreaudio.avfaudio', reason: 'error -10868'

Fixes #2814, but it feels clumsy as we have two syncronisation mechanisms - locks and queue.

We should probably rewrite a tap to work on one dispatch queue and remove all the locks.

macOS Version(s) Used to Build

macOS 13 Ventura

Xcode Version(s)

Xcode 14

Description

To reproduce, run tests with TSAN enabled

@josi

Crash Logs, Screenshots or Other Attachments (if applicable)

No response

• This currently fails so it is skipped
• Add a reference to bug report

I don't have clear idea in mind on how to fix this. I was thinking about possibly maintaining separate state within AudioEngine, but we don't have a reference to it from Mixer. Adding associated object to AVAudioEngine would probably work, but I am not sure if it is worth it. It is not that difficult to workaround at the callsite.

Description

AUAudioUnitBus has isEnabled property. If this property is disabled, pullBlock will not be called on an input bus and will save processing time.

This is working very well in Apple's Audio Units (e.g. AVAudioMixerNode) and depending where this bus in the chain is, you might experience significant performance improvements as none of the upstream nodes will be called.

Unfortunately this is not implemented in AudioKit units (e.g. DryWetMixer).

Proposed Solution

This check would need to happen here and here I believe.

I am not sure how to approach this since this property is in AUAudioUnitBusArray->AUAudioUnitBus which is not available within DSP.

Describe Alternatives You've Considered

I guess one approach would be to duplicate these properties on DSP, but that seems not very elegant.

Additional Context

Please note that this is different to bypass effect, which just bypasses processing in one node, but doesn't prevent further signal propagation.

Description

AVAudioEngine is:

• relatively buggy (still not handling dynamic connections well)
• opaque (we don't have access to its code)
• annoying (mapping AK's connections onto AVAudioEngine is tricky)
• hard to use (terrible error messages)

Proposed Solution

Mostly replace AVAudioEngine usage in AK with our own code. AVAudioEngine would still be used to handle input and output, but we would evaluate our audio graph within a single AU. Thus the AVAudioEngine graph would always be Input -> AK AU -> Output.

• Instead of the recursive pull-based style of typical AUs, sort the graph and call the AUs in sequence with dummy input blocks (this makes for simpler stack traces and easier profiling).
• Write everything in Swift (AK needs to run in Playgrounds. Make use of @noAllocations and @noLocks)
• Host AUs (actually the same AUs as before!)
• Don't bother trying to share buffers and update in-place (this seems to be of marginal benefit on modern hardware)
• Preserve almost exactly the same API

Describe Alternatives You've Considered

Continue to file bugs with Apple and hope they fix them. This is the status quo. It's not good enough.

Additional Context

The devil's in the details, and this could prove to be too much work after we investigate further.

Description

Re discussion in https://github.com/AudioKit/AudioKit/issues/298

From Logic, if you save your EXS24 instrument into the Samples/ folder in your Xcode project and also pick "include audio data" — the saved EXS24 instrument will contain the correct relative path to be able to load!

Proposed Solution

Since this is something a few people have run into, it might be worth adding something in the docs or FAQ about it?

Describe Alternatives You've Considered

This seems like the best thing

Additional Context

Happy to write up the FAQ post if you like