A simple, intuitive audio framework for iOS and OSX.
EZAudio has recently been deprecated in favor of AudioKit. However, since some people are still forking and using EZAudio I've decided to restore the README as it was. Check out the note below.
I'd really like to start creating a list of projects made using EZAudio. If you've used EZAudio to make something cool, whether it's an app or open source visualization or whatever, please email me at syedhali07[at]gmail.com and I'll add it to our wall of fame! To start it off:
- Detour - Gorgeous location-aware audio walks
- Jumpshare - Incredibly fast, real-time file sharing
- Piano Tuner (Home Edition)
- Piano Tuner (Pro Edition)
- Music Pitch Detector
- Piano Prober
- Multi Tuner
Awesome Components
I've designed six audio components and two interface components to allow you to immediately get your hands dirty recording, playing, and visualizing audio data. These components simply plug into each other and build on top of the high-performance, low-latency AudioUnits API and give you an easy to use API written in Objective-C instead of pure C.
A useful class for getting all the current and available inputs/output on any Apple device. The EZMicrophone
and EZOutput
use this to direct sound in/out from different hardware components.
A microphone class that provides its delegate audio data from the default device microphone with one line of code.
An output class that will playback any audio it is provided by its datasource.
An audio file class that reads/seeks through audio files and provides useful delegate callbacks.
A replacement for AVAudioPlayer
that combines an EZAudioFile
and a EZOutput
to perform robust playback of any file on any piece of hardware.
A recorder class that provides a quick and easy way to write audio files from any datasource.
A Core Graphics-based audio waveform plot capable of visualizing any float array as a buffer or rolling plot.
An OpenGL-based, GPU-accelerated audio waveform plot capable of visualizing any float array as a buffer or rolling plot.
Cross Platform
EZAudio
was designed to work transparently across all iOS and OSX devices. This means one universal API whether you're building for Mac or iOS. For instance, under the hood an EZAudioPlot
knows that it will subclass a UIView for iOS or an NSView for OSX and the EZMicrophone
knows to build on top of the RemoteIO AudioUnit for iOS, but defaults to the system defaults for input and output for OSX.
Within this repo you'll find the examples for iOS and OSX to get you up to speed using each component and plugging them into each other. With just a few lines of code you'll be recording from the microphone, generating audio waveforms, and playing audio files like a boss. See the full Getting Started guide for an interactive look into each of components.
EZAudioCoreGraphicsWaveformExample
Shows how to use the EZMicrophone
and EZAudioPlot
to visualize the audio data from the microphone in real-time. The waveform can be displayed as a buffer or a rolling waveform plot (traditional waveform look).
EZAudioOpenGLWaveformExample
Shows how to use the EZMicrophone
and EZAudioPlotGL
to visualize the audio data from the microphone in real-time. The drawing is using OpenGL so the performance much better for plots needing a lot of points.
EZAudioPlayFileExample
Shows how to use the EZAudioPlayer
and EZAudioPlotGL
to playback, pause, and seek through an audio file while displaying its waveform as a buffer or a rolling waveform plot.
EZAudioRecordWaveformExample
Shows how to use the EZMicrophone
, EZRecorder
, and EZAudioPlotGL
to record the audio from the microphone input to a file while displaying the audio waveform of the incoming data. You can then playback the newly recorded audio file using AVFoundation and keep adding more audio data to the tail of the file.
EZAudioWaveformFromFileExample
Shows how to use the EZAudioFile
and EZAudioPlot
to animate in an audio waveform for an entire audio file.
EZAudioPassThroughExample
Shows how to use the EZMicrophone
, EZOutput
, and the EZAudioPlotGL
to pass the microphone input to the output for playback while displaying the audio waveform (as a buffer or rolling plot) in real-time.
EZAudioFFTExample
Shows how to calculate the real-time FFT of the audio data coming from the EZMicrophone
and the Accelerate framework. The audio data is plotted using two EZAudioPlots
for the time and frequency displays.
The official documentation for EZAudio can be found here: http://cocoadocs.org/docsets/EZAudio/1.1.4/
You can also generate the docset yourself using appledocs by running the appledocs on the EZAudio source folder.
To begin using EZAudio
you must first make sure you have the proper build requirements and frameworks. Below you'll find explanations of each component and code snippets to show how to use each to perform common tasks like getting microphone data, updating audio waveform plots, reading/seeking through audio files, and performing playback.
iOS
- 6.0+
OSX
- 10.8+
iOS
- Accelerate
- AudioToolbox
- AVFoundation
- GLKit
OSX
- Accelerate
- AudioToolbox
- AudioUnit
- CoreAudio
- QuartzCore
- OpenGL
- GLKit
You can add EZAudio to your project in a few ways:
1.) The easiest way to use EZAudio is via <a href="http://cocoapods.org/", target="_blank">Cocoapods. Simply add EZAudio to your <a href="http://guides.cocoapods.org/using/the-podfile.html", target="_blank">Podfile like so:
pod 'EZAudio', '~> 1.1.4'
If you're also using the Amazing Audio Engine then use the EZAudio/Core
subspec like so:
pod 'EZAudio/Core', '~> 1.1.4'
2.) EZAudio now supports Carthage (thanks Andrew and Tommaso!). You can refer to Carthage's installation for a how-to guide: https://github.com/Carthage/Carthage
3.) Alternatively, you can check out the iOS/Mac examples for how to setup a project using the EZAudio project as an embedded project and utilizing the frameworks. Be sure to set your header search path to the folder containing the EZAudio source.
EZAudio
currently offers six audio components that encompass a wide range of functionality. In addition to the functional aspects of these components such as pulling audio data, reading/writing from files, and performing playback they also take special care to hook into the interface components to allow developers to display visual feedback (see the Interface Components below).
Provides a simple interface for obtaining the current and all available inputs and output for any Apple device. For instance, the iPhone 6 has three microphones available for input, while on OSX you can choose the Built-In Microphone or any available HAL device on your system. Similarly, for iOS you can choose from a pair of headphones connected or speaker, while on OSX you can choose from the Built-In Output, any available HAL device, or Airplay.
To get all the available input devices use the inputDevices
class method:
NSArray *inputDevices = [EZAudioDevice inputDevices];
or to just get the currently selected input device use the currentInputDevice
method:
// On iOS this will default to the headset device or bottom microphone, while on OSX this will
// be your selected inpupt device from the Sound preferences
EZAudioDevice *currentInputDevice = [EZAudioDevice currentInputDevice];
Similarly, to get all the available output devices use the outputDevices
class method:
NSArray *outputDevices = [EZAudioDevice outputDevices];
or to just get the currently selected output device use the currentInputDevice
method:
// On iOS this will default to the headset speaker, while on OSX this will be your selected
// output device from the Sound preferences
EZAudioDevice *currentOutputDevice = [EZAudioDevice currentOutputDevice];
Provides access to the default device microphone in one line of code and provides delegate callbacks to receive the audio data as an AudioBufferList and float arrays.
Relevant Example Projects
- EZAudioCoreGraphicsWaveformExample (iOS)
- EZAudioCoreGraphicsWaveformExample (OSX)
- EZAudioOpenGLWaveformExample (iOS)
- EZAudioOpenGLWaveformExample (OSX)
- EZAudioRecordExample (iOS)
- EZAudioRecordExample (OSX)
- EZAudioPassThroughExample (iOS)
- EZAudioPassThroughExample (OSX)
- EZAudioFFTExample (iOS)
- EZAudioFFTExample (OSX)
Create an EZMicrophone
instance by declaring a property and initializing it like so:
// Declare the EZMicrophone as a strong property
@property (nonatomic, strong) EZMicrophone *microphone;
...
// Initialize the microphone instance and assign it a delegate to receive the audio data
// callbacks
self.microphone = [EZMicrophone microphoneWithDelegate:self];
Alternatively, you could also use the shared EZMicrophone
instance and just assign its EZMicrophoneDelegate
.
// Assign a delegate to the shared instance of the microphone to receive the audio data
// callbacks
[EZMicrophone sharedMicrophone].delegate = self;
The EZMicrophone
uses an EZAudioDevice
instance to select what specific hardware destination it will use to pull audio data. You'd use this if you wanted to change the input device like in the EZAudioCoreGraphicsWaveformExample for iOS or OSX. At any time you can change which input device is used by setting the device property:
NSArray *inputs = [EZAudioDevice inputDevices];
[self.microphone setDevice:[inputs lastObject]];
Anytime the EZMicrophone
changes its device it will trigger the EZMicrophoneDelegate
event:
- (void)microphone:(EZMicrophone *)microphone changedDevice:(EZAudioDevice *)device
{
// This is not always guaranteed to occur on the main thread so make sure you
// wrap it in a GCD block
dispatch_async(dispatch_get_main_queue(), ^{
// Update UI here
NSLog(@"Changed input device: %@", device);
});
}
Note: For iOS this can happen automatically if the AVAudioSession changes the current device.
To tell the microphone to start fetching audio use the startFetchingAudio
function.
// Starts fetching audio from the default device microphone and sends data to EZMicrophoneDelegate
[self.microphone startFetchingAudio];
Once the EZMicrophone
has started it will send the EZMicrophoneDelegate
the audio back in a few ways.
An array of float arrays:
/**
The microphone data represented as non-interleaved float arrays useful for:
- Creating real-time waveforms using EZAudioPlot or EZAudioPlotGL
- Creating any number of custom visualizations that utilize audio!
*/
-(void) microphone:(EZMicrophone *)microphone
hasAudioReceived:(float **)buffer
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
{
__weak typeof (self) weakSelf = self;
// Getting audio data as an array of float buffer arrays that can be fed into the
// EZAudioPlot, EZAudioPlotGL, or whatever visualization you would like to do with
// the microphone data.
dispatch_async(dispatch_get_main_queue(),^{
// Visualize this data brah, buffer[0] = left channel, buffer[1] = right channel
[weakSelf.audioPlot updateBuffer:buffer[0] withBufferSize:bufferSize];
});
}
or the AudioBufferList representation:
/**
The microphone data represented as CoreAudio's AudioBufferList useful for:
- Appending data to an audio file via the EZRecorder
- Playback via the EZOutput
*/
-(void) microphone:(EZMicrophone *)microphone
hasBufferList:(AudioBufferList *)bufferList
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
{
// Getting audio data as an AudioBufferList that can be directly fed into the EZRecorder
// or EZOutput. Say whattt...
}
Pause or resume fetching audio at any time like so:
// Stop fetching audio
[self.microphone stopFetchingAudio];
// Resume fetching audio
[self.microphone startFetchingAudio];
Alternatively, you could also toggle the microphoneOn
property (safe to use with Cocoa Bindings)
// Stop fetching audio
self.microphone.microphoneOn = NO;
// Start fetching audio
self.microphone.microphoneOn = YES;
Provides flexible playback to the default output device by asking the EZOutputDataSource
for audio data to play. Doesn't care where the buffers come from (microphone, audio file, streaming audio, etc). As of 1.0.0 the EZOutputDataSource
has been simplified to have only one method to provide audio data to your EZOutput
instance.
// The EZOutputDataSource should fill out the audioBufferList with the given frame count.
// The timestamp is provided for sample accurate calculation, but for basic use cases can
// be ignored.
- (OSStatus) output:(EZOutput *)output
shouldFillAudioBufferList:(AudioBufferList *)audioBufferList
withNumberOfFrames:(UInt32)frames
timestamp:(const AudioTimeStamp *)timestamp;
Relevant Example Projects
- EZAudioPlayFileExample (iOS)
- EZAudioPlayFileExample (OSX)
- EZAudioPassThroughExample (iOS)
- EZAudioPassThroughExample (OSX)
Create an EZOutput
by declaring a property and initializing it like so:
// Declare the EZOutput as a strong property
@property (nonatomic, strong) EZOutput *output;
...
// Initialize the EZOutput instance and assign it a delegate to provide the output audio data
self.output = [EZOutput outputWithDataSource:self];
Alternatively, you could also use the shared output instance and just assign it an EZOutputDataSource
if you will only have one EZOutput
instance for your application.
// Assign a delegate to the shared instance of the output to provide the output audio data
[EZOutput sharedOutput].delegate = self;
The EZOutput
uses an EZAudioDevice
instance to select what specific hardware destination it will output audio to. You'd use this if you wanted to change the output device like in the EZAudioPlayFileExample for OSX. At any time you can change which output device is used by setting the device
property:
// By default the EZOutput uses the default output device, but you can change this at any time
EZAudioDevice *currentOutputDevice = [EZAudioDevice currentOutputDevice];
[self.output setDevice:currentOutputDevice];
Anytime the EZOutput
changes its device it will trigger the EZOutputDelegate
event:
- (void)output:(EZOutput *)output changedDevice:(EZAudioDevice *)device
{
NSLog(@"Change output device to: %@", device);
}
When providing audio data the EZOutputDataSource
will expect you to fill out the AudioBufferList provided with whatever inputFormat
that is set on the EZOutput
. By default the input format is a stereo, non-interleaved, float format (see defaultInputFormat for more information). If you're dealing with a different input format (which is typically the case), just set the inputFormat
property. For instance:
// Set a mono, float format with a sample rate of 44.1 kHz
AudioStreamBasicDescription monoFloatFormat = [EZAudioUtilities monoFloatFormatWithSampleRate:44100.0f];
[self.output setInputFormat:monoFloatFormat];
An example of implementing the EZOutputDataSource
is done internally in the EZAudioPlayer
using an EZAudioFile
to read audio from an audio file on disk like so:
- (OSStatus) output:(EZOutput *)output
shouldFillAudioBufferList:(AudioBufferList *)audioBufferList
withNumberOfFrames:(UInt32)frames
timestamp:(const AudioTimeStamp *)timestamp
{
if (self.audioFile)
{
UInt32 bufferSize; // amount of frames actually read
BOOL eof; // end of file
[self.audioFile readFrames:frames
audioBufferList:audioBufferList
bufferSize:&bufferSize
eof:&eof];
if (eof && [self.delegate respondsToSelector:@selector(audioPlayer:reachedEndOfAudioFile:)])
{
[self.delegate audioPlayer:self reachedEndOfAudioFile:self.audioFile];
}
if (eof && self.shouldLoop)
{
[self seekToFrame:0];
}
else if (eof)
{
[self pause];
[self seekToFrame:0];
[[NSNotificationCenter defaultCenter] postNotificationName:EZAudioPlayerDidReachEndOfFileNotification
object:self];
}
}
return noErr;
}
I created a sample project that uses the EZOutput
to act as a signal generator to play sine, square, triangle, sawtooth, and noise waveforms. Here's a snippet of code to generate a sine tone:
...
double const SAMPLE_RATE = 44100.0;
- (void)awakeFromNib
{
//
// Create EZOutput to play audio data with mono format (EZOutput will convert
// this mono, float "inputFormat" to a clientFormat, i.e. the stereo output format).
//
AudioStreamBasicDescription inputFormat = [EZAudioUtilities monoFloatFormatWithSampleRate:SAMPLE_RATE];
self.output = [EZOutput outputWithDataSource:self inputFormat:inputFormat];
[self.output setDelegate:self];
self.frequency = 200.0;
self.sampleRate = SAMPLE_RATE;
self.amplitude = 0.80;
}
- (OSStatus) output:(EZOutput *)output
shouldFillAudioBufferList:(AudioBufferList *)audioBufferList
withNumberOfFrames:(UInt32)frames
timestamp:(const AudioTimeStamp *)timestamp
{
Float32 *buffer = (Float32 *)audioBufferList->mBuffers[0].mData;
size_t bufferByteSize = (size_t)audioBufferList->mBuffers[0].mDataByteSize;
double theta = self.theta;
double frequency = self.frequency;
double thetaIncrement = 2.0 * M_PI * frequency / SAMPLE_RATE;
if (self.type == GeneratorTypeSine)
{
for (UInt32 frame = 0; frame < frames; frame++)
{
buffer[frame] = self.amplitude * sin(theta);
theta += thetaIncrement;
if (theta > 2.0 * M_PI)
{
theta -= 2.0 * M_PI;
}
}
self.theta = theta;
}
else if (... other shapes in full source)
}
For the full implementation of the square, triangle, sawtooth, and noise functions here: (https://github.com/syedhali/SineExample/blob/master/SineExample/GeneratorViewController.m#L220-L305)
Once the EZOutput
has started it will send the EZOutputDelegate
the audio back as float arrays for visualizing. These are converted inside the EZOutput
component from whatever input format you may have provided. For instance, if you provide an interleaved, signed integer AudioStreamBasicDescription for the inputFormat
property then that will be automatically converted to a stereo, non-interleaved, float format when sent back in the delegate playedAudio:...
method below:
An array of float arrays:
/**
The output data represented as non-interleaved float arrays useful for:
- Creating real-time waveforms using EZAudioPlot or EZAudioPlotGL
- Creating any number of custom visualizations that utilize audio!
*/
- (void) output:(EZOutput *)output
playedAudio:(float **)buffer
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
{
__weak typeof (self) weakSelf = self;
dispatch_async(dispatch_get_main_queue(), ^{
// Update plot, buffer[0] = left channel, buffer[1] = right channel
});
}
Pause or resume the output component at any time like so:
// Stop fetching audio
[self.output stopPlayback];
// Resume fetching audio
[self.output startPlayback];
Internally the EZOutput
is using an AUGraph to chain together a converter, mixer, and output audio units. You can hook into this graph by subclassing EZOutput
and implementing the method:
// By default this method connects the AUNode representing the input format converter to
// the mixer node. In subclasses you can add effects in the chain between the converter
// and mixer by creating additional AUNodes, adding them to the AUGraph provided below,
// and then connecting them together.
- (OSStatus)connectOutputOfSourceNode:(AUNode)sourceNode
sourceNodeOutputBus:(UInt32)sourceNodeOutputBus
toDestinationNode:(AUNode)destinationNode
destinationNodeInputBus:(UInt32)destinationNodeInputBus
inGraph:(AUGraph)graph;
This was inspired by the audio processing graph from CocoaLibSpotify (Daniel Kennett of Spotify has an excellent blog post explaining how to add an EQ to the CocoaLibSpotify AUGraph).
Here's an example of how to add a delay audio unit (kAudioUnitSubType_Delay
):
// In interface, declare delay node info property
@property (nonatomic, assign) EZAudioNodeInfo *delayNodeInfo;
// In implementation, overwrite the connection method
- (OSStatus)connectOutputOfSourceNode:(AUNode)sourceNode
sourceNodeOutputBus:(UInt32)sourceNodeOutputBus
toDestinationNode:(AUNode)destinationNode
destinationNodeInputBus:(UInt32)destinationNodeInputBus
inGraph:(AUGraph)graph
{
self.delayNodeInfo = (EZAudioNodeInfo *)malloc(sizeof(EZAudioNodeInfo));
// A description for the time/pitch shifter Device
AudioComponentDescription delayComponentDescription;
delayComponentDescription.componentType = kAudioUnitType_Effect;
delayComponentDescription.componentSubType = kAudioUnitSubType_Delay;
delayComponentDescription.componentManufacturer = kAudioUnitManufacturer_Apple;
delayComponentDescription.componentFlags = 0;
delayComponentDescription.componentFlagsMask = 0;
[EZAudioUtilities checkResult:AUGraphAddNode(graph,
&delayComponentDescription,
&self.delayNodeInfo->node)
operation:"Failed to add node for time shift"];
// Get the time/pitch shifter Audio Unit from the node
[EZAudioUtilities checkResult:AUGraphNodeInfo(graph,
self.delayNodeInfo->node,
NULL,
&self.delayNodeInfo->audioUnit)
operation:"Failed to get audio unit for delay node"];
// connect the output of the input source node to the input of the time/pitch shifter node
[EZAudioUtilities checkResult:AUGraphConnectNodeInput(graph,
sourceNode,
sourceNodeOutputBus,
self.delayNodeInfo->node,
0)
operation:"Failed to connect source node into delay node"];
// connect the output of the time/pitch shifter node to the input of the destination node, thus completing the chain.
[EZAudioUtilities checkResult:AUGraphConnectNodeInput(graph,
self.delayNodeInfo->node,
0,
destinationNode,
destinationNodeInputBus)
operation:"Failed to connect delay to destination node"];
return noErr;
}
// Clean up
- (void)dealloc
{
free(self.delayNodeInfo);
}
Provides simple read/seek operations, pulls waveform amplitude data, and provides the EZAudioFileDelegate
to notify of any read/seek action occuring on the EZAudioFile
. This can be thought of as the NSImage/UIImage equivalent of the audio world.
Relevant Example Projects
- EZAudioWaveformFromFileExample (iOS)
- EZAudioWaveformFromFileExample (OSX)
To open an audio file create a new instance of the EZAudioFile
class.
// Declare the EZAudioFile as a strong property
@property (nonatomic, strong) EZAudioFile *audioFile;
...
// Initialize the EZAudioFile instance and assign it a delegate to receive the read/seek callbacks
self.audioFile = [EZAudioFile audioFileWithURL:[NSURL fileURLWithPath:@"/path/to/your/file"] delegate:self];
The EZAudioFile allows you to quickly fetch waveform data from an audio file with as much or little detail as you'd like.
__weak typeof (self) weakSelf = self;
// Get a waveform with 1024 points of data. We can adjust the number of points to whatever level
// of detail is needed by the application
[self.audioFile getWaveformDataWithNumberOfPoints:1024
completionBlock:^(float **waveformData,
int length)
{
[weakSelf.audioPlot updateBuffer:waveformData[0]
withBufferSize:length];
}];
Reading audio data from a file requires you to create an AudioBufferList to hold the data. The EZAudio
utility function, audioBufferList
, provides a convenient way to get an allocated AudioBufferList to use. There is also a utility function, freeBufferList:
, to use to free (or release) the AudioBufferList when you are done using that audio data.
Note: You have to free the AudioBufferList, even in ARC.
// Allocate an AudioBufferList to hold the audio data (the client format is the non-compressed
// in-app format that is used for reading, it's different than the file format which is usually
// something compressed like an mp3 or m4a)
AudioStreamBasicDescription clientFormat = [self.audioFile clientFormat];
UInt32 numberOfFramesToRead = 512;
UInt32 channels = clientFormat.mChannelsPerFrame;
BOOL isInterleaved = [EZAudioUtilities isInterleaved:clientFormat];
AudioBufferList *bufferList = [EZAudioUtilities audioBufferListWithNumberOfFrames:numberOfFramesToRead
numberOfChannels:channels
interleaved:isInterleaved];
// Read the frames from the EZAudioFile into the AudioBufferList
UInt32 framesRead;
UInt32 isEndOfFile;
[self.audioFile readFrames:numberOfFramesToRead
audioBufferList:bufferList
bufferSize:&framesRead
eof:&isEndOfFile]
When a read occurs the EZAudioFileDelegate
receives two events.
An event notifying the delegate of the read audio data as float arrays:
-(void) audioFile:(EZAudioFile *)audioFile
readAudio:(float **)buffer
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
{
__weak typeof (self) weakSelf = self;
dispatch_async(dispatch_get_main_queue(), ^{
[weakSelf.audioPlot updateBuffer:buffer[0]
withBufferSize:bufferSize];
});
}
and an event notifying the delegate of the new frame position within the EZAudioFile
:
-(void)audioFile:(EZAudioFile *)audioFile updatedPosition:(SInt64)framePosition
{
__weak typeof (self) weakSelf = self;
dispatch_async(dispatch_get_main_queue(), ^{
// Update UI
});
}
You can seek very easily through an audio file using the EZAudioFile
's seekToFrame: method. The EZAudioFile
provides a totalFrames
method to provide you the total amount of frames in an audio file so you can calculate a proper offset.
// Get the total number of frames for the audio file
SInt64 totalFrames = [self.audioFile totalFrames];
// Seeks halfway through the audio file
[self.audioFile seekToFrame:(totalFrames/2)];
// Alternatively, you can seek using seconds
NSTimeInterval duration = [self.audioFile duration];
[self.audioFile setCurrentTime:duration/2.0];
When a seek occurs the EZAudioFileDelegate
receives the seek event:
-(void)audioFile:(EZAudioFile *)audioFile updatedPosition:(SInt64)framePosition
{
__weak typeof (self) weakSelf = self;
dispatch_async(dispatch_get_main_queue(), ^{
// Update UI
});
}
Provides a class that combines the EZAudioFile
and EZOutput
for file playback of all Core Audio supported formats to any hardware device. Because the EZAudioPlayer
internally hooks into the EZAudioFileDelegate
and EZOutputDelegate
, you should implement the EZAudioPlayerDelegate
to receive the playedAudio:...
and updatedPosition:
events. The EZAudioPlayFileExample projects for iOS and OSX shows how to use the EZAudioPlayer
to play audio files, visualize the samples with an audio plot, adjust the volume, and change the output device using the EZAudioDevice
class. The EZAudioPlayer
primarily uses NSNotificationCenter
to post notifications because often times you have one audio player and multiple UI elements that need to listen for player events to properly update.
// Declare the EZAudioFile as a strong property
@property (nonatomic, strong) EZAudioFile *audioFile;
...
// Create an EZAudioPlayer with a delegate that conforms to EZAudioPlayerDelegate
self.player = [EZAudioPlayer audioPlayerWithDelegate:self];
The EZAudioPlayer
uses an internal EZAudioFile
to provide data to its EZOutput
for output via the EZOutputDataSource
. You can provide an EZAudioFile
by just setting the audioFile
property on the EZAudioPlayer
will make a copy of the EZAudioFile
at that file path url for its own use.
// Set the EZAudioFile for playback by setting the `audioFile` property
EZAudioFile *audioFile = [EZAudioFile audioFileWithURL:[NSURL fileURLWithPath:@"/path/to/your/file"]];
[self.player setAudioFile:audioFile];
// This, however, will not pause playback if a current file is playing. Instead
// it's encouraged to use `playAudioFile:` instead if you're swapping in a new
// audio file while playback is already running
EZAudioFile *audioFile = [EZAudioFile audioFileWithURL:[NSURL fileURLWithPath:@"/path/to/your/file"]];
[self.player playAudioFile:audioFile];
As audio is played the EZAudioPlayerDelegate
will receive the playedAudio:...
, updatedPosition:...
, and, if the audio file reaches the end of the file, the reachedEndOfAudioFile:
events. A typical implementation of the EZAudioPlayerDelegate
would be something like:
- (void) audioPlayer:(EZAudioPlayer *)audioPlayer
playedAudio:(float **)buffer
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
inAudioFile:(EZAudioFile *)audioFile
{
__weak typeof (self) weakSelf = self;
// Update an EZAudioPlot or EZAudioPlotGL to reflect the audio data coming out
// of the EZAudioPlayer (post volume and pan)
dispatch_async(dispatch_get_main_queue(), ^{
[weakSelf.audioPlot updateBuffer:buffer[0]
withBufferSize:bufferSize];
});
}
//------------------------------------------------------------------------------
- (void)audioPlayer:(EZAudioPlayer *)audioPlayer
updatedPosition:(SInt64)framePosition
inAudioFile:(EZAudioFile *)audioFile
{
__weak typeof (self) weakSelf = self;
// Update any UI controls including sliders and labels
// display current time/duration
dispatch_async(dispatch_get_main_queue(), ^{
if (!weakSelf.positionSlider.highlighted)
{
weakSelf.positionSlider.floatValue = (float)framePosition;
weakSelf.positionLabel.integerValue = framePosition;
}
});
}
You can seek through the audio file in a similar fashion as with the EZAudioFile
. That is, using the seekToFrame:
or currentTime
property.
// Get the total number of frames and seek halfway
SInt64 totalFrames = [self.player totalFrames];
[self.player seekToFrame:(totalFrames/2)];
// Alternatively, you can seek using seconds
NSTimeInterval duration = [self.player duration];
[self.player setCurrentTime:duration/2.0];
Because the EZAudioPlayer
wraps the EZOutput
you can adjust the volume and pan parameters for playback.
// Make it half as loud, 0 = silence, 1 = full volume. Default is 1.
[self.player setVolume:0.5];
// Make it only play on the left, -1 = left, 1 = right. Default is 0.0 (center)
[self.player setPan:-1.0];
The EZAudioPlayer
wraps the EZAudioFile
and provides a high level interface for pulling values like current time, duration, the frame index, total frames, etc.
NSTimeInterval currentTime = [self.player currentTime];
NSTimeInterval duration = [self.player duration];
NSString *formattedCurrentTime = [self.player formattedCurrentTime]; // MM:SS formatted
NSString *formattedDuration = [self.player formattedDuration]; // MM:SS formatted
SInt64 frameIndex = [self.player frameIndex];
SInt64 totalFrames = [self.player totalFrames];
In addition, the EZOutput
properties are also offered at a high level as well:
EZAudioDevice *outputDevice = [self.player device];
BOOL isPlaying = [self.player isPlaying];
float pan = [self.player pan];
float volume = [self.player volume];
The EZAudioPlayer
provides the following notifications (as of 1.1.2):
/**
Notification that occurs whenever the EZAudioPlayer changes its `audioFile` property. Check the new value using the EZAudioPlayer's `audioFile` property.
*/
FOUNDATION_EXPORT NSString * const EZAudioPlayerDidChangeAudioFileNotification;
/**
Notification that occurs whenever the EZAudioPlayer changes its `device` property. Check the new value using the EZAudioPlayer's `device` property.
*/
FOUNDATION_EXPORT NSString * const EZAudioPlayerDidChangeOutputDeviceNotification;
/**
Notification that occurs whenever the EZAudioPlayer changes its `output` component's `pan` property. Check the new value using the EZAudioPlayer's `pan` property.
*/
FOUNDATION_EXPORT NSString * const EZAudioPlayerDidChangePanNotification;
/**
Notification that occurs whenever the EZAudioPlayer changes its `output` component's play state. Check the new value using the EZAudioPlayer's `isPlaying` property.
*/
FOUNDATION_EXPORT NSString * const EZAudioPlayerDidChangePlayStateNotification;
/**
Notification that occurs whenever the EZAudioPlayer changes its `output` component's `volume` property. Check the new value using the EZAudioPlayer's `volume` property.
*/
FOUNDATION_EXPORT NSString * const EZAudioPlayerDidChangeVolumeNotification;
/**
Notification that occurs whenever the EZAudioPlayer has reached the end of a file and its `shouldLoop` property has been set to NO.
*/
FOUNDATION_EXPORT NSString * const EZAudioPlayerDidReachEndOfFileNotification;
/**
Notification that occurs whenever the EZAudioPlayer performs a seek via the `seekToFrame` method or `setCurrentTime:` property setter. Check the new `currentTime` or `frameIndex` value using the EZAudioPlayer's `currentTime` or `frameIndex` property, respectively.
*/
FOUNDATION_EXPORT NSString * const EZAudioPlayerDidSeekNotification;
Provides a way to record any audio source to an audio file. This hooks into the other components quite nicely to do something like plot the audio waveform while recording to give visual feedback as to what is happening. The EZRecorderDelegate
provides methods to listen to write events and a final close event on the EZRecorder
(explained below).
Relevant Example Projects
- EZAudioRecordExample (iOS)
- EZAudioRecordExample (OSX)
To create an EZRecorder
you must provide at least 3 things: an NSURL representing the file path of where the audio file should be written to (an existing file will be overwritten), a clientFormat
representing the format in which you will be providing the audio data, and either an EZRecorderFileType
or an AudioStreamBasicDescription
representing the file format of the audio data on disk.
// Provide a file path url to write to, a client format (always linear PCM, this is the format
// coming from another component like the EZMicrophone's audioStreamBasicDescription property),
// and a EZRecorderFileType constant representing either a wav (EZRecorderFileTypeWAV),
// aiff (EZRecorderFileTypeAIFF), or m4a (EZRecorderFileTypeM4A) file format. The advantage of
// this is that the `fileFormat` property will be automatically filled out for you.
+ (instancetype)recorderWithURL:(NSURL *)url
clientFormat:(AudioStreamBasicDescription)clientFormat
fileType:(EZRecorderFileType)fileType;
// Alternatively, you can provide a file path url to write to, a client format (always linear
// PCM, this is the format coming from another component like the EZMicrophone's
// audioStreamBasicDescription property), a `fileFormat` representing your custom
// AudioStreamBasicDescription, and an AudioFileTypeID that corresponds with your `fileFormat`.
+ (instancetype)recorderWithURL:(NSURL *)url
clientFormat:(AudioStreamBasicDescription)clientFormat
fileFormat:(AudioStreamBasicDescription)fileFormat
audioFileTypeID:(AudioFileTypeID)audioFileTypeID;
Start by declaring an instance of the EZRecorder (you will have one of these per audio file written out)
// Declare the EZRecorder as a strong property
@property (nonatomic, strong) EZRecorder *recorder;
and initialize it using one of the two initializers from above. For instance, using the EZRecorderFileType
shortcut initializer you could create an instance like so:
// Example using an EZMicrophone and a string called kAudioFilePath representing a file
// path location on your computer to write out a M4A file.
self.recorder = [EZRecorder recorderWithURL:[NSURL fileURLWithPath:@"/path/to/your/file.m4a"]
clientFormat:[self.microphone audioStreamBasicDescription]
fileType:EZRecorderFileTypeM4A];
or to configure your own custom file format, say to write out a 8000 Hz, iLBC file:
// Example using an EZMicrophone, a string called kAudioFilePath representing a file
// path location on your computer, and an iLBC file format.
AudioStreamBasicDescription iLBCFormat = [EZAudioUtilities iLBCFormatWithSampleRate:8000];
self.recorder = [EZRecorder recorderWithURL:[NSURL fileURLWithPath:@"/path/to/your/file.caf"]
clientFormat:[self.microphone audioStreamBasicDescription]
fileFormat:iLBCFormat
audioFileTypeID:kAudioFileCAFType];
Once you've initialized your EZRecorder
you can append data by passing in an AudioBufferList and its buffer size like so:
// Append the microphone data coming as a AudioBufferList with the specified buffer size
// to the recorder
-(void) microphone:(EZMicrophone *)microphone
hasBufferList:(AudioBufferList *)bufferList
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
{
// Getting audio data as a buffer list that can be directly fed into the EZRecorder. This is
// happening on the audio thread - any UI updating needs a GCD main queue block.
if (self.isRecording)
{
// Since we set the recorder's client format to be that of the EZMicrophone instance,
// the audio data coming in represented by the AudioBufferList can directly be provided
// to the EZRecorder. The EZRecorder will internally convert the audio data from the
// `clientFormat` to `fileFormat`.
[self.recorder appendDataFromBufferList:bufferList
withBufferSize:bufferSize];
}
}
Once audio data has been successfully written with the EZRecorder
it will notify the EZRecorderDelegate
of the event so it can respond via:
// Triggers after the EZRecorder's `appendDataFromBufferList:withBufferSize:` method is called
// so you can update your interface accordingly.
- (void)recorderUpdatedCurrentTime:(EZRecorder *)recorder
{
__weak typeof (self) weakSelf = self;
// This will get triggerd on the thread that the write occured on so be sure to wrap your UI
// updates in a GCD main queue block! However, I highly recommend you first pull the values
// you'd like to update the interface with before entering the GCD block to avoid trying to
// fetch a value after the audio file has been closed.
NSString *formattedCurrentTime = [recorder formattedCurrentTime]; // MM:SS formatted
dispatch_async(dispatch_get_main_queue(), ^{
// Update label
weakSelf.currentTimeLabel.stringValue = formattedCurrentTime;
});
}
When you're recording is done be sure to call the closeAudioFile
method to make sure the audio file written to disk is properly closed before you attempt to read it again.
// Close the EZRecorder's audio file BEFORE reading
[self.recorder closeAudioFile];
This will trigger the EZRecorder's delegate method:
- (void)recorderDidClose:(EZRecorder *)recorder
{
recorder.delegate = nil;
}
EZAudio
currently offers two drop in audio waveform components that help simplify the process of visualizing audio.
Provides an audio waveform plot that uses CoreGraphics to perform the drawing. On iOS this is a subclass of UIView while on OSX this is a subclass of NSView. As of the 1.0.0 release, the waveforms are drawn using CALayers where compositing is done on the GPU. As a result, there have been some huge performance gains and CPU usage per real-time (i.e. 60 frames per second redrawing) plot is now about 2-3% CPU as opposed to the 20-30% we were experiencing before.
Relevant Example Projects
- EZAudioCoreGraphicsWaveformExample (iOS)
- EZAudioCoreGraphicsWaveformExample (OSX)
- EZAudioRecordExample (iOS)
- EZAudioRecordExample (OSX)
- EZAudioWaveformFromFileExample (iOS)
- EZAudioWaveformFromFileExample (OSX)
- EZAudioFFTExample (iOS)
- EZAudioFFTExample (OSX)
You can create an audio plot in the interface builder by dragging in a UIView on iOS or an NSView on OSX onto your content area. Then change the custom class of the UIView/NSView to EZAudioPlot
.
Alternatively, you can could create the audio plot programmatically
// Programmatically create an audio plot
EZAudioPlot *audioPlot = [[EZAudioPlot alloc] initWithFrame:self.view.frame];
[self.view addSubview:audioPlot];
All plots offer the ability to change the background color, waveform color, plot type (buffer or rolling), toggle between filled and stroked, and toggle between mirrored and unmirrored (about the x-axis). For iOS colors are of the type UIColor while on OSX colors are of the type NSColor.
// Background color (use UIColor for iOS)
audioPlot.backgroundColor = [NSColor colorWithCalibratedRed:0.816
green:0.349
blue:0.255
alpha:1];
// Waveform color (use UIColor for iOS)
audioPlot.color = [NSColor colorWithCalibratedRed:1.000
green:1.000
blue:1.000
alpha:1];
// Plot type
audioPlot.plotType = EZPlotTypeBuffer;
// Fill
audioPlot.shouldFill = YES;
// Mirror
audioPlot.shouldMirror = YES;
Also, as of iOS 8 you can adjust the background color, color, gain, shouldFill, and shouldMirror parameters directly in the Interface Builder via the IBInspectable attributes:
All plots have only one update function, updateBuffer:withBufferSize:
, which expects a float array and its length.
// The microphone component provides audio data to its delegate as an array of float buffer arrays.
- (void) microphone:(EZMicrophone *)microphone
hasAudioReceived:(float **)buffer
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
{
/**
Update the audio plot using the float array provided by the microphone:
buffer[0] = left channel
buffer[1] = right channel
Note: Audio updates happen asynchronously so we need to make sure
sure to update the plot on the main thread
*/
__weak typeof (self) weakSelf = self;
dispatch_async(dispatch_get_main_queue(), ^{
[weakSelf.audioPlot updateBuffer:buffer[0] withBufferSize:bufferSize];
});
}
Provides an audio waveform plot that uses OpenGL to perform the drawing. The API this class are exactly the same as those for the EZAudioPlot above. On iOS this is a subclass of the GLKView while on OSX this is a subclass of the NSOpenGLView. In most cases this is the plot you want to use, it's GPU-accelerated, can handle lots of points while displaying 60 frames per second (the EZAudioPlot starts to choke on anything greater than 1024), and performs amazingly on all devices. The only downside is that you can only have one OpenGL plot onscreen at a time. However, you can combine OpenGL plots with Core Graphics plots in the view hierachy (see the EZAudioRecordExample for an example of how to do this).
Relevant Example Projects
- EZAudioOpenGLWaveformExample (iOS)
- EZAudioOpenGLWaveformExample (OSX)
- EZAudioPlayFileExample (iOS)
- EZAudioPlayFileExample (OSX)
- EZAudioRecordExample (iOS)
- EZAudioRecordExample (OSX)
- EZAudioPassThroughExample (iOS)
- EZAudioPassThroughExample (OSX)
You can create an audio plot in the interface builder by dragging in a UIView on iOS or an NSView on OSX onto your content area. Then change the custom class of the UIView/NSView to EZAudioPlotGL
.
Alternatively, you can could create the EZAudioPlotGL
programmatically
// Programmatically create an audio plot
EZAudioPlotGL *audioPlotGL = [[EZAudioPlotGL alloc] initWithFrame:self.view.frame];
[self.view addSubview:audioPlotGL];
All plots offer the ability to change the background color, waveform color, plot type (buffer or rolling), toggle between filled and stroked, and toggle between mirrored and unmirrored (about the x-axis). For iOS colors are of the type UIColor while on OSX colors are of the type NSColor.
// Background color (use UIColor for iOS)
audioPlotGL.backgroundColor = [NSColor colorWithCalibratedRed:0.816
green:0.349
blue:0.255
alpha:1];
// Waveform color (use UIColor for iOS)
audioPlotGL.color = [NSColor colorWithCalibratedRed:1.000
green:1.000
blue:1.000
alpha:1];
// Plot type
audioPlotGL.plotType = EZPlotTypeBuffer;
// Fill
audioPlotGL.shouldFill = YES;
// Mirror
audioPlotGL.shouldMirror = YES;
Also, as of iOS 8 you can adjust the background color, color, gain, shouldFill, and shouldMirror parameters directly in the Interface Builder via the IBInspectable attributes:
All plots have only one update function, updateBuffer:withBufferSize:
, which expects a float array and its length.
// The microphone component provides audio data to its delegate as an array of float buffer arrays.
- (void) microphone:(EZMicrophone *)microphone
hasAudioReceived:(float **)buffer
withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels
{
/**
Update the audio plot using the float array provided by the microphone:
buffer[0] = left channel
buffer[1] = right channel
Note: Audio updates happen asynchronously so we need to make sure
sure to update the plot on the main thread
*/
__weak typeof (self) weakSelf = self;
dispatch_async(dispatch_get_main_queue(), ^{
[weakSelf.audioPlotGL updateBuffer:buffer[0] withBufferSize:bufferSize];
});
}
EZAudio is available under the MIT license. See the LICENSE file for more info.
Syed Haris Ali
www.syedharisali.com
syedhali07[at]gmail.com
The following people rock:
- My brother, Reza Ali, for walking me through all the gritty details of OpenGL and his constant encouragement through this journey to 1.0.0.
- Aure Prochazka for his amazing work on AudioKit and his encouragement to bring EZAudio to 1.0.0
- Daniel Kennett for writing this great blog post that inspired the rewrite of the
EZOutput
in 1.0.0. - Michael Tyson for creating the TPCircularBuffer and all his contributions to the community including the Amazing Audio Engine, Audiobus, and all the tasty pixel blog posts.
- Chris Adamson and Kevin Avila for writing the amazing Learning Core Audio book.
As of today, June 13, 2016, I’m officially deprecating EZAudio. I’d like to thank everyone for the support over the last few years I’ve been hacking on EZAudio and working to make it better.
The best alternative to EZAudio is now AudioKit. Note that The Amazing Audio Engine and The Amazing Audio Engine 2 have now both been retired as well. Any further contributions I make to iOS/macOS/tvOS audio programming will be to AudioKit.
EZAudio started as a pet project of mine in 2013 in an attempt to reduce the amount of duplicate code I was writing for my own iOS/Mac audio projects. I originally just wanted to record audio from an iPhone’s mic and plot a waveform. This over time quickly grew into the collection of classes you may know today (EZMicrophone, EZAudioPlot, etc).
I apologize for not being more active in addressing the issues and pull requests, but I’m hoping you all understand I’m only one person. EZAudio was solely written and maintained by me out of love during weeks of time I wasn’t making any money, all while living in one of the most expensive cities in the world. Like many of you, I spend the majority of my time working full-time to sustain myself (I have bills and rent to pay too!). I’m transitioning to non-audio ventures and will hopefully be able to share another cool open source project soon enough.
Getting the opportunity to work on EZAudio with all of you has been an incredibly insightful experience and I’m incredibly grateful to have gotten a chance to share it with all of you. Thank you.
As I’m writing this I’m a little hesitant to say it’s all broken (despite all the issues filed), but I’d recommend forking this repo from this point forward and using at your own risk. I’ll probably be continuing to use EZAudio’s EZAudioPlot for rendering waveforms in my future projects, however, please don’t expect any updates to this repo. Deprecated means I won’t be responding to issues, EZAudio-related emails, or pushing any new changes to this repo.