Bitesize Android KitKat: Week 5: Measuring the sound output level with Visualizer

Written by Sam Davies

The Visualizer class has been around since Gingerbread (API level 9), but KitKat sees the introduction of new functionality for measuring the current level of output sounds. It has a pretty simple API, and in this article we’ll take a look at how to use this new functionality by building a sound meter app which will measure the level of the device audio output.

The source code for this project is available in the GitHub repo at It’s a gradle based project, and can easily be imported into Android Studio. It has been tested using Android Studio 0.4.4.

This post is part of a series of articles – Bitesize Android KitKat, which takes a look at some of the new features available to developers as part of the KitKat release of Android. Each article is backed up with a sample app which demonstrates how to use the feature in a real scenario, with all the source code available on GitHub. Check the index for a list of all the articles published so far.

The Visualizer class

The Visualizer class is provided to enable the sampling of currently-playing data for some kind of visualization – hence the name. It has traditionally been able to provide samples in the temporal or frequency domains, but the advent of KitKat adds a measurementAPI. The measurements contain both peak and RMS values, and are accessed using the getMeasurementPeakRms() method. In order to get regularly updating measurement values, we’ll create a class which regularly requests these measurement values and then publishes them – SamplingSoundMeter:

Building the SamplingSoundMeter

The SamplingSoundMeter class has a constructor which takes just one argument – a so-called SoundLevelUpdateable object. We’ll take a look at this interface in a moment.

public class SamplingSoundMeter {
    private SoundLevelUpdateable updateable;
    private Handler handler;
    private Runnable sampler;
    private Visualizer visualizer;

    public SamplingSoundMeter(SoundLevelUpdateable updateableObject) {
        updateable = updateableObject;
        visualizer = new Visualizer(0);
        handler = new Handler();

There are member variables for several components – including saving the provided SoundLevelUpdateable object. We create a new visualizer object – providing the argument of 0. This value is the audioSession. If you have a particular MediaPlayer or AudioTrack then you can specify the audio session to which it belongs to attach the Visualizer to the same session. Here we use a session ID of 0 – which means that the visualizer will apply to the output mix of the Android device.

Then we enable measurement mode on the visualizer, by setting the measurement mode to MEASUREMENT_MODE_PEAK_RMS. At the moment, this is the only measurement mode available, but there could feasibly be additional options in the future.

SoundLevelUpdateable is a simple interface, which allows the setting of peak and RMS values:

public interface SoundLevelUpdateable {
    void setPeakValue(int peakValue);
    void setRmsValue(int rmsValue);

This will allow the on-screen values to be updated as new samples are received.

There are 2 additional public methods on the SamplingSoundMeter class – Start() and Stop().

public void Start(final int intervalMillis) {
    // Stop it if we're already running

    // Create a new runnable
    sampler = new Runnable() {
        public void run() {
            handler.postDelayed(sampler, intervalMillis);

    // Enable the visualiser and start the sampler

public void Stop() {

The Start() method creates a new Runnable object which, when run, will call the updateStatus() method. We’ll take a look at this later. This runnable gets also adds itself to the member Handler with a delay, so that the updateStatus() method is called repeatedly – with a time interval provided as an argument to Start().

The Start() method also enables the Visualizer – an action which is mirrored in the Stop() method, which also removes any remaining callbacks on the handler.

The updateStatus() method is as follows:

private void updateStatus() {
    // Capture the sample
    Visualizer.MeasurementPeakRms measurementPeakRms = new Visualizer.MeasurementPeakRms();
    // Notify our updateable

In order to retrieve a measurement from the Visualizer, we must first create an object into which the result can be passed – Visualizer.MeasurementPeakRms. This is then passed into the getMeasurementPeakRms() method, where it is populated with the current values – accessible as mPeak and mRms. These 2 values are passed to the SoundLevelUpdateable object, as defined in the interface.

Using the SamplingSoundMeter

Now that we’ve created this class, it’s pretty simple to wire it into a fragment or activity to get it working.

The following code snippets are based around a fragment which has two buttons (one for start and one for stop), and two text views (one to display the peak value, the other the RMS value).

The fragment should implement the SoundLevelUpdateable interface:

public class SoundMeterFragment extends Fragment implements SoundLevelUpdateable {

Which means there should be implementations of 2 update methods:

public void setPeakValue(int peakValue) {
    ((TextView)getView().findViewById("Peak: " + peakValue);

public void setRmsValue(int rmsValue) {
    ((TextView)getView().findViewById("RMS: " + rmsValue);

In OnCreateView() we create a SamplingSoundMeter, and then wire up the 2 buttons:

private SamplingSoundMeter samplingSoundMeter;

public View onCreateView(LayoutInflater inflater, ViewGroup container,
        Bundle savedInstanceState) {
    View rootView = inflater.inflate(R.layout.fragment_sound_meter, container, false);

    samplingSoundMeter = new SamplingSoundMeter(this);

    // Add some button handlers
    rootView.findViewById( View.OnClickListener() {
        public void onClick(View v) {

    rootView.findViewById( View.OnClickListener() {
        public void onClick(View v) {

    return rootView;

Because of the pulling out of the Visualizer code into a separate class, it means that the fragment is really simple.

Now if you run the app up, you should see something like this:

SoundMeter disabled

Tapping the start button will start the audio meter:

SoundMeter enabled

If you start some audio playback in another app and then come back to SoundOutputMeter then you’ll see the values update as the audio plays:

SoundMeter with sound

To demonstrate that it’s working you can see how the values change as the device volume is changed:

Sound Meter QuietSound Meter Loud


This is a relatively simple addition to the visualizer class, allowing you to perform simple measurements on live audio output sessions, but it can actually make the life of the developer quite a lot easier. Previously these values would have had to be calculated manually using the audio sample interface of the visualizer class, but the measurements API is a lot simpler. Using a similar architecture to the one we’ve created here today will allow you to create visual audio meters without much difficulty at all.

Don’t forget that all the code for today’s SoundOutputMeter project is available on Github at