Switch branches/tags
Nothing to show
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
152 lines (89 sloc) 4.18 KB sense Java Client (Beta version)

Importing sdk into your java project

Currently we do not support maven or gradle repository.

Just download "CochlearSense.jar", which, includes all the sdk classes and other dependencies (such as grpc io, etc..)

Import CochlearSense.jar file into your project, and add it as your project build path.

In case of eclipse java project, [right click on CochlearSense.jar -> Build Path -> Add to Build Path]

Using SDK

predict file (non-streaming)

For non streaming case, you should use CochlearSenseClient class

CochlearSenseClient cs = new CochlearSenseClient("", 9000, "YoUrApiKEyMustBeHERE");

Constructor requires ip addres of our server, port (must be 9000), and your api key. Each of type String, int and String.

You must give file as a Java InputStream object to our sdk object.

String filename = "Wistful Fear.wav";

File file = new File(filename);
InputStream in = new FileInputStream(file);

String res = cs.senseEvent(in, "babycry", "wav");

CochlearSenseClient have following methods

String senseEvent(InputStream is, String subtask, String format)

String senseMusicKey(InputStream is, String format)

String senseMusicTempo(InputStream is, String format)

String senseMusicGenre(InputStream is, String format)

String senseMusicMood(InputStream is, String format)

String senseMusicDetector(InputStream is, String format)

String senseSpeechDetector(InputStream is, String format)

Most of the require InputStream for the file, and format (such as "wav" or "mp3") as an argument. Note that specifically, senseEvent requires 'subtask' String as a second argument.

String res = cs.senseEvent(in, "babycry", "wav");

Simply get a String returned from the method call, and try printing out.

predict streaming

For streaming prediction, you must use CochlearSenseStreamClient class.

CochlearSenseStreamClient cs = new CochlearSenseStreamClient("", 9000, "YoUrApiKEyMustBeHERE");

Predict methods require AudioInputStream java obeject as a single parameter.

We will demonstrate two ways to get AudioInputStream.

AIS from file

String filename = "Wistful Fear.wav";

File file = new File(filename);
AudioInputStream ais= AudioSystem.getAudioInputStream(file);

Precisely speaking, using streaming prediction for file like above, is almost meaningless. As we provide simpler api for file prediction.

AIS from microphone

AudioFormat format = new AudioFormat(44100, 16, 2, true, true);
DataLine.Info tinfo = new DataLine.Info(TargetDataLine.class, format);
TargetDataLine microphone = (TargetDataLine)AudioSystem.getLine(tinfo);;

AudioInputStream ais = new AudioInputStream(microphone);

Now calling predict method and returning streamed response are as follows

cs.senseSpeechDetectorStream(ais); // Send call

for (String result:cs){

Note that CochlearSenseStreamClient itself works as a Python generator(The thing that uses 'yields' statemet). The object itself returns the iterator for the String typed result.

List of the streamed client methods are as follows

CochlearSenseStreamClient senseEventStream(AudioInputStream is, String subtask)

CochlearSenseStreamClient senseAgeGenderStream(AudioInputStream is)

CochlearSenseStreamClient senseMusicGenreStream(AudioInputStream is)

CochlearSenseStreamClient senseMusicMoodStream(AudioInputStream is)

CochlearSenseStreamClient senseMusicDetectorStream(AudioInputStream is)

CochlearSenseStreamClient senseSpeechDetectorStream(AudioInputStream is)

Also note that senseEventStream method also requires additional string parameter 'subtask'

Java builder like usage of streaming client

CochlearSenseStreamClient cs = new CochlearSenseStreamClient("", 9000, apiKey)

for(String out : cs) {

Streaming client supports builder code as above.


This sdk was compiled and tested under Java 1.8 version. Precedding Java versions (such as Java 1.7) may not be supportive.