Using the Microphone capabilities in Adobe AIR 2

　　AIR 2 introduces the ability to manipulate and record sound directly from a microphone or the line-in on the user's sound card. Previously, the developer had to involve the use of a remote server such as the Adobe Flash Media Server to access any of these inputs. AIR 2 exposes the SampleDataEvent.SAMPLE_DATA event for microphone or line-in inputs, which enables you to access the raw data from the sound card as if you were reading it from a file on the file system via the Sound.extract() method. With this raw data you can read this data to add effects, or, in the case of the example discussed here, search for certain audio patterns. Additionally, using the file-system access that the AIR runtime gives you, you can save this raw data to the hard drive or encapsulate it as a WAV file. [b]Setting up and working with the microphone [/b]
　　Up to the release of Adobe AIR 2, just like in Flash 10 and below, developers had no direct access to any of the data coming in from the user's microphone. The only way to have access to this data was to connect the AIR runtime or Flash Player to a remote server, such as Adobe Flash Media Server, to which the microphone data could be sent. AIR 2 introduces the SampleDataEvent.SAMPLE_DATA event so that we can get access to this raw data to manipulate or store it. For example, you can now add simple echo effects, audio gates, or other sound functions to enhance the sound coming from the input. You can attach as many microphone (or line-in) devices as you want. You can also attach to a microphone that is already working with a NetStream.attachMicrophone() output, allowing you to locally store any data that is being sent to the server. The raw data returned is formatted as a ByteArray of single-channel (mono) float value, uncompressed PCM samples. [b]Selecting the microphone input [/b]
　　Before you start to receive data from your microphone input, you will need to select the input from which you want to get the data. You do this by first importing the new flash.media.Microphone class into your project. Next, you will want to display a list of available microphones to end users so they can select the proper input. Unlike Flash Player, AIR will not display the security dialog box where users can select the input; you must do this yourself. The list of available microphones are available in the Microphone singleton, within the names property. This array contains the names (as strings) of the input sources that are returned by the operating system. You will need to know the index number of the microphone that the user selected from this array. The system's default microphone will have an index of 0. [b]Setting up the microphone instance [/b]
　　After you have figured out which microphone the user wants to use, you will need to instantiate a new Microphone object. Set your microphone instance to a copy of the Microphone singleton's getMicrophone() function, passing in the index of the selected microphone (as shown in the following snippet). After a microphone is selected and the object is instantiated, this object will spawn SampleDataEvent.SAMPLE_DATA events that contain the raw PCM data. To start receiving microphone events, you will need to attach an event handler to the SAMPLE_DATA event. To stop receiving the microphone event, you need to remove that event handler. Array; protected var microphone:Microphone; protected var isRecording:Boolean = false; protected function setupMicrophoneList():void { microphoneList = Microphone.names; } protected function setupMicrophone():void { microphone = Microphone.getMicrophone(comboMicList.selectedInde x); } protected function startMicRecording():void { isRecording = true; microphone.addEventListener(SampleDataEvent.SAMPLE _DATA, gotMicData); } protected function stopMicRecording():void { isRecording = false; microphone.removeEventListener(SampleDataEvent.SAM PLE_DATA, gotMicData); } private function gotMicData(micData:SampleDataEvent):void { // micData.data contains a ByteArray with our sample. } ]]>
　　[b]Doing something with the sampled data [/b]
　　When you have the event firing with active data, you will most likely want to do something with it. To finish up this first example, you will simply play back the sound that you are recording. Add the following code to your existing project, and don't forget to add a button to call the playbackData() function. import flash.media.Sound; import flash.utils.ByteArray; protected var soundRecording:ByteArray; protected var soundOutput:Sound; protected function playbackData():void { soundRecording.position = 0; soundOutput = new Sound(); soundOutput.addEventListener(SampleDataEvent.SAMPL E_DATA, playSound); soundOutput.play(); } private function playSound(soundOutput:SampleDataEvent):void { if (!soundRecording.bytesAvailable > 0) return; for (var i:int = 0; i 0) sample = soundRecording.readFloat(); soundOutput.data.writeFloat(sample); soundOutput.data.writeFloat(sample); } }
　　In addition to simply playing back the audio, you can also encapsulate your raw data into a WAV file that you can store to the use's hard drive. With the help of the WAVWriter class included in the samples downloads, you can transcode your recording into a specific bitrate, and encapsulate the data with the proper headers for writing to disk. Since WAV files are just ByteArray streams of PCM data with a header on the front, it is very easy to take your raw data and create these files. To save WAV files in your application, import the com.adobe.audio.format.WAVWriter class and add the following function: protected function saveFile():void { var outputFile:File = File.desktopDirectory.resolvePath("recording.wav") ; var outputStream:FileStream = new FileStream(); var wavWriter:WAVWriter = new WAVWriter(); soundRecording.position = 0; // rewind to the beginning of the sample wavWriter.numOfChannels = 1; // set the inital properties of the Wave Writer wavWriter.sampleBitRate = 16; wavWriter.samplingRate = 44100; wavWriter.processSamples(outputStream, soundRecording, 44100, 1); // convert our ByteArray to a WAV file. outputStream.open(outputFile, FileMode.WRITE); //write out our file to disk. outputStream.close(); }
　　[b]Setting the microphone properties [/b]
　　The Microphone object makes many properties available to help you control the quality and type of data you get back from it. For example, you can set driver-level settings such as the microphone gain, silence level, and rate. If you are familiar with Adobe Connect, these settings are all set via the Audio Wizard and can be fine-tuned to give the user more control of their microphone enviroment. All of these settings can be made via the microphone instance. You may also want to enable the user to mute the microphone. (If you do this, your SampleDataEvent will simply send back ByteArrays full of zeros.) Quality settings can be set via the rate property, which determines how many samples the microphone wll pass back to you per second. You should not change any of these properties after you have added your event listener, or you may have mismatched data in your ByteArray. [b]Doing something with your data [/b]
　　You are not limited to what you can do with your recorded data. Many developers will find that they will want to display a volume meter or a spectrum analyzer of the data that is currently being recorded or displayed. You can use the SoundMixer.ComputeSpectrum, for example, to compute the values needed to display a spectrum analyzer or a sine-wave representation of the current audio. Note, however, that the ComputeSpectrum class will only work with audio that is currently being played to the end user via the Flash Player primary audio mixer. Remember that if you get the float values of the stream, they represent the wave form as integers of -1 to 1 (which can be used to graph the audio wave). [b]Building the DTMF parser [/b]
　　Once you have your audio recorded to a ByteArray, you can take the representation of the audio and search it for patterns. DTMF tones, commonly found on telephones, are a combination of sine waves that are embedded into a sound recording. DTMF tones are made of two sine waves at different frequencies that are overlapped to make the touch-tone noise that we know. In the following example, you will enable the default microphone and records a bit of data to a ByteArray. You will then pass this ByteArray to a function that will attempt to find each of the valid tones. If it finds a combination of the two, it knows what number was pushed. To start and stop the microphone recording, execute the following code: private function startMicRecording():void { micRecording = new ByteArray(); myMicrophone = Microphone.getMicrophone(); myMicrophone.rate = 44; myMicrophone.setLoopBack(enableLoopback); myMicrophone.addEventListener(SampleDataEvent.SAMP LE_DATA, gotMicData); } private function stopMicRecording():void { myMicrophone.removeEventListener(SampleDataEvent.S AMPLE_DATA, gotMicData); }
　　The searchDTMF() function parses the audio using what is known as the Goertzel alrorithm to see if a certain tone exists. You will get back each tone's strength, and rank them. The tones with the strongest representaiton will be looked up and assigned their DTMF values. When you run this example, start the microphone recording. Place a phone (or cellular telephone) next to the microphone and press a number on the keypad. The number represenging the button you pushed should show up in the myDisplay TextInput. If it does not, adjust the dB sensitivity using the slider on the bottom of the app. This article covered the new SampleDataEvent.SAMPLE_DATA functionality that exists in AIR 2. I demonstrated how to enable the microphone, record, and manipulate the results. http://www.adobe.com/devnet/air/flex/articles/usin g_mic_api.html