<< Chapter < Page Chapter >> Page >

Determine the frequency to represent a y-value of zero

Listing 14 determines the frequency that will represent a y-value of 0. If the data are all positive, the data values will be biased so that the lowestvalue will sound at 220 Hz and the highest value will sound at 1760 Hz. Beyond that, the code in Listing 14 shouldn't require an explanation.

Listing 14 . Determine the frequency to represent a y-value of zero.
double zeroFreq; if(lowData>= 0.0){ //Bias all data so that the lowest value will display at 220Hz. Skip the// first three data values which are always zero. for(int cnt = 3;cnt<inputData.length;cnt++){ inputData[cnt]-= lowData; }//end for loop//zeroFreq has little meaning in this case because it is probably off the// bottom of the page. Set it to the bottom of the range. zeroFreq = lowFreq;}else{ //Set zeroFreq to a value that represents a value of zero for bipolar// data. zeroFreq = lowFreq +(Math.abs(lowData)/dataRange) * freqRange;}//end else System.out.println("zeroFreq = " + zeroFreq);

Compute audio sample values and deposit them in the output array

Listing 15 computes audio sample values and deposits them in the output array.

You have seen code similar to this in previous modules. The only thing that is new and different here is the algebra that is used to map the sample valuesto frequencies and the algebra that is used to cause the sound to progress from the left speaker to the right speaker in proportion to the value of x currentlybeing evaluated.

Note that the sound that is produced for a y-value of zero is different from the sound that is produced for all other values. In the case of a y-value of zero, the soundthat is produced is the weighted sum of three sounds covering two octaves centered on the frequency that is considered to be the zero frequency. Thismakes it possible to audibly identify those points where the function being graphed crosses the horizontal or zero axis (provided that there is actually a data value of zero) .

Beyond that, the embedded comments should be sufficient to explain the code in Listing 15 .

Listing 15 . Compute audio sample values and deposit them in the output array.
int sampLength = melody.length/audioParams.channels/bytesPerSampPerChan; for(int cnt = 0; cnt<sampLength; cnt++){ //Compute the time in seconds for this sample.double time = cnt/audioParams.sampleRate;double yValue = 0; if(cnt%pulseLengthInSamples == 0){//It is time for a new pulse. Get the next y value from the data array // and use it to compute the frequency of the next pulse.yValue = inputData[cnt/pulseLengthInSamples];freq = zeroFreq + (yValue/dataRange)*freqRange; }//end if//Deposit audio data in the melody array for each channel. Shape the // amplitude of each pulse with a triangular scale factor (rooftop shape)// to minimize the undesirable pops and clicks that occur when there // are abrupt change in the frequency from one pulse to the next. The// following gain factor ranges from 0.0 at the ends to maximum in the // center of the pulse.gain = (cnt%pulseLengthInSamples)/(double)pulseLengthInSamples; if(gain>0.5){ //Change to a negative slope.gain = (pulseLengthInSamples - cnt%pulseLengthInSamples)/(double)pulseLengthInSamples;}//end if//Set the final gain to a value that is compatible with 16-bit audio // data.gain = 8000*gain; //Cause the sound to progress from the left speaker to the right speaker// in proportion to the value of x. rightGain = gain * ((double)cnt/sampLength);leftGain = gain - rightGain; if(freq == zeroFreq){//Compute scaled pulse values and deposit them into the melody. Mark // the zeroFreq by adding frequency components at a reduced level one// octave above and one octave below the zeroFreq. This will make it // sound special in the output.byteBuffer.putShort((short)(leftGain*Math.sin(2*Math.PI*freq*time) + (leftGain*Math.sin(2*2*Math.PI*freq*time))/3 +(leftGain*Math.sin(2*Math.PI*freq*time/2))/3)); byteBuffer.putShort((short)(rightGain*Math.sin(2*Math.PI*freq*time) +(rightGain*Math.sin(2*2*Math.PI*freq*time))/3 + (rightGain*Math.sin(2*Math.PI*freq*time/2))/3));}else{ //Compute scaled pulse values and deposit them into the melody.byteBuffer.putShort((short)(leftGain*Math.sin(2*Math.PI*freq*time))); byteBuffer.putShort((short)(rightGain*Math.sin(2*Math.PI*freq*time)));}//end else }//end for loopreturn melody;}//end method getMelody //-------------------------------------------------------------------------//}//end class AudioGraph01

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Accessible objected-oriented programming concepts for blind students using java. OpenStax CNX. Sep 01, 2014 Download for free at https://legacy.cnx.org/content/col11349/1.17
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Accessible objected-oriented programming concepts for blind students using java' conversation and receive update notifications?

Ask