kinect 传感器 语音识别

[chenrr2305]

各位大神好，c# 新手，这个是语音识别代码，基于微软 XBOX 的，不知道这个波束形成和置信度式子含义，求助！！！！


1、这个kinect麦克风阵列的数据怎么采集进来，是光强还是其他物理形式？？？数据输入口在哪？
这个是置信度模块，具体式子这种到底什么原理？？？vs2010中的代码
private void AudioSourceSoundSourceAngleChanged(object sender, SoundSourceAngleChangedEventArgs e)
{
const double MinGradientWidth = 0.04;

double halfWidth = Math.Max((1 - e.ConfidenceLevel), MinGradientWidth) / 2;

this.sourceGsPre.Offset = Math.Max(this.sourceGsMain.Offset - halfWidth, 0);
this.sourceGsPost.Offset = Math.Min(this.sourceGsMain.Offset + halfWidth, 1);

sourceRotation.Angle = -e.Angle;

sourceAngleText.Text = string.Format(CultureInfo.CurrentCulture, Properties.Resources.SourceAngle, e.Angle.ToString("0", CultureInfo.CurrentCulture));
sourceConfidenceText.Text = string.Format(CultureInfo.CurrentCulture, Properties.Resources.SourceConfidence, e.ConfidenceLevel.ToString("0.00", CultureInfo.CurrentCulture));
}

2.这个是读线程模块，里头式子含义，急求！！！！新手。大神求帮忙
private void AudioReadingThread()
{
const double EnergyNoiseFloor = 0.2;

while (this.reading)
{
int readCount = audioStream.Read(audioBuffer, 0, audioBuffer.Length);
lock (this.energyLock)
{
for (int i = 0; i < readCount; i += 2)
{
short audioSample = BitConverter.ToInt16(audioBuffer, i);
this.accumulatedSquareSum += audioSample * audioSample;
++this.accumulatedSampleCount;

if (this.accumulatedSampleCount < SamplesPerColumn)
{
continue;
double meanSquare = this.accumulatedSquareSum / SamplesPerColumn;
double amplitude = Math.Log(meanSquare) / Math.Log(int.MaxValue);
this.energy[this.energyIndex] = Math.Max(0, amplitude - EnergyNoiseFloor) / (1 - EnergyNoiseFloor);
this.energyIndex = (this.energyIndex + 1) % this.energy.Length;