private HTS_GStreamSet m_GSS; // set of generated parameter streams
//-----------------------------------------------------------
public HTS_Engine()
{
m_Condition = new HTS_Condition();
m_ModelSet = new HTS_ModelSet();
m_Label = new HTS_Label();
m_SSS = new HTS_SStreamSet();
m_PSS = new HTS_PStreamSet();
m_GSS = new HTS_GStreamSet();
Initialize();
}
// HTS_GStreamSet_create: generate speech
public bool Create(HTS_PStreamSet pss, int stage, bool use_log_gain, int sampling_rate, int tFramePeriod, double alpha, double beta, double volume)
{
int i, j, k;
int msd_frame;
int nlpf = 0;
double[] lpf = null;
HTS_Vocoder tVocoder = new HTS_Vocoder();
// check
if (m_GStream != null || m_GSpeech != null)
{
Debug.LogError("HTS_GStreamSet_create: HTS_GStreamSet is not initialized.");
return(false);
}
// initialize
m_NumericOfStream = pss.GetNumericOfStream();
m_TotalFrame = pss.GetTotalFrame();
m_TotalSample = tFramePeriod * m_TotalFrame;
m_GStream = new HTS_GStream[m_NumericOfStream];
for (i = 0; i < m_NumericOfStream; i++)
{
m_GStream[i] = new HTS_GStream();
m_GStream[i].vector_length = pss.GetVectorLength(i);
m_GStream[i].par = new double[m_TotalFrame][];
for (j = 0; j < m_TotalFrame; j++)
{
m_GStream[i].par[j] = new double[m_GStream[i].vector_length];
}
}
m_GSpeech = new double[m_TotalSample];
// copy generated parameter
for (i = 0; i < m_NumericOfStream; i++)
{
if (pss.IsMsd(i) == true)
{
// for MSD
for (j = 0, msd_frame = 0; j < m_TotalFrame; j++)
{
if (pss.GetMsdFlag(i, j) == true)
{
for (k = 0; k < m_GStream[i].vector_length; k++)
{
m_GStream[i].par[j][k] = pss.GetParameter(i, msd_frame, k);
}
msd_frame++;
}
else
{
for (k = 0; k < m_GStream[i].vector_length; k++)
{
m_GStream[i].par[j][k] = HTS_NODATA;
}
}
}
}
else
{
// for non MSD
for (j = 0; j < m_TotalFrame; j++)
{
for (k = 0; k < m_GStream[i].vector_length; k++)
{
m_GStream[i].par[j][k] = pss.GetParameter(i, j, k);
}
}
}
}
// check
if (m_NumericOfStream != 2 && m_NumericOfStream != 3)
{
Debug.LogError("HTS_GStreamSet_create: The number of streams should be 2 or 3.");
Initialize();
return(false);
}
if (pss.GetVectorLength(1) != 1)
{
Debug.LogError("HTS_GStreamSet_create: The size of lf0 static vector should be 1.");
Initialize();
return(false);
}
if (m_NumericOfStream >= 3 && m_GStream[2].vector_length % 2 == 0)
{
Debug.LogError("HTS_GStreamSet_create: The number of low-pass filter coefficient should be odd numbers.");
Initialize();
return(false);
}
Debug.LogWarning("================= Vocoder !!! : " + m_TotalFrame);
// synthesize speech waveform
tVocoder.Prepare(m_GStream[0].vector_length - 1, stage, use_log_gain, sampling_rate, tFramePeriod);
if (m_NumericOfStream >= 3)
{
nlpf = m_GStream[2].vector_length;
}
// トータルフレーム数分ループする
for (i = 0; i < m_TotalFrame; i++)
{
j = i * tFramePeriod;
if (m_NumericOfStream >= 3)
{
// ０じゃなかったらヤバかった
// lpf = this.gstream[ 2 ].par[ i ][ 0 ] ;
lpf = m_GStream[2].par[i];
}
// オフセット補正あり
// HTS_Vocoder_synthesize(&v, gss->gstream[0].vector_length - 1, gss->gstream[1].par[i][0], &gss->gstream[0].par[i][0], nlpf, lpf, alpha, beta, volume, &gss->gspeech[j], audio);
if (i == 0)
{
HTS_Engine.SetDebugFlag(0, 1);
}
tVocoder.Synthesize(m_GStream[0].vector_length - 1, m_GStream[1].par[i][0], m_GStream[0].par[i], nlpf, lpf, alpha, beta, volume, m_GSpeech, j);
if (i == 0)
{
HTS_Engine.SetDebugFlag(0, 0);
}
}
tVocoder.Initialize();
return(true);
}
