corr_02.cc

这是一个从音频信号里提取特征参量的程序

// file: $isip/class/algo/Correlation/corr_02.cc
// version: $Id: corr_02.cc,v 1.20 2002/04/13 21:22:33 huang Exp $
//

// isip include files
//
#include "Correlation.h"

// method: diagnose
//
// arguments:
//  Integral::DEBUG level: (input) debug level for diagnostics
//
// return: a boolean value indicating status
//
boolean Correlation::diagnose(Integral::DEBUG level_a) {

  //---------------------------------------------------------------------------
  //
  // 0. preliminaries
  //
  //---------------------------------------------------------------------------

  // output the class name
  //
  if (level_a > Integral::NONE) {
    SysString output(L"diagnosing class ");
    output.concat(CLASS_NAME);
    output.concat(L": ");
    Console::put(output);
    Console::increaseIndention();
  }

  //--------------------------------------------------------------------------
  //
  // 1. required public methods
  //     class constructors
  //
  //--------------------------------------------------------------------------

  // set indentation
  //  
  if (level_a > Integral::NONE) {
    Console::put(L"testing required public methods: class constructors...\n");
    Console::increaseIndention();
  }
  
  Correlation cor0;
  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(FACTORED);
  cor0.setNormalization(NONE);
  Correlation cor1(cor0);
  
  if (!cor1.eq(cor0)) {
    return Error::handle(name(), L"copy constructor/eq", Error::TEST,
			 __FILE__, __LINE__);
  }

  // test large allocation construction and deletion
  //
  if (level_a == Integral::ALL) {
    
    Console::put(L"\ntesting large chunk memory allocation and deletion:\n");
    
    // set the memory to a strange block size so we can hopefully catch any
    // frame overrun errors
    //
    
    Correlation::setGrowSize((long)500);
    
    Correlation* pcor = new Correlation();

    for (long j = 1; j <= 100; j++) {
      Correlation** pcors = new Correlation*[j * 100];
      
      // create the objects
      //
    for (long i = 0; i < j * 100; i++) {
      pcors[i] = new Correlation();
    }
    
    // delete objects
    //
    for (long i = (j * 100) - 1; i >= 0; i--) {
      delete pcors[i];
    }
    
    delete [] pcors;
    } 
    
    delete pcor;
  }

  // reset indentation
  //
  if (level_a > Integral::NONE) {
    Console::decreaseIndention();
  }
  
  //---------------------------------------------------------------------------
  //
  // 2. required public methods
  //     i/o methods 
  //
  //---------------------------------------------------------------------------

  // set indentation
  //  
  if (level_a > Integral::NONE) {
    Console::put(L"testing required public methods: io methods...\n");
    Console::increaseIndention();
  }

  // declare a reference object
  //
  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(FACTORED);
  cor0.setNormalization(NONE);

  // we need binary and text sof files
  //
  String tmp_filename0;
  Integral::makeTemp(tmp_filename0);
  String tmp_filename1;
  Integral::makeTemp(tmp_filename1);

  // open files in write mode
  //
  Sof tmp_file0;
  tmp_file0.open(tmp_filename0, File::WRITE_ONLY, File::TEXT);
  Sof tmp_file1;
  tmp_file1.open(tmp_filename1, File::WRITE_ONLY, File::BINARY);

  cor0.write(tmp_file0, (long)0);
  cor0.write(tmp_file1, (long)0);
  
  // close the files
  //
  tmp_file0.close();
  tmp_file1.close();

  // open the files in read mode
  //
  tmp_file0.open(tmp_filename0);
  tmp_file1.open(tmp_filename1);

  // read the object back
  //
  cor1.read(tmp_file0, (long)0);
  cor1.init();

  if (!cor0.eq(cor1)) {
    return Error::handle(name(), L"i/o", Error::TEST,
			 __FILE__, __LINE__);
  }
    
  cor1.read(tmp_file1, (long)0);
  cor1.init();

  if (!cor0.eq(cor1)) {
    return Error::handle(name(), L"i/o", Error::TEST,
			 __FILE__, __LINE__);
  }
    
  // close and delete the temporary files
  //
  tmp_file0.close();
  tmp_file1.close();
  File::remove(tmp_filename0);
  File::remove(tmp_filename1);

  // reset indentation
  //
  if (level_a > Integral::NONE) {
    Console::decreaseIndention();
  }
  
  //---------------------------------------------------------------------------
  //
  // 3. class-specific public methods
  //     set and get methods
  //
  //---------------------------------------------------------------------------

  // set indentation
  //  
  if (level_a > Integral::NONE) {
    Console::put(L"testing class-specific public methods: set and get methods...\n");
    Console::increaseIndention();
  }

  // set the parameters manually
  //
  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(FACTORED);
  cor0.setNormalization(LENGTH);
  cor0.setOrder(27);

  // check that the values were set
  //
  if (cor0.algorithm_d != AUTO) {
    return Error::handle(name(), L"setAlgorithm", Error::TEST,
			 __FILE__, __LINE__);
  }
  else if (cor0.implementation_d != FACTORED) {
    return Error::handle(name(), L"setImplementation", Error::TEST,
			 __FILE__, __LINE__);
  }
  else if (cor0.normalization_d != LENGTH) {
    return Error::handle(name(), L"setNormalization", Error::TEST,
			 __FILE__, __LINE__);
  }
  else if (cor0.order_d != (Long)27) {
    return Error::handle(name(), L"setOrder", Error::TEST,
			 __FILE__, __LINE__);
  }

  // reset indentation
  //
  if (level_a > Integral::NONE) {
    Console::decreaseIndention();
  }
  
  //---------------------------------------------------------------------------
  //
  // 4. class-specific public methods
  //     computational methods
  //
  //---------------------------------------------------------------------------

  // set indentation
  //
  if (level_a > Integral::NONE) {
    Console::put(L"testing class-specific public methods: computational methods...\n");
    Console::increaseIndention();
  }

  // establish some signals
  //
  VectorFloat input(L"0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15");
  VectorFloat output;
  VectorFloat zeros(L"0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0");
  VectorFloat ones(L"1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1");
  VectorFloat ans;

  // algorithm: AUTO
  //  check that FACTORED and UNFACTORED produce the same result
  //
  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(UNFACTORED);
  cor0.setNormalization(NONE);
  cor0.setOrder(4);
  cor0.compute(ans, input);

  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(FACTORED);
  cor0.setNormalization(NONE);
  cor0.setOrder(4);
  cor0.compute(output, input);

  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(),
  			  L"compute (AUTO) from sampled data",
  			  ERR, __FILE__, __LINE__);
  }
  
  // algorithm: AUTO_TAPERED
  //  check that FACTORED and UNFACTORED produce the same result
  //
  cor0.setAlgorithm(AUTO_TAPERED);
  cor0.setImplementation(UNFACTORED);
  cor0.setNormalization(NONE);
  cor0.setOrder(4);
  cor0.compute(ans, input);

  cor0.setAlgorithm(AUTO_TAPERED);
  cor0.setImplementation(FACTORED);
  cor0.setNormalization(NONE);
  cor0.setOrder(4);
  cor0.compute(output, input);

  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(),
			  L"compute (AUTO_TAPERED) from sampled data",
			  ERR, __FILE__, __LINE__);
  }

  // check that normalization works
  //
  VectorFloat input_norm(input);
  input_norm.mult(1.0 / input_norm.rms());
  ans.assign(L"1.0000, 0.9070, 0.8140, 0.7210, 0.6281");

  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(UNFACTORED);
  cor0.setNormalization(UNIT_ENERGY);
  cor0.setOrder(4);
  cor0.compute(output, input_norm);

  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(),
			  L"compute (AUTO) with normalization",
			  ERR, __FILE__, __LINE__);
  }

  ans.assign(L"1.31828, 1.1957, 1.07312, 0.950538, 0.827957");
  cor0.setNormalization(LENGTH);
  cor0.setOrder(4);
  cor0.compute(output, input_norm);

  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(),
			  L"compute (AUTO) with normalization",
			  ERR, __FILE__, __LINE__);
  }

  // compute a correlation function using predictor coefficients
  //
  input.assign(L"1.0, -2.34644, 1.65697,-0.00599, 0.32305, -1.48213, 1.15463, -0.18966, -0.05899");

  cor0.setOrder(8);
  cor0.setAlgorithm(AUTO);
  cor0.compute(output, input, AlgorithmData::PREDICTION);
  
  // compute using the corresponding reflection coefficients
  //
  input.assign(L"-0.94217, 0.92386, -0.56198, -0.09454, 0.20218, 0.53595, -0.32922, -0.05899");
  VectorFloat tmp;

  cor0.compute(tmp, input, AlgorithmData::REFLECTION);

  if (!output.almostEqual(tmp, 0.3)) {
    tmp.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(),
			  L"compute (AUTO) from prediction coefficients",
			  Error::TEST, __FILE__, __LINE__);
  }

  // test for degenerate cases
  //  
  ans.assign(0.0);
  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(FACTORED);
  cor0.setOrder(4);
  cor0.compute(output, zeros);
  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(), L"compute for a zero-valued signal",
			  Error::TEST, __FILE__, __LINE__);
  }
  
  cor0.setImplementation(UNFACTORED);
  cor0.compute(output, zeros);
  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(), L"compute (AUTO) for a zero-valued signal",
			  Error::TEST, __FILE__, __LINE__);
  }
  
  ans.assign(L"11.0, 11.0, 11.0, 11.0, 11.0");
  cor0.setAlgorithm(AUTO);
  cor0.setImplementation(UNFACTORED);
  cor0.setNormalization(NONE);
  cor0.compute(output, ones);
  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(), L"compute (AUTO) for a one-valued signal",
			  Error::TEST, __FILE__, __LINE__);
  }

  cor0.setImplementation(FACTORED);
  cor0.setOrder(4);
  cor0.compute(output, ones);

  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(), L"compute (AUTO) for a one-valued signal",
			  Error::TEST, __FILE__, __LINE__);
  }
  
  // check the CROSS algorithm
  //
  input.assign(L"1, 2, 3, 2, 1");
  VectorFloat ref(L"1, 2");
  ans.assign(L"0.512989, 0.820783, 0.718185, 0.410391, 0.102598");

  cor0.setAlgorithm(CROSS);
  cor0.setImplementation(UNFACTORED);
  cor0.setNormalization(UNIT_ENERGY);
  cor0.compute(output, input, ref);
  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return Error::handle(name(), L"compute (CROSS)",
			 ERR, __FILE__, __LINE__);
  }

  // check the CONV algorithm
  //
  cor0.setAlgorithm(CONV);
  ans.assign(L"1.0, 4.0, 7.0, 8.0, 5.0");

  cor0.setAlgorithm(CONV);
  cor0.setImplementation(UNFACTORED);
  cor0.setNormalization(NONE);
  cor0.compute(output, input, ref);
  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return Error::handle(name(), L"compute (CONV)",
			 ERR, __FILE__, __LINE__);
  }

  // test for degenerate cases
  //  
  ans.setLength(12);
  ans.assign(0.0);
  cor0.setAlgorithm(CROSS);
  cor0.setImplementation(UNFACTORED);
  cor0.setNormalization(UNIT_ENERGY);
  cor0.compute(output, zeros, ref);

  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(), L"compute (CROSS) for a zero-valued signal",
			  Error::TEST, __FILE__, __LINE__);
  }
  
  cor0.setAlgorithm(CONV);
  cor0.setNormalization(LENGTH);
  cor0.compute(output, zeros, ref);
  
  if (!output.almostEqual(ans)) {
    ans.debug(L"correct answer");
    output.debug(L"wrong answer");
    return  Error::handle(name(), L"compute (CONV) for a zero-valued signal",
			  Error::TEST, __FILE__, __LINE__);
  }
  
  // reset indentation
  //
  if (level_a > Integral::NONE) {
    Console::decreaseIndention();
  }

  //---------------------------------------------------------------------------
  //
  // 5. print completion message
  //
  //---------------------------------------------------------------------------

  // reset indentation
  //
  if (level_a > Integral::NONE) {
    Console::decreaseIndention();
  }

  // possibly print completion message
  //  
  if (level_a > Integral::NONE) {
    String output(L"diagnostics completed successfully for class ");
    output.concat(name());
    output.concat(L"\n");
    Console::put(output);
  }

  // exit gracefully
  //
  return true;
}