Copying double array to vector

[phorce]

Hey, I have a function in C that I am trying to convert over to C++. I have done everything, apart from this line:

memcpy(mel[i],&temp[melstart[i]],mellength[i]*sizeof(double));

I need a way of copying the elements in temp[melstart[i]] over to a vector of doubles..

Here is the C code:

void Setup_Mel(int fft_size, int sample_rate) {
        int i,j,k,tap;
        double fmax;
        double dphi;
        double fsample;
        double freq;
        double temp[fft_size/2];

        fmax=2595*log10(8000.0f/700+1);
        dphi = fmax/17;
        freq = (double)sample_rate/fft_size;


        for (i=0; i<16; i++) {
                melstart[i]=fft_size/2;
                mellength[i]=0;
                memset(temp,0,sizeof(double)*fft_size/2);
                for (j=0; j<fft_size/2; j++) {
                        fsample = 2595*log10(freq*j/700 + 1);

                        if ((dphi*i <= fsample) && (fsample < dphi*(i+1))) temp[j] = (fsample-dphi*i)/(dphi*(i+1)-dphi*i);
                        if ((dphi*(i+1) <= fsample) && (fsample < dphi*(i+2))) temp[j] = (fsample-dphi*(i+2))/(dphi*(i+1)-dphi*(i+2));
                        if ((temp[j] != 0) && (melstart[i] > j)) melstart[i] = j;
                        if (temp[j] != 0) mellength[i]++;
                }

                mel[i] = malloc(sizeof(double)*mellength[i]);           
                memcpy(mel[i],&temp[melstart[i]],mellength[i]*sizeof(double));

                printf("%f\n", temp[melstart[i]]);
                for(k=0; (k < mellength[i]); k++)
                {
                    //printf("%f\n", mel[i][k]);
                }
//              for (k=0; k<mellength[i]; tap++,k++) printf("mel filter: %d, %d, %d, %f, %f\n",i,melstart[i]+k,tap,mel[i][k],(melstart[i]+k)*freq);
        }
}

And the C++ function I'm working on:

void setUp_Mel(int fft_size, int sample_rate, vector< vector< double > > &mel, int *melstart, int *mellength)
{
    double fmax;
    double dphi;
    double fsample;
    double freq;
    double temp[fft_size/2];

    fmax = 2595*log10(8000.0f/700+1);
    dphi = fmax / 17;
    freq = (double)sample_rate/fft_size;

      for(int i=0; (i < 16); i++)
      { 
           melstart[i]=fft_size/2;
           mellength[i]=0;
           memset(temp,0,sizeof(double)*fft_size/2);


          for(int j=0; (j < fft_size/2); j++)
          {
              fsample = 2595*log10(freq*j/700 + 1);

              if ((dphi*i <= fsample) && (fsample < dphi*(i+1))) temp[j] = (fsample-dphi*i)/(dphi*(i+1)-dphi*i);
                        if ((dphi*(i+1) <= fsample) && (fsample < dphi*(i+2))) temp[j] = (fsample-dphi*(i+2))/(dphi*(i+1)-dphi*(i+2));
                        if ((temp[j] != 0) && (melstart[i] > j)) melstart[i] = j;
                        if (temp[j] != 0) mellength[i]++;


          }

          mel[i].resize(mellength[i]);

          mel[i].assign(temp[melstart[0]], temp[melstart[0]] + mellength[i]);

          for(int k=0; (k < mellength[i]); k++)
          {
            cout << mel[i][k] << endl;
          }
          //mel[i].assign(temp[melstart[i]], temp[melstart[i]] + mellength[i]);
          //cout << temp[melstart[i]] << endl;
      }
}

This throws up an error. Sorry there is a lot of code, but, the problem should be simple.. I just cannot figure it out. Thanks guys

[mike_2000_17]

Shouldn't it be melstart[i] and not melstart[0]?

Also, you should use the reserve() function instead of resize if you are going to use assign() after, that will avoid the default-construction of all the objects that resize() would trigger. So, I guess this should work:

      mel[i].reserve(mellength[i]);
      mel[i].assign(temp[melstart[i]], temp[melstart[i]] + mellength[i]);

Also, I don't think that you need the temporary array temp. Why not just use mel[i] directly?

[phorce]

Hey thanks for the reply.

No, it didn't work! Gives this error:

MFCCexample.cpp:46: instantiated from here
/usr/include/c++/4.2.1/bits/stl_iterator_base_types.h:129: error: ‘double’ is not a class, struct, or union type

I even tried assigning it directly, I just get a segmentation fault. Any ideas?

[mike_2000_17]

Oh yeah, sorry, this should work:

mel[i].assign(&temp[melstart[i]], &temp[melstart[i]] + mellength[i]);

notice the & to take the address of the values in temp.