compute-kaldi-pitch-feats.cc File Reference

#include "base/kaldi-common.h"
#include "util/common-utils.h"
#include "feat/pitch-functions.h"
#include "feat/wave-reader.h"

Include dependency graph for compute-kaldi-pitch-feats.cc:

Go to the source code of this file.

Functions
int	main (int argc, char *argv[])

Function Documentation

main()

int main	(	int	argc,
		char *	argv[]
	)

Definition at line 27 of file compute-kaldi-pitch-feats.cc.

References kaldi::ComputeKaldiPitch(), WaveData::Data(), SequentialTableReader< Holder >::Done(), ParseOptions::GetArg(), KALDI_ASSERT, KALDI_ERR, KALDI_LOG, KALDI_VLOG, KALDI_WARN, SequentialTableReader< Holder >::Key(), SequentialTableReader< Holder >::Next(), ParseOptions::NumArgs(), MatrixBase< Real >::NumRows(), ParseOptions::PrintUsage(), ParseOptions::Read(), PitchExtractionOptions::Register(), PitchExtractionOptions::samp_freq, WaveData::SampFreq(), SequentialTableReader< Holder >::Value(), and TableWriter< Holder >::Write().

                                 {
  try {
    using namespace kaldi;
    const char *usage =
        "Apply Kaldi pitch extractor, starting from wav input.  Output is 2-dimensional\n"
        "features consisting of (NCCF, pitch in Hz), where NCCF is between -1 and 1, and\n"
        "higher for voiced frames.  You will typically pipe this into\n"
        "process-kaldi-pitch-feats.\n"
        "Usage: compute-kaldi-pitch-feats [options...] <wav-rspecifier> <feats-wspecifier>\n"
        "e.g.\n"
        "compute-kaldi-pitch-feats --sample-frequency=8000 scp:wav.scp ark:- \n"
        "\n"
        "See also: process-kaldi-pitch-feats, compute-and-process-kaldi-pitch-feats\n";

    ParseOptions po(usage);
    PitchExtractionOptions pitch_opts;
    int32 channel = -1; // Note: this isn't configurable because it's not a very
                        // good idea to control it this way: better to extract the
                        // on the command line (in the .scp file) using sox or
                        // similar.

    pitch_opts.Register(&po);

    po.Read(argc, argv);

    if (po.NumArgs() != 2) {
      po.PrintUsage();
      exit(1);
    }

    std::string wav_rspecifier = po.GetArg(1),
        feat_wspecifier = po.GetArg(2);

    SequentialTableReader<WaveHolder> wav_reader(wav_rspecifier);
    BaseFloatMatrixWriter feat_writer(feat_wspecifier);

    int32 num_done = 0, num_err = 0;
    for (; !wav_reader.Done(); wav_reader.Next()) {
      std::string utt = wav_reader.Key();
      const WaveData &wave_data = wav_reader.Value();

      int32 num_chan = wave_data.Data().NumRows(), this_chan = channel;
      {
        KALDI_ASSERT(num_chan > 0);
        // reading code if no channels.
        if (channel == -1) {
          this_chan = 0;
          if (num_chan != 1)
            KALDI_WARN << "Channel not specified but you have data with "
                       << num_chan  << " channels; defaulting to zero";
        } else {
          if (this_chan >= num_chan) {
            KALDI_WARN << "File with id " << utt << " has "
                       << num_chan << " channels but you specified channel "
                       << channel << ", producing no output.";
            continue;
          }
        }
      }

      if (pitch_opts.samp_freq != wave_data.SampFreq())
        KALDI_ERR << "Sample frequency mismatch: you specified "
                  << pitch_opts.samp_freq << " but data has "
                  << wave_data.SampFreq() << " (use --sample-frequency "
                  << "option).  Utterance is " << utt;


      SubVector<BaseFloat> waveform(wave_data.Data(), this_chan);
      Matrix<BaseFloat> features;
      try {
        ComputeKaldiPitch(pitch_opts, waveform, &features);
      } catch (...) {
        KALDI_WARN << "Failed to compute pitch for utterance "
                   << utt;
        num_err++;
        continue;
      }

      feat_writer.Write(utt, features);
      if (num_done % 50 == 0 && num_done != 0)
        KALDI_VLOG(2) << "Processed " << num_done << " utterances";
      num_done++;
    }
    KALDI_LOG << "Done " << num_done << " utterances, " << num_err
              << " with errors.";
    return (num_done != 0 ? 0 : 1);
  } catch(const std::exception &e) {
    std::cerr << e.what();
    return -1;
  }
}