Simon Giddings

Using a seek table increases processing by 10 to 15 times

Cannot build FLAC with cmake

I am creating a config object by calling pConfig = cmd_ln_init(nullptr, ps_args(), TRUE, "-hmm", strModel, "-dict", strDictionary, "-samplerate", "16000", "-topn", strMaxTopGaussians, "-wbeam", m_strWBeam[iBeamIndex], "-beam", m_strBeam[iBeamIndex], "-maxhmmpf", "-1", "-frate", std::to_string(m_ucFrameRate), "-fsgusefiller", "no", "-backtrace", "yes", "-remove_noise", "no", nullptr); However, this appears highly ineficient when doing multiple searches one after the other (10 - 15), as each call will...

I am in the process of migrating pocketsphinx python code over to C++. I am therefor using the pocketsphinx api. However, I do not see how to transform the calls in python over to equivalant C++ code. Here is the heart of what is happening in python : cmd_forced_align = [pocketsphinx_cmd, '-hmm', acmod, '-infile', tmp_user.path("audiofile.wav"), "-jsgf", tmp_user.path("audiofile-forcing.jspf"), "-dict", phoneme_dict, "-backtrace", "yes", "-fsgusefiller", "no", "-wbeam", "{}".format(wbeam[niter]),...

I am in the process of migrating pocketsphinx python code over to C++. I am therefor using the pocketsphinx api. However, I do not see how to transform the calls in python over to equivalant C++ code. Here is the heart of what is happening in python : cmd_forced_align = [pocketsphinx_cmd, '-hmm', acmod, '-infile', tmp_user.path("audiofile.wav"), "-jsgf", tmp_user.path("audiofile-forcing.jspf"), "-dict", phoneme_dict, "-backtrace", "yes", "-fsgusefiller", "no", "-wbeam", "{}".format(wbeam[niter]),...

I am in the process of migrating pocketsphinx python code over to C++. I am therefor using the pocketsphinx api. However, I do not see how to transform the calls in python over to equivalant C++ code. Here is the heart of what is happening in python : !/usr/bin/python cmd_forced_align = [pocketsphinx_cmd, '-hmm', acmod, '-infile', tmp_user.path("audiofile.wav"), "-jsgf", tmp_user.path("audiofile-forcing.jspf"), "-dict", phoneme_dict, "-backtrace", "yes", "-fsgusefiller", "no", "-wbeam", "{}".format(wbeam[niter]),...