|
prints the current best results to the specified stream
This method prints the results of the lipsync (word alignment) to an output stream of choice.
It is printed in a format that can be directly loaded in "The Lipsync Tool" http://www.annosoft.com/demos.htm. You can download the free version to see the results (with annosoft's nice smoothing of the phonemes)
The format of the results see .anno file format.
For the output, since we don't have an energy value, we just use "75" percent. Other readers can ignore this parameter. - Note:
- sapi_lipsync::finalize_phoneme_alignment should be called before calling this method.
- Parameters:
-
| os | - [in] stream to print the results to |
Reimplemented in sapi_textbased_lipsync.
Definition at line 347 of file sapi_lipsync.cpp. 00348 {
00349
00350 std::vector<alignment_result>::iterator p, pEnd;
00351 p = m_results.begin();
00352 pEnd = m_results.end();
00353
00354 if (m_strAudioFile.size())
00355 {
00356 os << "audio " << wstring_2_string(m_strAudioFile) << std::endl;
00357 }
00358 while (p != pEnd)
00359 {
00360
00361 if (p->m_orthography.size())
00362 {
00363 os << "word " << p->m_msStart << ' ' << p->m_msEnd <<
00364 ' ' << wstring_2_string(p->m_orthography) << std::endl;
00365 }
00366
00367 long pos = p->m_msStart;
00368 for (unsigned long j = 0; j < p->m_phonemes.size(); j++)
00369 {
00370 long start = pos;
00371 long end = p->m_phonemeEndTimes[j];
00372 os << "phn " << start << ' ' << end << ' ' << 75 <<
00373 ' ' << wstring_2_string(p->m_phonemes[j]) << std::endl;
00374 pos = end;
00375 }
00376 p++;
00377 }
00378 }
|
