Ȩ : »çÀÌÆ®¸Ê : ¹®ÀǸÞÀÏ : ÀüÀÚÀú³Î
      ¿¬±¸À±¸®À§¿øȸ ±ÔÁ¤
      ÆíÁýÀ§¿øȸ ±ÔÁ¤
      ³í¹®Åõ°í¾È³»/±ÔÁ¤
      ³í¹®ÀÛ¼º¾ç½Ä
      ³í¹®Åõ°í½Åû
      ³í¹®ÀÚ·á½Ç
      ÇÐȸÁö°ü·Ã FAQ
 
 
 
Ȩ > ÇÐȸÁö > ³í¹®ÀÚ·á½Ç
 
Á¦¸ñ A Critical Study on the Accuracy and Reliability of an Automated Analysis on H1-H2 for Korean Phonation Types
ÀúÀÚ Mi-Ryoung Kim
±Ç / È£ 24±Ç / 4È£
Ãâó 103-127
³í¹®°ÔÀçÀÏ 2016.12.31.
ÃÊ·Ï Kim, Mi-Ryoung. (2016). A Critical Study on the Accuracy and Reliability of an Automated Analysis on H1-H2 for Korean Phonation Types. The Linguistic Association of Korea Journal, 24(4), 103-127. This study examined the amplitude difference between the first harmonic and the second harmonic (H1-H2 or spectral tilt) for Korean phonation types (initial stops) and critically evaluated the accuracy and reliability of an automated analysis based on a PRAAT script-based algorithm (Remijsen, 2014). Speech data were collected from seven speakers of native Seoul Korean. The output was compared to those of the manual outcomes. The results showed that, from both the automated and by-hand method, H1-H2 was highest (i.e., positive) for lax stops, higher for aspirated stops, and lowest (i.e., negative) for tense stops, showing a pattern of tense < aspirated < lax stops at vowel onset. However, the automated results neither corresponded to the manual outputs nor were fully reliable. Among individual outcomes, serious errors were observed for some speakers outcomes in that the H1-H2 ranges for the same aspirated stops were -32 dB to 1 dB and the error rate was above 70%. The results indicate that, including Remijsens (2014) Praat script-based algorithm employed in this study, any automated analysis on acoustic parameters cannot fully be reliable but must be always accompanied by hand-corrections to enhance the accuracy and reliability.
÷ºÎ
  6.±è¹Ì·É.pdf
  6.±è¹Ì·É.hwp
 
 
 
 °³ÀÎÁ¤º¸º¸È£Á¤Ã¥ : À̸ÞÀϹ«´Ü¼öÁý°ÅºÎ : »çÀÌÆ®¸Ê : À̸ÞÀϹ®ÀÇÇϱâ