| Additional Discussion
Takehiro Moriya, NTT, made an additional presentation concerning this CE. An analysis of the relative frequency of ACEP frames as a function of bit rate for both the 15 test items and for only the speech and mix of speech and music indicated that the CE could save 2% of bitrate for speech and mix of speech and music content. The audio expert from Samsung (a native Korean speaker) reported that she participated in an informal blind listening test done during the week and verified that the Korean-language speech item “Arirang_speech.wav” was improved.
Roch L, VoiceAge, noted that the CI is larger at NTT. Is this OK to pool? Further, there is no consistent agreement on improved items across test sites. Hence, he does not see strong evidence of performance in this CE.
Dostları ilə paylaş: |
