Previous research by the authors showed that signal compression codecs used in remote meetings and mobile communications have a substantial negative effect on perceived speaker charisma. Moreover, this effect size varied as a function of speaker gender. Following up from this previous study, we conducted a multiparametric acoustic analysis of a set of sentences elicited from male and female speakers in order to detail the effect of speech-signal compression on charisma-related acoustic-prosodic feature settings. Results show that all compression algorithms caused significant acoustic changes compared to the baseline condition. Almost all of them go in an unfavorable direction concerning speaker charisma. The six compression methods also performed differently well. While OPUS and MP3 caused the fewest negative effects, SPEEX and AMRNB resulted in the most negative effects; GSMFR took a middle position. Moreover, evidence is found for gender-specific effects in terms of both the number of negatively affected acoustic features and their type. The results are discussed with respect to their conceptual implications of perceived speaker charisma and the further development of codecs.
%0 Generic
%1 siegert2021speech
%A Siegert, Ingo
%A Niebuhr, Oliver
%D 2021
%K artificial_intelligence audio_codec bias charisma degraded_speech discrimination distant_meetings gender_bias gender_discrimination phonetics prosody sex_differences speech_signal_compression video_conference
%T Speech Signal Compression Deteriorates Acoustic Cues to Perceived Speaker Charisma
%U http://www.essv.de/essv2021/pdfs/06_siegert_v2.pdf
%X Previous research by the authors showed that signal compression codecs used in remote meetings and mobile communications have a substantial negative effect on perceived speaker charisma. Moreover, this effect size varied as a function of speaker gender. Following up from this previous study, we conducted a multiparametric acoustic analysis of a set of sentences elicited from male and female speakers in order to detail the effect of speech-signal compression on charisma-related acoustic-prosodic feature settings. Results show that all compression algorithms caused significant acoustic changes compared to the baseline condition. Almost all of them go in an unfavorable direction concerning speaker charisma. The six compression methods also performed differently well. While OPUS and MP3 caused the fewest negative effects, SPEEX and AMRNB resulted in the most negative effects; GSMFR took a middle position. Moreover, evidence is found for gender-specific effects in terms of both the number of negatively affected acoustic features and their type. The results are discussed with respect to their conceptual implications of perceived speaker charisma and the further development of codecs.
@presentation{siegert2021speech,
abstract = {Previous research by the authors showed that signal compression codecs used in remote meetings and mobile communications have a substantial negative effect on perceived speaker charisma. Moreover, this effect size varied as a function of speaker gender. Following up from this previous study, we conducted a multiparametric acoustic analysis of a set of sentences elicited from male and female speakers in order to detail the effect of speech-signal compression on charisma-related acoustic-prosodic feature settings. Results show that all compression algorithms caused significant acoustic changes compared to the baseline condition. Almost all of them go in an unfavorable direction concerning speaker charisma. The six compression methods also performed differently well. While OPUS and MP3 caused the fewest negative effects, SPEEX and AMRNB resulted in the most negative effects; GSMFR took a middle position. Moreover, evidence is found for gender-specific effects in terms of both the number of negatively affected acoustic features and their type. The results are discussed with respect to their conceptual implications of perceived speaker charisma and the further development of codecs.},
added-at = {2021-05-29T18:35:11.000+0200},
author = {Siegert, Ingo and Niebuhr, Oliver},
biburl = {https://www.bibsonomy.org/bibtex/23b0a4788967344e284a4b98f72aec24e/meneteqel},
eventdate = {2021-03-03},
eventtitle = {32nd Conference on Electronical Speech Signal Processing (ESSV) 2021},
interhash = {f94c28652b390c0fb832d9b50a08c6ea},
intrahash = {3b0a4788967344e284a4b98f72aec24e},
keywords = {artificial_intelligence audio_codec bias charisma degraded_speech discrimination distant_meetings gender_bias gender_discrimination phonetics prosody sex_differences speech_signal_compression video_conference},
language = {en-US},
timestamp = {2021-05-29T18:54:57.000+0200},
title = {Speech Signal Compression Deteriorates Acoustic Cues to Perceived Speaker Charisma},
url = {http://www.essv.de/essv2021/pdfs/06_siegert_v2.pdf},
venue = {Berlin},
year = 2021
}