Presenting the VENEC corpus : Development of a cross-cultural corpus of vocal emotion expressions and a novel method of annotating emotion appraisals
Presenting the VENEC corpus : Development of a cross-cultural corpus of vocal emotion expressions and a novel method of annotating emotion appraisals
复制标题
DOI：
--
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
Petri Laukka;Hillary Anger Elfenbein;Wanda Chui;Nutankumar S. Thingujam;Frederick K. Iraki;Thomas Rockstuhl;Jean Althoff
中科院分区：
文献类型：
--
作者：
Petri Laukka;Hillary Anger Elfenbein;Wanda Chui;Nutankumar S. Thingujam;Frederick K. Iraki;Thomas Rockstuhl;Jean Althoff
我们介绍了跨文化十九种情感的声音表达（VENEC）语料库，并使用一种注释情感评估的新方法呈现了初步评估工作的结果。 VENEC 语料库由来自 5 种英语文化（美国、印度、肯尼亚、新加坡和澳大利亚）的 100 名专业演员组成，他们用声音表达了 19 种不同的情感/情感（喜爱、娱乐、愤怒、蔑视、厌恶、苦恼、恐惧、内疚、快乐、兴趣、欲望、消极惊喜、中立、积极惊喜、自豪、宽慰、悲伤、平静和羞耻），每种情绪都有 3 个级别的情感强度，通过在不同的场景中表现自己引发情绪的情境。总之，该语料库包含大约 6,500 个刺激，由于说话者、文化和情绪强度的影响，为每个情绪类别提供了各种各样的表达风格。所有刺激都进一步进行了音调、强度、语音质量和持续时间线索的声学分析。在评估评级研究中，听众根据情绪引发情境的特征对 VENEC 刺激的选择进行评级，并根据 8 个情绪评估维度（新颖性、内在愉悦性、目标有利性、紧迫性、力量、自我和他人责任以及规范兼容性）进行描述。首先，结果表明，除了责任之外，所有量表的评估者间信度都是可以接受的。其次，对不同声音表达的感知评价概况通常与基于评价理论的预测一致。最后，听众对每个量表的评价与几个声学特征显着相关。结果表明，听众可以从声音情感表达中可靠地推断出情绪引发情况的几个方面，从而表明声音情感表达可能携带认知表征信息。
We introduce the Vocal Expressions of Nineteen Emotions across Cultures (VENEC) corpus and present results from initial evaluation efforts using a novel method of annotating emotion appraisals. The VENEC corpus consists of 100 professional actors from 5 English speaking cultures (USA, India, Kenya, Singapore, and Australia) who vocally expressed 19 different affects/emotions (affection, amusement, anger, contempt, disgust, distress, fear, guilt, happiness, interest, lust, negative surprise, neutral, positive surprise, pride, relief, sadness, serenity, and shame), each with 3 levels of emotion intensity, by enacting finding themselves in various emotion-eliciting situations. In all, the corpus contains approximately 6,500 stimuli offering great variety of expressive styles for each emotion category due to speaker, culture, and emotion intensity effects. All stimuli have further been acoustically analyzed regarding pitch, intensity, voice quality, and durational cues. In the appraisal rating study, listeners rated a selection of VENEC-stimuli with regard to the characteristics of the emotion eliciting situation, described in terms of 8 emotion appraisal dimensions (novelty, intrinsic pleasantness, goal conduciveness, urgency, power, self- and other-responsibility, and norm compatibility). First, results showed that the inter-rater reliability was acceptable for all scales except responsibility. Second, the perceived appraisal profiles for the different vocal expressions were generally in accord with predictions based on appraisal theory. Finally, listeners’ appraisal ratings on each scale were significantly correlated with several acoustic characteristics. The results show that listeners can reliably infer several aspects of emotion-eliciting situations from vocal affect expressions, and thus suggest that vocal affect expressions may carry cognitive representational information.