• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.7
• /
• pp.303-314
• /
• 2023

Voice conversion, a technology that allows an individual's speech data to be regenerated with the acoustic properties(tone, cadence, gender) of another, has countless applications in education, communication, and entertainment. This paper proposes an approach based on the StarGAN-VC model that generates realistic-sounding speech without requiring parallel utterances. To overcome the constraints of the existing StarGAN-VC model that utilizes one-hot vectors of original and target speaker information, this paper extracts feature vectors of target speakers using a pre-trained version of Rawnet3. This results in a latent space where voice conversion can be performed without direct speaker-to-speaker mappings, enabling an any-to-any structure. In addition to the loss terms used in the original StarGAN-VC model, Wasserstein distance is used as a loss term to ensure that generated voice segments match the acoustic properties of the target voice. Two Time-Scale Update Rule (TTUR) is also used to facilitate stable training. Experimental results show that the proposed method outperforms previous methods, including the StarGAN-VC network on which it was based.

• Speech Sciences
• /
• v.10 no.1
• /
• pp.71-84
• /
• 2003

In the acoustic modeling for large vocabulary speech recognition, a sparse data problem caused by a huge number of context-dependent (CD) models usually leads the estimated models to being unreliable. In this paper, we develop a new clustering method based on the C45 decision-tree learning algorithm that effectively encapsulates the CD modeling. The proposed scheme essentially constructs a supervised decision rule and applies over the pre-clustered triphones using the C45 algorithm, which is known to effectively search through the attributes of the training instances and extract the attribute that best separates the given examples. In particular, the data driven method is used as a clustering algorithm while its result is used as the learning target of the C45 algorithm. This scheme has been shown to be effective particularly over the database of low unknown-context ratio in terms of recognition performance. For speaker-independent, task-independent continuous speech recognition task, the proposed method reduced the percent accuracy WER by 3.93% compared to the existing rule-based methods.

• Journal of English Language & Literature
• /
• v.60 no.2
• /
• pp.339-363
• /
• 2014

This essay reconsiders the subject of Romantic self-consciousness in a post-de Manian perspective. Self-consciousness is an attribute of Romantic lyricism whereby the poetic speaker I remains conscious of how (s)he feels or lives here and now. This self-reflective feature of Romantic poetry has been controversially interpreted either as self-centered solipsism or as self-expressive objectivism. The question is stirring more disputes among Romantic critics after the advent of New Historicism and Feminism. These two historicistic approaches reprove Romantic poetry for a lack of the sense of history and ascribes it to Romantic self-consciousness. They argue that Romantic poets in general displace historical materiality into an object of internal consciousness, so negating absurd social realities "merely to gain their own immortal soul." This essay targets to overcome this negative stance on Romantic self-consciousness with a "subversive" return to Paul de Man's criticism of Romantic internality.

• Culinary science and hospitality research
• /
• v.19 no.3
• /
• pp.57-72
• /
• 2013

This study carried out an empirical analysis on the long-term effects of a marketing communication strategy used as a means of promotion(advertisement, free gift) of Christmas cake despite its high costs. Such verification results were considered to provide as important core elements for successful management in the bakery industry. By utilizing SPSS and AMOS, exploratory factor analysis(EFA) and confirmatory factor analysis(CFA) were conducted for the validity of the measurement items, and Structural Equation Model(SEM) was performed to verify the hypothesis. The analysis result is as follows. First, advertisement attribute(${\beta}$=0.269, t=4.512, p,<0.001) and free gift attitude(${\beta}$=0.293, t=5.043, p<0.001) had a positive influence on trust. Second, trust had a positive influence on purchase intention(${\beta}$=0.513, t=8.677, p<0.001). Third, advertisement attribute(${\beta}$=0.104, t=1.966, p,<0.05) and free gift attitude(${\beta}$=0.204, t=3.817, p<0.001) had a positive influence on purchase intention. Lastly, advertisement attitude appeared different by advertisement types(actors and animation), and ads with an actor were evaluated high than animation ads. Also, gift attitude appeared different by gift types(speaker doll/character hat and turtleneck muffler), and a turtleneck muffler was evaluated high than a speaker doll or a character hat.