Summary made by Quivr/GPT-4

This document is a research study that investigates whether the language environment of a pet dog influences its vocalizations. The researchers focused on Shiba Inu dogs living in English and Japanese speaking environments. They collected a large amount of data from YouTube videos, which they used to create a dataset called EJShibaVoice.

The researchers used machine learning models to analyze the data and identify key acoustic features that distinguish the barks of Shiba Inus in the two language environments. They found that the frequency and speed of the barks were the most significant distinguishing factors. Specifically, they found that dogs in English-speaking environments barked at a lower pitch (frequency) but slower (speed) than dogs in Japanese-speaking environments. This mirrors the characteristics of human speech in these two languages, where English is generally spoken at a lower pitch and slower speed than Japanese.

The researchers also conducted a survey with human participants to assess whether people could perceive the differences in dog barks. The results showed that the differences were most noticeable in terms of pitch, which aligns with the acoustic analysis.

The findings of this study suggest that the language environment of a dog may influence its vocalizations. This could have implications for understanding dog behavior and communication. The researchers suggest that future research could expand the dataset to include more breeds and more language environments.

A warning for those expecting a website, this is a pdf download