自动语音识别(ASR)
Search documents
微信支持潮汕话语音转文字了
经济观察报· 2025-11-16 09:42
Core Viewpoint - WeChat has recently launched support for Chaozhou dialect in its voice-to-text feature, marking it as the second Chinese dialect supported after Cantonese, reflecting the platform's strategy to enhance user engagement among specific high-demand user groups [3][4]. Group 1: WeChat's Development and User Engagement - WeChat's monthly active user count reached 1.414 billion as of Q3 2023, showing significant growth since its inception in 2011 [4]. - The growth rate of WeChat's monthly active users has been declining, with a projected year-on-year growth rate of only 2% by 2025, compared to 0.8% in the previous year [5]. - The introduction of Chaozhou dialect support is part of WeChat's strategy to maintain user engagement as the platform matures [4][5]. Group 2: Chaozhou Dialect and ASR Technology - Chaozhou dialect, a branch of Minnan language, has over 10 million speakers but is classified as a "low-resource language" due to data scarcity and complex tonal variations, making ASR development challenging [4][5]. - The automatic speech recognition (ASR) for Chaozhou dialect took six years to develop, with the feature being launched in November 2025 after initially supporting Mandarin and Cantonese [5]. - Despite the large user base, Chaozhou dialect is not the second largest Chinese dialect by speaker count, as Mandarin, Cantonese, and Wu dialects have higher numbers of native speakers [6].
微信支持潮汕话语音转文字了
Jing Ji Guan Cha Bao· 2025-11-16 09:28
Core Insights - WeChat has launched support for the Chaoshan dialect, marking it as the second Chinese dialect supported after Cantonese, enhancing user engagement among specific high-demand user groups [2][3] User Engagement and Growth - The monthly active user count (MAU) for WeChat and Weixin has reached 1.414 billion as of Q3 2023, reflecting significant growth since its inception in 2011 [2] - The MAU growth rate has been declining, with projections indicating a year-on-year growth of only 2% by 2025, compared to a quarterly growth rate of 0.2% [3] Technological Development - The automatic speech recognition (ASR) for Chaoshan dialect has been in development for six years, with the initial support for Mandarin launched in June 2019, followed by Cantonese in September 2020 [3][4] - Chaoshan dialect is classified as a "low-resource language" in the field of AI and speech recognition due to its complex tonal system and data scarcity, making its ASR development technically challenging [3][4] Demographics and Usage - The Chaoshan dialect has over 10 million speakers, primarily in the three cities of Shantou, Chaozhou, and Jieyang, which have a combined population of 13.835 million as of 2024 [3] - According to Ethnologue, Mandarin has 941 million native speakers, while Cantonese and Wu dialects each have over 80 million speakers, with Min Nan (which includes Chaoshan dialect) having approximately 74.02 million speakers [4]