公开(公告)号：KR100774708B1

公开(公告)日：2007-11-08

申请号：KR1020060066766

申请日：2006-07-18

Applicant: 한국정보통신대학교 산학협력단

Inventor： 김회린 ,  박만수 ,  김성탁 ,  김상호

IPC: H04B1/40 ,  G11B20/10

CPC classification number: G06F17/30743 ,  H04M3/42017

Abstract: A bell sound and/or RBT(Ring Back Tone) generation system of a user preference melody section through real-time music recognition and a method are provided to detect a sound source and an offset position through music recognition if a user inputs a particular melody section of the preferred sound source as an audio query, thereby offering the particular melody section of the sound source as a bell sound and/or an RBT according to the personal preference of a user. An audio query input unit(104) inputs a melody section of particular music selected by a user as an audio query. An audio fingerprint extractor(105) extracts audio fingerprints for a music signal outputted from the input unit(104). A music DB(103) indexes meta data of a lot of music and audio fingerprints to store the indexed data and fingerprints. A music recognition and information searcher(106) recognizes music by matching the extracted audio fingerprints with the stored audio fingerprints, and searches meta information of the recognized music. An offset searcher(107) detects a user's preferred particular melody section from the DB(103) for the recognized music. A bell sound and/or RBT generator(108) generates a bell sound and/or an RBT as the detected particular melody section.

Abstract translation: 提供通过实时音乐识别的用户偏好旋律部分的铃声和/或RBT（回铃音）生成系统，以及如果用户输入特定旋律，通过音乐识别来检测声源和偏移位置 作为音频查询的优选声源的部分，从而根据用户的个人喜好将声源的特定旋律部分提供为铃声和/或RBT。 音频查询输入单元（104）输入由用户选择的特定音乐的旋律部分作为音频查询。 音频指纹提取器（105）提取用于从输入单元（104）输出的音乐信号的音频指纹。 音乐DB（103）索引许多音乐和音频指纹的元数据以存储索引的数据和指纹。 音乐识别和信息搜索器（106）通过将所提取的音频指纹与存储的音频指纹进行匹配来识别音乐，并且搜索识别的音乐的元信息。 偏移搜索器（107）从DB（103）中检测出用于所识别的音乐的用户优选的特定旋律部分。 钟声和/或RBT发生器（108）产生铃声和/或RBT作为检测到的特定旋律部分。