Font Size: a A A

Animation Automatically Generates Systems Research Controlled By Voice Content Features

Posted on:2012-08-24Degree:MasterType:Thesis
Country:ChinaCandidate:W ZhouFull Text:PDF
GTID:2178330332989682Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
In recent years, with the rapid development of multimedia technology and network technique, Internet information resources is becoming more and more, and the main multimedia messages are images and audio informations. Recently, media information is becoming more and more diversified and comprehensived. Network transmission of Informations only contain simple texts, images or sounds in the past times. With the appearation of Flash, voice and images have started to integrate, so audio and visual information would become a trend come. On this basis, audio information into a visual information would become the urgent demand.This text extracts audio feature parameter based on content, achieves a system for real-time animation, including the module of audio feature parameter extraction, animation object selection and animated movement mapping. Users only need to choose music, choose the animation objects and their corresponding movements, can enjoy the music visualizations, experience the hearing scene to the visual scene of the music.The module of audio feature parameter extraction achieves the characteristics of the parameter based on the content. Audio feature parameters can be divided into the time domain , frequency domain and time-frequency domain. Continuous audio signal are sampled to get the sample points. In the extraction of the audio characteristics in time domain. Each sample point includes all the information of the audio signal, so we can extract the audio features directly, don't need to make any further processing. Time domin feature extraction of the audio, contains Temporary Average Energy, Zero-crossing Rate and so on. Generally, voice or music signal contains some environmental noises. It is difficult to separate out the noise in the time domin, but in the frequency domain, it is easy to pick up the main audio information by analyzing the spectrum of signals.Extraction of characteristics in the frequency domain is to achieve fourier changing of audio signal, to analysis different frequency and different value of harmonics composed in the audio signal, to extract the characteristics of harmonics. The characteristics in Frequency domain contains Energy Spectrum, Cepstrum and so on. In real life, some audio signal has a strong feature as time changes. In a period of time, it shows the characteristics of a period signal, but it shows the noise properties at another time. For these audio signals, it can not only analysed in the Time domain or in the Frequency domain. Because fourier changing only considers the global natures, ignores Local characteristics, so the characteristics of the parameter is extracted in the time- frequency domain. This paper in the specific algorithm of the extraction of the audio characteristics, uses VC++ and Matlab mixed programming. Finally, the parameters extracted contains Temporary Average Energy, Temporary Zero-crossing Rate, Temporary Energy Spectrum and Temporary Cepstrum and so on.The module of animation object achieves the construction and choose of the animation object. This paper constructs animation objects using vector and bitmap pictures. Vector charts are drawed by OpenGL. OpenGL has a strong graphic diagram ability, contains functions of constructing objects, starting light, managing bitmap, texture mapping, animation, images enhancement and interaction technique and so on. As graphical hardware and software interfaces, OpenGL mainly achieves the three-dimensional object into a two-dimensional graphic, then to display the processed pixels. Bitmap is made of points, called pixel. These points can be arranged and dyed to different pictures. Images animation have two methods, one is for the pixel operation of the image, realize the image color, position and configuration changes by the individual pixel processing. But this method needs a large calculating, so it will slow down the operation speed, and can not achieve the real-time animation operation. So we adopt a different method—OpenGL image texture textue and map.The experiment results show that this paper can achieve the extraction of the audio parameter characteristics, also control the animation objects using parameters extracted, achieve the animation controlled by the sound characteristics based on content. At tht last , the paper reviewed the work and propose the further studies to explore.
Keywords/Search Tags:content-based, audio, parameters, animation, OpenGL, Visual C++
PDF Full Text Request
Related items