Enhancement of Spatial Learning-Memory in Developing Rats via Mozart Music

This paper studies the effect of musical stimulations on the capability of the spatial learningmemory in developing rats by behavioral and electrophysiological techniques. Rats, which are exposed to Mozart's Sonata for Two Pianos in D Major, complete learning tasks of the Moriss water maze with significantly shorter latencies, and the power spectrum of alpha band of electrohippocampogram (EHG) significantly increase, compared with the control rats and rats exposed to the horror music. The results indicate that if given the stimulation of Mozart music in the developmental period of the auditory cortex, the capability of the spatial learning-memory can be significantly changed. The enhancement of alpha band of EHG may be related to the change of this function mainly.     

建筑物中背景音响与消防广播的解决方案

从设计的角度 ,针对建筑物中背景音响与消防广播怎样进行有机的结合 ,使两者很好地统一 ,从而达到既能满足消防广播的要求 ,又能舒适方便地使用背景音响指出了解决方案 ,并对广播的调音及开关控制进行探讨。     

动画声音构成元素剖析

国内对于动画电影声音的理论与实践研究尚处初级阶段。声音对于视听艺术的动画电影而言,有着不可替代的重要作用。研究声音的音乐、音响与语言则是提高动画声音创作的重要基础。     

Learning effectiveness of applying automated music composition software in the high grades of elementary school

The paper is mainly to explore the automated music composition experiences for the students in the high grades of elementary school who are unfamiliar with the professional music theory. The proposed Automated Composition for Music Education (ACME) software was developed to help the students to compose the rhythm and melody themselves, while chord configuration and arrangement are automatically generated. Subsequently, sixth-year students of a non-experimental school were recruited to use the software for the completion of its development. Analysis of learning effectiveness at the reaction level, analysis of learning-level learning effectiveness, correlation analysis of the reaction- and learning-levels learning effectiveness has been performed with a significant result. The correlations between each component of learning effectiveness at the reaction and learning levels were also found. Finally the students in the experimental group using the ACME teaching materials showed a better learning effectiveness at both reaction level and learning level, than those of the students in the comparison group.     

建筑工程司法鉴定实务中的农村自建房鉴定标准化初探

我国农村自建房建设没有严格的监管,质量问题和因此带来的纠纷时有发生。为给农村自建房质量司法鉴定提供依据,在分析农村自建房司法鉴定中的具体问题后,本文从农村自建房结构安全和抗震性能基本要求、检测方法和质量评价原则等方面提出了标准化的具体思路,给下一步制订相关的技术标准提供了技术路线。     

PCA-改进RPROP方法的BP算法在音乐信号分类中的应用

针对标准的BP神经网络仅从预测误差负梯度方向修正权值和阈值,学习过程收敛缓慢,并且容易陷入局部最小值,导致泛化能力不足的问题,提出了一种基于学习经验变学习速率改进的RPROP方法作为BP神经网络权值和阈值更新方法,并与主成分分析法(Principal Component Analysis,PCA)相结合,形成了PCA-改进神经网络算法。同时,采用Matlab软件对四类音乐信号进行分类实验。实验结果表明,改进算法比标准算法的稳定识别率提高2.6%,当稳定识别率达到90%时,用时节省75%,表明该算法可以加快网络的收敛过程,提高泛化能力。     

一种语音与背景乐音信号盲分离优化算法

在盲分离算法中,许多智能优化算法被应用,以克服独立性准则函数的优化进入局部最优位置,但这些优化算法的性能依赖控制参数的选择。因此,提出利用一种单参数的纯随机搜索的单形进化优化算法(Surface-Simplex Swarm Evolution,SSSE),克服算法参数对优化算法性能的影响,提高盲分离算法的应用有效性,并将该改进盲分离算法应用于语音与背景乐音信号的盲分离。实验中,以四阶累积量作为独立分量分析(Independent Component Analysis,ICA)中的准则函数。实验结果表明,该改进算法有效分离出语音成分与背景乐音成分,而且在稳定性和分离效果方面具有较好的性能。     

隐式特征和循环神经网络的多声部音乐生成系统

音乐生成是一种使用算法来生成音乐序列的研究。本文针对音乐样本特征提取以及自动作曲问题提出了一种基于音乐隐式特征和循环神经网络（recurrent neural network, RNN）的多声部音乐生成算法。该方法通过使用栈式自编码器对多声部音乐序列每个时间步的音符隐式特征进行提取,结合长短期记忆循环神经网络（long short-term memory, LSTM）,以序列预测的方式搭建了基于隐式特征的音乐生成模型。仿真结果表明,该音乐生成算法在使用相同风格的音乐数据训练后,得到的模型可以生成旋律与和弦匹配较好的多声部音乐数据。     

蓝牙耳机的主观评测方法

蓝牙无线耳机的使用主要分为音乐和通话两种场景。本文提出了一种主观评测音乐播放效果的方法,结果表明耳机的音乐播放效果都是很不错的。在通话测试部分,按照3GPP和沃达丰的测试规范进行上下行响度的测试,并对噪声消除进行了主观评测,结果表明,语音通话部分还有很大的改进空间。