技术领域technical field
本发明涉及音乐声学、心理声学、信号处理、计算机模式识别等交叉技术领域，特别涉及一种结合曲风的钢琴演奏音质评价系统及方法。The invention relates to the interdisciplinary technical fields of music acoustics, psychoacoustics, signal processing, computer pattern recognition, etc., and in particular relates to a piano performance sound quality evaluation system and method combined with music styles.
背景技术Background technique
随着经济、社会的不断发展，人们的物质生活水平得到了迅速提高，精神文化方面受到越来越多的关注和重视，需求日益增大。其中，音乐艺术是人们日常生活中接触得最多的方面，而钢琴作为“乐器之王”占据了演奏音乐艺术的很大一部分比重。由于钢琴结构复杂、音域广阔，对其音质、音色的研究和评价难度较大。目前，钢琴的音质好坏基本上依赖于主观判断，即由具备一定钢琴乐理知识和演奏经验的专家进行听音，并根据一些指标进行评分。这种评价方法会因为测试者的听觉疲劳、个人偏好以及现场听音环境的不同，导致评价准确度波动等问题。此外，在很多实时的情况下，听音测试的人力成本和时间成本也相对昂贵。With the continuous development of economy and society, people's material living standards have been rapidly improved, and spiritual and cultural aspects have received more and more attention and attention, and the demand has increased day by day. Among them, the art of music is the most contacted aspect in people's daily life, and the piano, as the "king of musical instruments", occupies a large proportion of the art of playing music. Due to the complex structure and wide range of the piano, it is difficult to study and evaluate its sound quality and timbre. At present, the sound quality of a piano basically depends on subjective judgment, that is, experts with certain knowledge of piano music theory and performance experience listen to the sound and score it according to some indicators. This evaluation method will lead to problems such as fluctuations in evaluation accuracy due to the hearing fatigue of the testers, personal preferences, and differences in the on-site listening environment. In addition, in many real-time situations, the labor and time costs of listening tests are relatively expensive.
随着计算机科学的发展以及信号处理技术与音乐声学之间更多的结合研究，人们开始从理性和科学的角度来分析乐器各方面的特征。应用小波分析等数字信号信号处理技术或神经网络等机器学习知识可实现乐器演奏音信号的音乐信号特征的精确提取，而以往的研究多是使用这些特征进行乐器识别、演奏正误评价等方面的应用，或是直接分析这些特征进行简单的音质评价，无法很好地实现自动化地智能地进行音质评价。若可以将主观先验知识和客观信号特征分析结合，找到两者之间的映射关系，再利用更接近人类认知和判断机制的神经网络、模糊推理等技术来设计一个智能评价系统，最终只需要给计算机输入一段钢琴演奏音信号，就可以节省人力成本，无需人工对钢琴演奏音质进行评价。With the development of computer science and more combined research between signal processing technology and music acoustics, people began to analyze the characteristics of various aspects of musical instruments from a rational and scientific perspective. The application of digital signal processing technology such as wavelet analysis or machine learning knowledge such as neural network can realize the accurate extraction of musical signal features of musical instrument performance sound signals, while previous studies mostly use these features for musical instrument recognition and performance evaluation. , or directly analyze these features for simple sound quality evaluation, it is impossible to realize automatic and intelligent sound quality evaluation. If it is possible to combine subjective prior knowledge with objective signal feature analysis, find the mapping relationship between the two, and then design an intelligent evaluation system using neural network, fuzzy reasoning and other technologies that are closer to human cognition and judgment mechanisms, the final result is only It is necessary to input a piece of piano performance sound signal to the computer, which can save labor costs, and does not need to manually evaluate the sound quality of piano performance.
在音乐信号特征提取和分析方面，钢琴音乐主观上的先验知识与客观上的数字信号特征之间的关系得到了越来越多的关注和研究。在之前的研究中，大多是基于内容的音乐识别的研究，且多是使用单个麦克风收集钢琴演奏音的信号，而结合麦克风阵列的研究中，多是收集空间场的信号，应用于空间场的声场模拟和声场重现，很少利用麦克风阵列在不同位置处接收钢琴演奏音的信号，即空域信号的测量和采集，而空域信号与整体听觉感受息息相关，同时在音板特性对钢琴音质的影响的研究中，空域信号的空域特征与音板特性也存在紧密关系。因此，考虑采集多个位置的空域信号并综合分析，就能更全面评估一架钢琴的音质。In terms of music signal feature extraction and analysis, the relationship between subjective prior knowledge of piano music and objective digital signal features has received more and more attention and research. In the previous studies, most of them were content-based music recognition studies, and most of them used a single microphone to collect the signal of the piano performance sound, while in the research combined with the microphone array, most of them were to collect the signal of the space field and apply it to the space field. Sound field simulation and sound field reproduction, seldom use the microphone array to receive the signal of the piano playing sound at different positions, that is, the measurement and collection of the spatial signal, and the spatial signal is closely related to the overall auditory experience. In the research of , there is also a close relationship between the spatial characteristics of the spatial signal and the characteristics of the soundboard. Therefore, it is possible to more comprehensively evaluate the sound quality of a piano by considering the collection of airspace signals from multiple locations and analyzing them comprehensively.
而关于钢琴音质的研究，之前的研究在分析音乐数据时大多仅使用了单音符、双音符或简单的没有曲风标签的演奏样本，演奏者主观上的偏好即曲风这一因素没有考虑到其中。而如今社会环境下，个性化需求受到钢琴厂家和销售商更多的重视，因为每个钢琴爱好者都有自己的偏好，希望能找到最符合自己习惯和喜好的钢琴。而无论是初学者还是有经验的专家，演奏者在同一架钢琴上演奏不同风格的钢琴曲时，钢琴表现出来的音质也不相同。因此，评价一架钢琴的音质好坏，应该考虑曲风这一具备个性化特性的因素。As for the research on piano sound quality, most of the previous studies only used single-note, double-note or simple performance samples without genre labels when analyzing music data, and the player’s subjective preference, namely the genre, was not considered. in. However, in today's social environment, individual needs are paid more attention by piano manufacturers and sellers, because every piano lover has his own preferences, hoping to find a piano that best suits his habits and preferences. Regardless of whether it is a beginner or an experienced expert, when a player plays different styles of piano music on the same piano, the sound quality of the piano will be different. Therefore, to evaluate the sound quality of a piano, one should consider the individual characteristic of music style.
发明内容Contents of the invention
本发明的目的在于克服钢琴音质评价时依赖于主观评价以及没有考虑曲风导致分析不够全面的缺陷，并解决仅利用了单个麦克风进行数据采集而缺少对钢琴演奏音信号的空域特征分析的问题，提出一种结合曲风的钢琴演奏音质评价系统， 并采用麦克风阵列录音装置结合高质量麦克风录音装置的演奏音信号采集装置。同时，本发明采用模糊神经网络构建音质评价模块，当接收到用户在一台钢琴上演奏特定风格曲子的演奏音信号后，通过分析这些数据在域、时域、频域、时频图的相应特征，并将得到的特征向量和曲风标签作为模糊神经网络的输入，最终输出一个分数，对所弹钢琴在选定曲风时进行音质评价。The purpose of the present invention is to overcome the defect that the evaluation of piano sound quality relies on subjective evaluation and does not consider the style of music, which leads to insufficient analysis, and solves the problem that only a single microphone is used for data collection and lacks the spatial feature analysis of the piano performance sound signal, A piano performance sound quality evaluation system combined with music style is proposed, and a performance sound signal acquisition device combined with a microphone array recording device and a high-quality microphone recording device is used. At the same time, the present invention uses a fuzzy neural network to build a sound quality evaluation module. After receiving the performance sound signal of a specific style of music played by a user on a piano, the data is analyzed in the domain, time domain, frequency domain, and time-frequency diagram. Features, and the obtained feature vector and style label are used as the input of the fuzzy neural network, and finally a score is output to evaluate the sound quality of the piano played when the style is selected.
为了实现上述目的和功能，本发明提出的系统在使用之前，需要先经过系统建立的过程，即采集不同钢琴上不同曲风的演奏音信号，收集专家主观评价数据并提取和分析信号特征，确定模糊神经网络的结构并进行训练；训练好的模糊神经网络即可使用。本发明具体技术方案如下。In order to achieve the above purpose and functions, before the system proposed by the present invention is used, it needs to go through the process of system establishment, that is, to collect performance sound signals of different styles on different pianos, collect expert subjective evaluation data and extract and analyze signal characteristics, determine Fuzzy the structure of the neural network and train it; the trained fuzzy neural network is ready to use. The specific technical scheme of the present invention is as follows.
一种结合曲风的钢琴演奏音质评价系统，其包括钢琴曲曲库、麦克风阵列录音装置、高质量麦克风录音装置、音乐数据库、专家听音评价模块、信号特征提取模块、样本库、信号特征分析模块和音质评价模块；A piano performance sound quality evaluation system combined with music style, which includes a piano music library, a microphone array recording device, a high-quality microphone recording device, a music database, an expert listening evaluation module, a signal feature extraction module, a sample library, and signal feature analysis module and sound quality evaluation module;
所述的钢琴曲曲库带有曲风标签，曲风标签是指在曲风分类的基础上对每种曲风进行编号标记；钢琴曲曲库有两个功能，一是在系统建立过程中，提供大量的钢琴曲资源，用于后续的分析和模糊神经网络的训练；二是在用户使用过程中，为用户提供可选的钢琴曲；The said piano melody storehouse has genre labels, and the melody label refers to numbering and marking each style of music on the basis of genre classification; the piano melody storehouse has two functions. , providing a large number of piano music resources for subsequent analysis and training of fuzzy neural networks; the second is to provide users with optional piano music during use;
所述的麦克风阵列录音装置包括在不同的空间位置上放置的麦克风，多个麦克风采集钢琴在不同位置的演奏音信号即空域信号，实现在宏观上对钢琴的演奏音质进行分析；The microphone array recording device includes microphones placed in different spatial positions, and multiple microphones collect the performance sound signals of the piano at different positions, that is, spatial signals, so as to analyze the sound quality of the piano performance macroscopically;
所述信号特征提取模块通过对麦克风阵列录音装置和高质量麦克风录音装置采集到的钢琴演奏音形成的音频文件进行信号特征提取，提取得到宏观上和微观上的信号特征，这些信号特征包括空域特征、时域特征、频域特征以及时频图特征；然后建立样本库，用于存储模糊神经网络的输入样本，输入样本包括训练样本和待评价样本两种样本类型；样本内容为信号特征向量，包括提取到的信号特征和对应的曲风标签；The signal feature extraction module extracts signal features from the audio files formed by the piano performance sounds collected by the microphone array recording device and the high-quality microphone recording device, and extracts macroscopic and microscopic signal features, and these signal features include spatial features. , time-domain features, frequency-domain features, and time-frequency map features; then establish a sample library for storing input samples of the fuzzy neural network, and the input samples include two types of samples: training samples and samples to be evaluated; the sample content is a signal feature vector, Including the extracted signal features and corresponding genre labels;
所述的信号特征分析模块实现模糊集和模糊推理规则建立的功能，包括将提取到的信号特征与从专家听音评价模块中得到的主观评价数据进行统计对比分析，由此建立信号特征和主观评价数据的模糊集，同时建立评价过程的模糊推理规则；The described signal feature analysis module realizes the function of establishing fuzzy sets and fuzzy inference rules, including statistically comparing and analyzing the extracted signal features with the subjective evaluation data obtained from the expert listening evaluation module, thereby establishing signal features and subjective evaluation data. Evaluate the fuzzy set of data and establish fuzzy inference rules for the evaluation process;
所述的音质评价模块实现将样本库中的样本经过模糊神经网络的处理后，输出音质评价分数的功能；此音质评价模块在使用前，需采用所述的信号特征分析模块建立的模糊推理规则来确定模糊神经网络的结构，并且将所述样本库的所有样本作为模糊神经网络的训练样本，而每个样本对应的音频文件的主观评价数据作为期望输出，即监督信号，然后对模糊神经网络进行训练；网络训练好后，音质评价模块即可使用；用户使用音质评价模块时，无需人工获取主观评价数据，就可实现在获取到待评价的样本后智能输出评价分数的功能。The sound quality evaluation module realizes the function of outputting the sound quality evaluation score after the samples in the sample library are processed by the fuzzy neural network; before the sound quality evaluation module is used, it needs to adopt the fuzzy inference rules established by the signal characteristic analysis module To determine the structure of the fuzzy neural network, and use all the samples of the sample library as the training samples of the fuzzy neural network, and the subjective evaluation data of the audio file corresponding to each sample as the expected output, that is, the supervisory signal, and then the fuzzy neural network Carry out training; after the network is trained, the sound quality evaluation module can be used; when users use the sound quality evaluation module, there is no need to manually obtain subjective evaluation data, and the function of intelligently outputting evaluation scores after obtaining samples to be evaluated can be realized.
基于上述的技术方案，本发明提出的利用一种结合曲风的钢琴演奏音质评价系统的评价方法，包括两个过程，一是系统建立过程，二是用户使用过程。Based on the above-mentioned technical solution, the evaluation method of the piano performance sound quality evaluation system proposed by the present invention includes two processes, one is the system establishment process, and the other is the user's use process.
系统在使用之前，需要采集不同钢琴上不同曲风的演奏音信号，收集专家主观评价数据并提取和分析信号特征，确定模糊神经网络的结构并进行训练。Before the system is used, it is necessary to collect performance sound signals of different styles on different pianos, collect expert subjective evaluation data and extract and analyze signal features, determine the structure of the fuzzy neural network and conduct training.
系统建立过程的步骤如下：The steps of the system establishment process are as follows:
（1）通过分析具有代表性的曲风分类规则，建立带有曲风标签的钢琴曲曲库，曲风标签是指在曲风分类的基础上对各曲风类型进行编号标记。(1) By analyzing the representative genre classification rules, a piano music library with genre tags is established. The genre label refers to the numbering and marking of each genre type on the basis of genre classification.
（2）使用麦克风阵列录音装置和高质量麦克风录音装置现场录制在多台钢琴上演奏的演奏音信号，演奏的曲子包含了钢琴曲曲库中的所有曲子。将麦克风阵列录音装置采集的演奏音信号进过处理之后形成多声道音频文件，高质量麦克风录音装置采集得到的演奏音信号经过相应处理之后形成高保真音频文件。(2) Use a microphone array recording device and a high-quality microphone recording device to record the performance sound signals played on multiple pianos on the spot, and the pieces played include all the pieces in the piano music library. The performance sound signal collected by the microphone array recording device is processed to form a multi-channel audio file, and the performance sound signal collected by the high-quality microphone recording device is processed accordingly to form a high-fidelity audio file.
（3）将高保真音频文件在听音室中进行回放，让多名专业人士进行听音实验，对其音质的好坏进行评价，收集主观评价数据，并将统计得到的主观评价数据、音频文件和曲风标签对应地存储到音乐数据库中。(3) Play back high-fidelity audio files in the listening room, let a number of professionals conduct listening experiments, evaluate the sound quality, collect subjective evaluation data, and count the subjective evaluation data, audio Files and genre tags are correspondingly stored in the music database.
（4）基于音频信号的短时平稳性，对多声道音频文件和高保真音频文件进行预加重处理和分帧处理。(4) Based on the short-term stationarity of audio signals, pre-emphasis processing and frame processing are performed on multi-channel audio files and high-fidelity audio files.
（5）将音乐数据库中的所有音频文件输入到信号特征提取模块进行信号特征的提取，并建立样本库，将提取到的信号特征和对应的曲风标签一起形成信号特征向量，对应地存储到样本库中，作为训练样本。(5) Input all audio files in the music database to the signal feature extraction module to extract signal features, and establish a sample library, form signal feature vectors with the extracted signal features and corresponding genre labels, and store them in corresponding sample library as training samples.
（6）将提取到的信号特征进行统计，与步骤（3）中的主观评价数据进行对比分析，从而建立信号特征和主观评价数据的模糊集，同时建立评价过程的模糊推理规则；所述模糊集和模糊推理规则是指判断音质好坏所需要的条件和判断逻辑在模糊数学中的描述；如可以做如下描述：如果基频泛音比例在x1范围，且空间均衡度在y1范围，则音质评价得分在z1范围；x1，y1，z1都属于各自的模糊集X、Y、Z。(6) Perform statistics on the extracted signal features, and compare and analyze them with the subjective evaluation data in step (3), so as to establish a fuzzy set of signal features and subjective evaluation data, and at the same time establish fuzzy inference rules for the evaluation process; the fuzzy Set and fuzzy inference rules refer to the description of the conditions and judgment logic required for judging the sound quality in fuzzy mathematics; for example, the following description can be made: if the overtone ratio of the fundamental frequency is in the range of x1, and the degree of spatial balance is in the range of y1, the sound quality Evaluation scores are in the range of z1; x1, y1, and z1 all belong to their respective fuzzy sets X, Y, and Z.
（7）根据步骤（5）中得到的信号特征向量的大小以及步骤（6）中分析得到的模糊推理规则，确定模糊神经网络的结构，将样本库的所有样本作为模糊神经网络的训练样本，而每个样本对应的音频文件的主观评价数据作为期望输出，即监督信号，然后对模糊神经网络进行训练，音质评价模块即完成建立。(7) Determine the structure of the fuzzy neural network according to the size of the signal feature vector obtained in step (5) and the fuzzy inference rules analyzed in step (6), and use all the samples in the sample library as the training samples of the fuzzy neural network, The subjective evaluation data of the audio file corresponding to each sample is used as the expected output, that is, the supervisory signal, and then the fuzzy neural network is trained, and the sound quality evaluation module is established.
将系统建立好之后，用户即可以使用此系统，无需人工获取主观评价数据，就可实现在获取到待评价的样本后智能输出评价分数的功能。用户使用系统的步骤如下：After the system is established, users can use the system, without manual acquisition of subjective evaluation data, and can realize the function of intelligently outputting evaluation scores after obtaining samples to be evaluated. The steps for users to use the system are as follows:
（1）让用户选择钢琴曲曲库中的一首曲子并在待评价的钢琴上演奏，麦克风阵列录音装置和高质量麦克风录音装置同时采集演奏音信号。(1) Let the user select a piece of music in the piano music library and play it on the piano to be evaluated. The microphone array recording device and the high-quality microphone recording device simultaneously collect performance sound signals.
（2）将演奏音信号经过处理后形成音频文件，与曲风标签一起存储到音乐数据库中；(2) Process the performance sound signal to form an audio file, and store it in the music database together with the genre label;
（3）将音频文件进行预处理后输入信号特征提取模块，将提取出的信号特征和曲风标签一起形成信号特征向量，存储到样本库中作为待评价样本。(3) After the audio file is preprocessed, it is input to the signal feature extraction module, and the extracted signal features and genre labels form a signal feature vector, which is stored in the sample library as a sample to be evaluated.
（4）将待评价样本输入到建立好的音质评价模块中，最终音质评价模块会输出一个在[0,100]范围内的评价分数，作为所选钢琴的演奏音质评价结果。(4) Input the sample to be evaluated into the established sound quality evaluation module, and the final sound quality evaluation module will output an evaluation score in the range of [0,100] as the evaluation result of the selected piano's playing sound quality.
与现有音质评价系统相比，本发明具有如下优点：Compared with the existing sound quality evaluation system, the present invention has the following advantages:
（1）本发明提出的结合曲风的钢琴演奏音质评价系统，在系统建立好之后，只需要给计算机输入一段钢琴演奏音信号，就可以无需后续人工对钢琴演奏音质进行评价，与专家人工主观评价相比，便利性、智能性大大提升，减少了人力成本和时间成本。(1) The piano performance sound quality evaluation system combined with music style proposed by the present invention, after the system is established, it only needs to input a piece of piano performance sound signal to the computer, and it can evaluate the piano performance sound quality without manual follow-up, which is different from the manual subjective evaluation of experts. Compared with the evaluation, the convenience and intelligence are greatly improved, and the labor cost and time cost are reduced.
（2）本发明提出的系统在系统建立过程中充分考虑了主观评价数据和客观信号特征，其中客观信号特征包含了空域特征、时域特征、频域特征和时频图特征，多特征分析方法让评价系统的客观性大大提升，避免了人工主观评价引起的评价标准浮动等问题。(2) The system proposed by the present invention fully considers the subjective evaluation data and objective signal features during the system establishment process, wherein the objective signal features include air domain features, time domain features, frequency domain features and time-frequency map features, and the multi-feature analysis method The objectivity of the evaluation system is greatly improved, and problems such as floating evaluation standards caused by manual subjective evaluation are avoided.
（3）本发明提出的系统还考虑了曲风这一种具备个性化特性的因素，并将曲风标签带入钢琴演奏中，在进过系统处理之后，用户可以得到自己偏好的曲风在待评价的钢琴上的音质好坏。因此，用户可以对比多台意向钢琴在自己偏好曲风上的评价分数，更好地辅助用户选择到真正适合自己的钢琴，让演奏达到最佳的效果；而钢琴的生产厂家根据用户的喜好，在生产个性化定制钢琴时，可以利用此系统进行辅助调试钢琴的各项参数。(3) The system proposed by the present invention also considers the music style, a factor with individual characteristics, and brings the music style label into the piano performance. After entering the system processing, the user can get the music style he prefers in the The sound quality on the piano to be evaluated is good or bad. Therefore, users can compare the evaluation scores of multiple intended pianos on their preferred music style, and better assist users to choose the piano that really suits them, so that the performance can achieve the best effect; and the piano manufacturer according to the user's preferences, When producing personalized customized pianos, this system can be used to assist in debugging various parameters of the piano.
附图说明Description of drawings
图1是实施例中一种结合曲风的钢琴演奏音质评价系统的总体结构框图。Fig. 1 is a block diagram of the overall structure of a piano performance sound quality evaluation system combined with musical genres in an embodiment.
图2是麦克风阵列录音装置结构框图。Fig. 2 is a structural block diagram of the microphone array recording device.
图3是实例中系统建立流程图。Fig. 3 is a flow chart of system establishment in the example.
图4是实例中用户使用流程图。Figure 4 is a flow chart of user usage in the example.
具体实施方式Detailed ways
为了使本发明的目的、技术方案及创新点和优点更清楚明白，下面结合附图对本发明的具体实施方式作进一步说明。In order to make the purpose, technical solution, innovations and advantages of the present invention clearer, the specific implementation manners of the present invention will be further described below in conjunction with the accompanying drawings.
下面结合附图对本发明的具体实施方式作进一步说明，但本发明的实施不限于此。The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings, but the implementation of the present invention is not limited thereto.
如图1所示，本发明系统包括九个模块：钢琴曲曲库，麦克风阵列录音装置，高质量麦克风录音装置，音乐数据库，专家听音评价模块，信号特征提取模块，样本库，信号特征分析模块，音质评价模块。As shown in Figure 1, the system of the present invention comprises nine modules: piano music library, microphone array recording device, high-quality microphone recording device, music database, expert listening evaluation module, signal feature extraction module, sample library, signal feature analysis module, sound quality evaluation module.
本实例的一种结合曲风的钢琴演奏音质评价系统，主要包括如下几个部分：A kind of piano performance sound quality evaluation system combined with music style in this example mainly includes the following parts:
（1）钢琴曲曲库：包括带有曲风标签的多首钢琴曲，曲风标签是指在曲风分类的基础上对各曲风进行编号标记。(1) Piano music library: including multiple piano songs with genre tags. Genre tags refer to the numbering and marking of each genre based on the classification of genres.
（2）麦克风阵列录音装置：该装置用来采集钢琴演奏音的空域信号以用于宏观上的分析。(2) Microphone array recording device: This device is used to collect the airspace signal of piano performance for macroscopic analysis.
（3）高质量麦克风录音装置：该用来采集钢琴高保真演奏音的信号以用于微观上的分析。(3) High-quality microphone recording device: it is used to collect the signal of high-fidelity piano performance for microscopic analysis.
（4）专家听音评价模块：在系统建立过程中，需采集钢琴曲曲库中所有钢琴曲在不同钢琴上的演奏音信号，然后将演奏音信号进行相应处理之后形成音频文件；然后通过音频文件的回放及专家听音评价这个过程，获取主观评价数据。(4) Expert listening evaluation module: During the establishment of the system, it is necessary to collect the performance sound signals of all piano pieces in the piano music library on different pianos, and then perform corresponding processing on the performance sound signals to form audio files; The playback of files and the process of listening and evaluating by experts to obtain subjective evaluation data.
（5）音乐数据库：将上述的曲风标签、音频文件和主观评价数据对应存储到音乐数据库中；在用户使用系统过程中则无需音频文件的回放及专家听音评价这个过程，因此存储到音乐库中的数据仅包含音频文件和曲风标签。(5) Music database: store the above-mentioned genre labels, audio files and subjective evaluation data in the music database; in the process of using the system, there is no need for audio file playback and expert listening evaluation, so it is stored in the music database. The data in the library only includes audio files and genre tags.
（6）信号特征提取模块：对音乐数据库中的演奏音音频文件进行预处理之后，通过该模块提取钢琴演奏音宏观上和微观上的信号特征，这些信号特征主要包括空域特征、时域特征、频域特征以及时频图特征。(6) Signal feature extraction module: After preprocessing the performance sound audio files in the music database, the macro and micro signal features of the piano performance sound are extracted through this module. These signal features mainly include spatial features, time domain features, Frequency domain features and time-frequency map features.
（7）样本库：此部分用于存储模糊神经网络的输入样本，样本内容为信号特征向量，包括提取到的信号特征和对应的曲风标签，样本类型包括训练样本和待评价样本两种。(7) Sample library: This part is used to store the input samples of the fuzzy neural network. The sample content is the signal feature vector, including the extracted signal features and the corresponding genre labels. The sample types include training samples and samples to be evaluated.
（8）信号特征分析模块：该模块主要用于系统建立过程中，实现模糊推理规则建立的功能，包括将提取到的信号特征与主观评价数据进行统计对比分析，由此建立信号特征和主观评价数据的模糊集，同时建立评价过程的模糊推理规则。(8) Signal feature analysis module: This module is mainly used in the process of system establishment to realize the function of establishing fuzzy inference rules, including statistical comparison and analysis of the extracted signal features and subjective evaluation data, thereby establishing signal features and subjective evaluation Fuzzy sets of data, and establish fuzzy inference rules for the evaluation process.
（9）音质评价模块：该模块实现了将样本库中的样本经过模糊神经网络的处理后，输出音质评价分数的功能。此模块在使用前，需采用（8）中的信号特征分析模块建立的模糊推理规则来确定模糊神经网络的结构，并且将（7）中样本库的所有样本作为模糊神经网络的训练样本，而每个样本对应的音频文件的主观评价数据作为期望输出，即监督信号，然后对模糊神经网络进行训练。网络训练好后，音质评价模块即可使用。用户使用该模块时，无需人工获取主观评价数据，就可实现在获取到待评价的样本后智能输出评价分数的功能。(9) Sound quality evaluation module: This module realizes the function of outputting sound quality evaluation scores after the samples in the sample library are processed by the fuzzy neural network. Before using this module, it is necessary to use the fuzzy inference rules established by the signal feature analysis module in (8) to determine the structure of the fuzzy neural network, and use all the samples in the sample library in (7) as the training samples of the fuzzy neural network, and The subjective evaluation data of the audio file corresponding to each sample is used as the expected output, that is, the supervisory signal, and then the fuzzy neural network is trained. After the network is trained, the sound quality evaluation module can be used. When users use this module, they can realize the function of intelligently outputting evaluation scores after obtaining samples to be evaluated without manually obtaining subjective evaluation data.
本发明提出的系统在使用之前，需要先经过系统建立的过程，即采集不同钢琴上不同曲风的演奏音信号，收集专家主观评价数据并提取和分析信号特征，确定模糊神经网络的结构并进行训练；训练好的模糊神经网络即可使用。因此，系统建立过程包括所有九个模块，而用户使用过程只包括其中七个模块：钢琴曲曲库，麦克风阵列录音装置，高质量麦克风录音装置，音乐数据库，信号特征提取模块，样本库，音质评价模块。Before the system proposed by the present invention is used, it needs to go through the process of system establishment, that is, to collect the playing sound signals of different styles on different pianos, collect the subjective evaluation data of experts and extract and analyze the signal characteristics, determine the structure of the fuzzy neural network and carry out Training; the trained fuzzy neural network is ready to use. Therefore, the system establishment process includes all nine modules, while the user use process only includes seven of them: piano music library, microphone array recording device, high-quality microphone recording device, music database, signal feature extraction module, sample library, sound quality evaluation module.
本实例中的钢琴曲曲库可包括按照钢琴曲发展时期而分类形成的四类风格：巴洛克风格、古典主义风格、浪漫主义风格、现代主义风格，曲风标签分别编号1,2,3,4；每类风格选取三首典型的练习曲；因此曲库中一共包括12首钢琴曲。The piano music library in this example can include four types of styles classified according to the development period of the piano music: baroque style, classic style, romantic style, modern style, and the style tags are respectively numbered 1, 2, 3, and 4 ; Select three typical etudes for each style; therefore, the music library includes a total of 12 piano pieces.
本实施例中的麦克风阵列录音装置如图2所示，包含18个麦克风，接口集合板，USB转串口信号传输控制电路，放大带通电路，A/D模块，ARM信号处理存储模块。其中放大带通电路，A/D模块，ARM信号处理存储模块集合成一块集合电路，每块集合电路板接收3个麦克风采集的演奏音信号，共6块这样的集合电路板，包括了1个主机集合电路和5个从机集合电路；所采用的采样频率是100kHz，量化精度是12bit；而USB转串口信号传输控制电路主要负责接收计算机端录制起止的信号，当计算机发出开始接收的指令时，通过此模块将指令传输给主机，并且主机通过和从机之间连接的信号线，将指令同时传给每个从机，然后其他模块便开始工作；所述ARM信号处理存储模块中数据以bin文件的形式存储在SD存储卡上，通过相应程序将其转化为多声道wav音频文件。The microphone array recording device in this embodiment, as shown in Figure 2, includes 18 microphones, an interface assembly board, a USB-to-serial port signal transmission control circuit, an amplification bandpass circuit, an A/D module, and an ARM signal processing storage module. Among them, the amplification band-pass circuit, A/D module, and ARM signal processing and storage module are integrated into an integrated circuit, and each integrated circuit board receives performance sound signals collected by three microphones, and there are 6 such integrated circuit boards in total, including 1 The host assembly circuit and 5 slave assembly circuits; the sampling frequency used is 100kHz, and the quantization precision is 12bit; and the USB-to-serial port signal transmission control circuit is mainly responsible for receiving the recording start and stop signals from the computer. , the instruction is transmitted to the master through this module, and the master transmits the instruction to each slave at the same time through the signal line connected with the slave, and then other modules start to work; the ARM signal processes the data in the storage module to Bin files are stored on the SD memory card, and converted into multi-channel wav audio files through corresponding programs.
高质量麦克风录音装置包含可以采集高保真演奏音信号的单个麦克风，并结合Adobe Audition软件生成高保真wav音频文件，所选采样频率为96kHz，量化精度16bit。The high-quality microphone recording device includes a single microphone that can collect high-fidelity performance sound signals, combined with Adobe Audition software to generate high-fidelity wav audio files, the selected sampling frequency is 96kHz, and the quantization precision is 16bit.
本实例中的音乐数据库主要用于存储多声道wav音频文件和高保真wav音频文件，这些音频文件与经过专家听音评价模块后获得的主观评价数据，以及曲风标签对应存储到该数据库中，从而建立具有多种风格、多台钢琴、不同音质评价分数标签的音乐数据库。The music database in this example is mainly used to store multi-channel wav audio files and high-fidelity wav audio files. These audio files are stored in the database corresponding to the subjective evaluation data obtained after the expert listening evaluation module, as well as genre labels , so as to establish a music database with multiple styles, multiple pianos, and different sound quality evaluation score labels.
而其中所述专家听音评价模块主要包括以下内容：将用高保真wav音频文件截取部分在听音室中回放，邀请5名专家进行听音实验，根据声音稳定性、丰富度、明亮度、饱满度等指标对每一个截取音频进行评价，然后对两两音频文件之间音质对比并进整体评分，并对收集到的主观评价数据进行统计和分析。Wherein said expert listening evaluation module mainly includes the following contents: the intercepted part of the high-fidelity wav audio file will be played back in the listening room, and 5 experts will be invited to carry out the listening experiment, according to sound stability, richness, brightness, Indexes such as fullness evaluate each intercepted audio, and then compare the sound quality between two audio files and make an overall score, and conduct statistics and analysis on the collected subjective evaluation data.
本实例中的音乐特征提取模块，主要实现了将音乐数据库中音频文件作为输入，提取得到空域、时域、频域、时频图等方面的特征的功能；这些特征与音质指标间的关系如下：The music feature extraction module in this example mainly implements the function of extracting the features of airspace, time domain, frequency domain, and time-frequency graph from audio files in the music database as input; the relationship between these features and sound quality indicators is as follows :
1）基于多声道wav音频文件提取的空间均衡度等宏观上的空域信号特征与声音的稳定性、音区及过渡区自然度有关；1) Based on the spatial balance extracted from the multi-channel wav audio file, the macroscopic spatial domain signal characteristics are related to the stability of the sound, the sound range and the naturalness of the transition area;
2）基于高保真wav音频文件提取的起振时间、单音时值等微观上的时域包络特征与声音的音色明亮或低沉有关，也直接影响音质好坏；2) Based on the high-fidelity wav audio file extraction, the microscopic time-domain envelope characteristics such as the start-up time and the single-tone time value are related to the bright or deep timbre of the sound, and also directly affect the sound quality;
3）基于高保真wav音频文件提取的基频和泛音的振幅及振幅比例等微观上的频域特征与声音的音色表现力有关；3) Based on the high-fidelity wav audio files, the microscopic frequency domain features such as the fundamental frequency and overtone amplitude and amplitude ratio are related to the timbre expressiveness of the sound;
4）时频图中显示出基频、泛音的整体特征与声音的音色饱满度、协调性有关。4) The time-frequency diagram shows that the overall characteristics of the fundamental frequency and overtone are related to the fullness and coordination of the sound.
本实例中的样本库，用于存储音质评价模块的输入样本，包括训练样本和待评价样本两种样本类型；样本内容为信号特征向量，包括提取到的信号特征和对应的曲风标签。The sample library in this example is used to store the input samples of the sound quality evaluation module, including training samples and samples to be evaluated; the sample content is a signal feature vector, including the extracted signal features and corresponding genre labels.
本实例中的音质评价模块为本系统的主要模块，其主结构为多输入单输出的五层神经网络，内部逻辑和权函数的选择依赖于信号特征分析模块建立的模糊集和模糊推理规则。The sound quality evaluation module in this example is the main module of this system, and its main structure is a five-layer neural network with multiple inputs and single outputs. The selection of internal logic and weight function depends on the fuzzy set and fuzzy inference rules established by the signal feature analysis module.
而其中所述音乐信号特征分析模块主要实现模糊推理规则建立的功能，包括将提取到的信号特征与主观评价数据进行统计对比分析，由此建立信号特征和主观评价数据的模糊集，同时建立评价过程的模糊推理规则；而所述模糊集和模糊规则是指判断音质好坏所需要的条件和判断逻辑在模糊数学中的描述；如可以做如下描述：如果基频泛音比例在x1范围，且空间均衡度在y1范围，则音质评价得分在z1范围。x1，y1，z1都属于各自的模糊集X、Y、Z。Wherein said music signal feature analysis module mainly realizes the function of establishing fuzzy inference rules, including carrying out statistical comparative analysis of the signal features extracted and subjective evaluation data, thereby establishing a fuzzy set of signal features and subjective evaluation data, and simultaneously establishing evaluation The fuzzy inference rules of the process; and the fuzzy set and fuzzy rules refer to the description in fuzzy mathematics of the conditions and judgment logic needed for judging the sound quality; as can be described as follows: if the fundamental frequency overtone ratio is in the x1 range, and If the spatial balance is in the y1 range, the sound quality evaluation score is in the z1 range. x1, y1, z1 all belong to their respective fuzzy sets X, Y, Z.
本实例的系统在使用之前，需要先经过系统建立的过程，即采集不同钢琴上不同曲风的演奏音信号，收集专家主观评价数据并提取和分析信号特征，确定模糊神经网络的结构并进行训练；训练好的模糊神经网络即可使用。系统建立过程的流程图如图3所示。Before the system in this example is used, it needs to go through the process of system establishment, that is, to collect the playing sound signals of different styles on different pianos, collect the subjective evaluation data of experts and extract and analyze the signal characteristics, determine the structure of the fuzzy neural network and conduct training ; The trained fuzzy neural network can be used. The flow chart of the system establishment process is shown in Figure 3.
系统建立的流程如下：The process of system establishment is as follows:
（1）建立钢琴曲曲库：如前所述，曲库中一共包括四种风格的12首钢琴曲。(1) Establish a piano music library: As mentioned above, the music library includes a total of 12 piano songs in four styles.
（2）录音装置开启，演奏者演奏曲库中的曲子：使用麦克风阵列录音装置和高质量麦克风录音装置现场录制在多台钢琴上演奏的演奏音信号，演奏的曲子包含了钢琴曲曲库中的所有曲子。(2) The recording device is turned on, and the performer plays the pieces in the music library: use the microphone array recording device and the high-quality microphone recording device to record the performance sound signals played on multiple pianos on the spot, and the pieces played include the piano music library all the tunes.
（3）形成音频文件：将麦克风阵列录音装置采集的演奏音信号进过处理之后形成多声道音频文件，高质量麦克风录音装置采集得到的演奏音信号经过相应处理之后形成高保真音频文件。为了对应本实例的钢琴曲曲库中的每种风格的每首练习曲，因此让3名演奏者在4架差异明显的钢琴上演奏，共可获得48个高保真音频文件和48个16声道的音频文件。(3) Forming audio files: The performance sound signals collected by the microphone array recording device are processed to form multi-channel audio files, and the performance sound signals collected by the high-quality microphone recording device are processed accordingly to form high-fidelity audio files. In order to correspond to each etude of each style in the piano music library of this example, let 3 performers play on 4 pianos with obvious differences, and a total of 48 high-fidelity audio files and 48 16-voice channel audio files.
（4）音频文件回放，专家听音评价，建立音乐数据库：将高保真音频文件在听音室中进行回放，让多名专业人士进行听音实验，对其音质的好坏进行评价，收集主观评价数据，并将统计得到的主观评价数据、音频文件和曲风标签对应地存储到音乐数据库中。(4) Audio file playback, expert listening evaluation, and establishment of a music database: Play back high-fidelity audio files in the listening room, let a number of professionals conduct listening experiments, evaluate the sound quality, and collect subjective evaluation data, and correspondingly store the statistically obtained subjective evaluation data, audio files and genre labels in the music database.
（5）音频文件预处理：基于音频信号的短时平稳性，对多声道音频文件和高保真音频文件进行预加重处理和分帧处理。(5) Audio file preprocessing: Based on the short-term stationarity of audio signals, pre-emphasis processing and frame processing are performed on multi-channel audio files and high-fidelity audio files.
（6）提取信号特征，形成训练样本：将音乐数据库中的所有音频文件输入到信号特征提取模块进行信号特征的提取，并建立样本库，将提取到的信号特征和对应的曲风标签一起形成信号特征向量，对应地存储到样本库中，作为训练样本。(6) Extract signal features to form training samples: input all audio files in the music database to the signal feature extraction module to extract signal features, and establish a sample library, and form the extracted signal features and corresponding genre labels together The signal feature vectors are correspondingly stored in the sample library as training samples.
（7）分析信号特征和主观评价数据，建立模糊集和模糊推理规则：将提取到的信号特征进行统计，与（4）得到的主观评价数据进行对比分析，从而建立信号特征和主观评价数据的模糊集，同时建立评价过程的模糊推理规则；所述模糊集和模糊推理规则是指判断音质好坏所需要的条件和判断逻辑在模糊数学中的描述；如可以做如下描述：如果基频泛音比例在x1范围，且空间均衡度在y1范围，则音质评价得分在z1范围；x1，y1，z1都属于各自的模糊集X、Y、Z。(7) Analyze signal characteristics and subjective evaluation data, and establish fuzzy sets and fuzzy inference rules: make statistics on the extracted signal characteristics, and compare and analyze with the subjective evaluation data obtained in (4), so as to establish the relationship between signal characteristics and subjective evaluation data. Fuzzy set, set up the fuzzy inference rule of evaluation process simultaneously; Described fuzzy set and fuzzy inference rule refer to the description in fuzzy mathematics of the condition that needs to judge sound quality and judgment logic; As can be described as follows: If fundamental frequency overtone If the ratio is in the range of x1, and the degree of spatial balance is in the range of y1, then the sound quality evaluation score is in the range of z1; x1, y1, and z1 all belong to their respective fuzzy sets X, Y, and Z.
（8）确定模糊神经网络的结构，训练模糊神经网络：根据步（6）中得到的信号特征向量的大小以及步骤（7）中分析得到的模糊推理规则，确定模糊神经网络的结构，将样本库的所有样本作为模糊神经网络的训练样本，而每个样本对应的音频文件的主观评价数据作为期望输出，即监督信号，然后对模糊神经网络进行训练。(8) Determine the structure of the fuzzy neural network and train the fuzzy neural network: According to the size of the signal feature vector obtained in step (6) and the fuzzy inference rules analyzed in step (7), determine the structure of the fuzzy neural network, and the sample All the samples in the library are used as the training samples of the fuzzy neural network, and the subjective evaluation data of the audio file corresponding to each sample is used as the expected output, that is, the supervisory signal, and then the fuzzy neural network is trained.
验证验证评价结果已达最优后，音质评价模块即完成建立。系统成功建立后，用户即可以使用该系统，无需人工获取主观评价数据，就可实现在获取到待评价的样本后智能输出评价分数的功能；用户使用系统的流程图4所示。After verifying and verifying that the evaluation result has reached the optimum, the establishment of the sound quality evaluation module is completed. After the system is successfully established, users can use the system without manually obtaining subjective evaluation data, and can realize the function of intelligently outputting evaluation scores after obtaining the samples to be evaluated; the flow chart 4 of the user using the system is shown.
用户使用系统过程的流程如下：The flow of the user using the system process is as follows:
（1）用户选择偏好的曲风：用户在钢琴曲曲库中选择偏好的曲风的曲子，如用户选择的是浪漫主义风格的曲子，曲风标签编号设置为3，在希望获得音质评价的钢琴上正确地演奏所选曲子。(1) The user selects the preferred style of music: the user selects the preferred style of music in the piano music library. For example, if the user chooses a piece of romantic style, the label number of the style is set to 3. If you want to obtain sound quality evaluation Play the selected piece correctly on the piano.
（2）录音装置开启，用户在待评价的钢琴上演奏所选曲风的曲子：麦克风阵列录音装置和高质量麦克风录音装置同时采集用户的演奏音信号。(2) The recording device is turned on, and the user plays the selected style of music on the piano to be evaluated: the microphone array recording device and the high-quality microphone recording device simultaneously collect the user's performance sound signal.
（3）形成音频文件，结合曲风标签存储到音乐数据库：演奏结束后，将采集的到的演奏音信号经过处理后形成音频文件，与曲风标签一起存储到音乐数据库中。(3) Form an audio file and store it in the music database in combination with the genre tag: After the performance, the collected performance sound signal is processed to form an audio file, which is stored in the music database together with the genre tag.
（4）音频文件预处理：基于音频信号的短时平稳性，对音频文件进行预加重处理和分帧处理。(4) Audio file preprocessing: Based on the short-term stationarity of audio signals, pre-emphasis processing and frame processing are performed on audio files.
（5）提取信号特征，形成待评价样本：将预处理后的音频文件输入信号特征提取模块，将提取出的信号特征和曲风标签一起形成信号特征向量，存储到样本库中作为待评价样本。(5) Extract signal features to form samples to be evaluated: input the preprocessed audio files into the signal feature extraction module, and form signal feature vectors with the extracted signal features and genre labels, and store them in the sample library as samples to be evaluated .
（6）进行音质评价，输出音质评价分数：将待评价样本输入到建立好的音质评价模块中，最终音质评价模块会输出一个在[0,100]范围内的评价分数，如65分，此分数即为所弹钢琴在所选曲风下的演奏音质评价结果。(6) Perform sound quality evaluation and output sound quality evaluation scores: Input the sample to be evaluated into the established sound quality evaluation module, and finally the sound quality evaluation module will output an evaluation score in the range of [0,100], such as 65 points, this score is It is the evaluation result of the playing sound quality of the piano played under the selected music style.
由此，用户可以对比多台意向钢琴在自己偏好曲风上的评价分数，以此来辅助选择到真正适合自己的钢琴，让演奏达到最佳的效果；而钢琴的生产厂家根据用户的喜好，在生产个性化定制钢琴时，可以利用此系统进行辅助调试钢琴的各项参数，从而减少很大一部分的人工评价成本和时间成本。In this way, users can compare the evaluation scores of multiple intended pianos in their preferred style, so as to help them choose the piano that really suits them, so that the performance can achieve the best effect; and the piano manufacturer according to the user's preferences, When producing personalized customized pianos, this system can be used to assist in debugging various parameters of the piano, thereby reducing a large part of manual evaluation costs and time costs.
上述实施例需求为本发明较佳的实施方式，但本发明的实施方式并不受上述实施例的限制，其他的任何未背离本发明的精神实质与原理下所作的改变、修饰、替代、组合、简化，均应为等效的置换方式，都包含在本发明的保护范围。The above-mentioned embodiment needs to be the preferred implementation mode of the present invention, but the implementation mode of the present invention is not limited by the above-mentioned embodiment, and any other changes, modifications, substitutions and combinations made without departing from the spirit and principle of the present invention , simplification, all should be equivalent replacement methods, and are all included in the protection scope of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810311753.2ACN108735192B (en) | 2018-04-09 | 2018-04-09 | System and method for evaluating piano playing tone quality by combining music |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810311753.2ACN108735192B (en) | 2018-04-09 | 2018-04-09 | System and method for evaluating piano playing tone quality by combining music |
| Publication Number | Publication Date |
|---|---|
| CN108735192Atrue CN108735192A (en) | 2018-11-02 |
| CN108735192B CN108735192B (en) | 2021-09-21 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810311753.2AExpired - Fee RelatedCN108735192B (en) | 2018-04-09 | 2018-04-09 | System and method for evaluating piano playing tone quality by combining music |
| Country | Link |
|---|---|
| CN (1) | CN108735192B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109991842A (en)* | 2019-03-14 | 2019-07-09 | 合肥工业大学 | Method and system for piano tuning based on neural network |
| CN110516104A (en)* | 2019-08-27 | 2019-11-29 | 腾讯音乐娱乐科技（深圳）有限公司 | Song recommendations method, apparatus and computer storage medium |
| CN110675879A (en)* | 2019-09-04 | 2020-01-10 | 平安科技（深圳）有限公司 | Big data-based audio evaluation method, system, device and storage medium |
| CN113108894A (en)* | 2021-03-30 | 2021-07-13 | 华南理工大学 | Transient sound field measuring system and method of musical instrument soundboard based on audio array acquisition |
| CN114842876A (en)* | 2022-04-27 | 2022-08-02 | 厦门大学 | Lute tone quality evaluation method based on vibroacoustics |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105719661A (en)* | 2016-01-29 | 2016-06-29 | 西安交通大学 | Automatic discrimination method for playing timbre of string instrument |
| CN107146497A (en)* | 2016-08-02 | 2017-09-08 | 浙江大学 | A Grading System for Piano Examination |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105719661A (en)* | 2016-01-29 | 2016-06-29 | 西安交通大学 | Automatic discrimination method for playing timbre of string instrument |
| CN107146497A (en)* | 2016-08-02 | 2017-09-08 | 浙江大学 | A Grading System for Piano Examination |
| Title |
|---|
| OLGA SLIZOVSKAIA ET AL.: ""musical instrument recognition in user-generated videos using a multimodal convolutional neural network architecture"", 《ACM ON INTERNATIONAL CONFERENCE ON MULTIMEDIA RETRIEVAL》* |
| 刘乐: ""钢琴演奏评价系统研究"", 《中国优秀博硕士学位论文全文数据库（硕士） 信息科技辑》* |
| 吴晶晶: ""钢琴音乐信号的特征识别"", 《中国优秀硕士学位论文全文数据库 哲学与人文科学辑》* |
| 黄阿罗等: ""声乐演唱中音色的评价标准"", 《黄河之声》* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109991842A (en)* | 2019-03-14 | 2019-07-09 | 合肥工业大学 | Method and system for piano tuning based on neural network |
| CN110516104A (en)* | 2019-08-27 | 2019-11-29 | 腾讯音乐娱乐科技（深圳）有限公司 | Song recommendations method, apparatus and computer storage medium |
| CN110675879A (en)* | 2019-09-04 | 2020-01-10 | 平安科技（深圳）有限公司 | Big data-based audio evaluation method, system, device and storage medium |
| CN113108894A (en)* | 2021-03-30 | 2021-07-13 | 华南理工大学 | Transient sound field measuring system and method of musical instrument soundboard based on audio array acquisition |
| CN114842876A (en)* | 2022-04-27 | 2022-08-02 | 厦门大学 | Lute tone quality evaluation method based on vibroacoustics |
| Publication number | Publication date |
|---|---|
| CN108735192B (en) | 2021-09-21 |
| Publication | Publication Date | Title |
|---|---|---|
| CN108735192B (en) | System and method for evaluating piano playing tone quality by combining music | |
| CN105719661B (en) | A method for automatically judging the sound quality of stringed instruments | |
| CN109147804A (en) | A kind of acoustic feature processing method and system based on deep learning | |
| CN102723079B (en) | Music and chord automatic identification method based on sparse representation | |
| Saitis et al. | Brightness perception for musical instrument sounds: Relation to timbre dissimilarity and source-cause categories | |
| KR20070059102A (en) | Content Creation Device and Content Creation Method | |
| Fritz et al. | Perceptual evaluation of musical instruments: State of the art and methodology | |
| Chourdakis et al. | A machine-learning approach to application of intelligent artificial reverberation | |
| KR101459324B1 (en) | Evaluation method of sound source and Apparatus for evaluating sound using it | |
| Turchet et al. | Music emotion recognition: intention of composers-performers versus perception of musicians, non-musicians, and listening machines | |
| CN109189982A (en) | A kind of music emotion classification method based on SVM Active Learning | |
| Poirson et al. | Study of the brightness of trumpet tones | |
| Herbst | Distortion and rock guitar harmony: the influence of distortion level and structural complexity on acoustic features and perceived pleasantness of guitar chords | |
| Liu et al. | Research on the correlation between the timbre attributes of musical sound and visual color | |
| CN119629543A (en) | Channel calibration method, device, equipment and storage medium | |
| CN113781989B (en) | Audio animation playing and rhythm stuck point identifying method and related device | |
| CN115662465A (en) | Voice recognition algorithm and device suitable for national stringed instruments | |
| CN108513227B (en) | Modern electronic organ manufacturing method based on loudspeaker array design | |
| Raman et al. | Bach, Mozart, and Beethoven: Sorting piano excerpts based on perceived similarity using DiSTATIS | |
| CN105895079A (en) | Voice data processing method and device | |
| Jiang et al. | Analysis and modeling of timbre perception features of chinese musical instruments | |
| Talavera et al. | Transcription of guitar chords from acoustic audio | |
| Hantschel et al. | The use of prototype theory for understanding the perception and concept formation of musical styles | |
| Liu et al. | Comparison and analysis of timbre fusion for chinese and western musical instruments | |
| Thilakan et al. | Classification of the perceptual impression of source-level blending between violins in a joint performance |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | Granted publication date:20210921 | |
| CF01 | Termination of patent right due to non-payment of annual fee |