|
|
Variable segmentation for sound signal processing
Garai, Szabolcs ; Míča, Ivan (referee) ; Balík, Miroslav (advisor)
This paper describes the methods used mainly in the filtration of audio signals -- noise reduction. It realizes segmentation with variable lenght of segment, constant overlap add and segmentation with variable lenght of segment and overlap add. These methods are then compared with comon methods of segmentation in dependance of the lenght of segment and used window function by the thresholding method. For this purpose it uses the database of audio records. In the first part it desrcibes the technique of audio signal processing with comon method of segmentation. According to this method it continues in design of signal processing by variable segmentation method, in which it is needed to modify the shape of window function, which influences and attributes are explained in next chapter. In practical part it describes the implemented methods in MATLAB programming language with each steps of testing. It continues with chart of enclosed files and the evaluation of the results of hearing tests.
|
|
|
Spectral method for motor parameters identification
Sedláček, Filip ; Buchta, Luděk (referee) ; Pohl, Lukáš (advisor)
This project deals with the problem of spectral method for system identification. Introductory part acquaints reader with basic concept of identification and explains its application in practice. It describes and mathematically derives two non-parametric methods for identification – correlation method and spectral analysis. The main aim of this final project is testing and implementation of spectral analysis in system parameters identification, which is specifically described in remaining sections of this project. These sections discuss a fundamental rules, procedures and principles in order to prepare a reader as much as possible for its final application in REAL-TIME system identification.
|
|
| |
|
|
Simulation of sound source motion for multichannel speaker topology
Dvořák, Petr ; Kurc, David (referee) ; Orlovský, Kristián (advisor)
This paper describes the simulation of the motion of a sound source in space with several speakers. It deals mainly with the study and subsequent implementation of events necessary for simulating the motion of the virtual sound source. The theoretical part is therefore concerned with finding these phenomena and with their description. Mainly that means description of panning of the sound and of the Doppler effect. In the following practical part of the work, the previously described phenomena are applied into Matlab. In the created interface , it is then possible to select the parameters of the simulation, such as the trajectory of the motion of the virtual sound source, the number of the points of the trajectory, etc. Analyses of the generated signal are then made of the selected referential parameters of the simulation.
|
|
|
Audio effect for pitch modification
Čermák, Václav ; Sysel, Petr (referee) ; Schimmel, Jiří (advisor)
This thesis deals with the realisation of an effect for real-time pitch detection and pitch shifting. At first, properties and basic operations of audio signals are analyzed for time and frequency analysis. An emphasis is placed on signal conversion from the time domain to the frequency domain. Methods for audio pitch shifting, fundamental frequency detection and voiced parts determination are discussed next. The practical part describes design of relevant algorithms, their implementation in Matlab environment and evaluation of the results. Evaluation is followed by the creation of a VST plug-in for real-time pitch shifting, which contains chosen algorithms from the previous section and is the final output of this thesis.
|
|
| |
|
|
Spectral method for motor parameters identification
Sedláček, Filip ; Buchta, Luděk (referee) ; Pohl, Lukáš (advisor)
This project deals with the problem of spectral method for system identification. Introductory part acquaints reader with basic concept of identification and explains its application in practice. It describes and mathematically derives two non-parametric methods for identification – correlation method and spectral analysis. The main aim of this final project is testing and implementation of spectral analysis in system parameters identification, which is specifically described in remaining sections of this project. These sections discuss a fundamental rules, procedures and principles in order to prepare a reader as much as possible for its final application in REAL-TIME system identification.
|
|
|
Simulation of sound source motion for multichannel speaker topology
Dvořák, Petr ; Kurc, David (referee) ; Orlovský, Kristián (advisor)
This paper describes the simulation of the motion of a sound source in space with several speakers. It deals mainly with the study and subsequent implementation of events necessary for simulating the motion of the virtual sound source. The theoretical part is therefore concerned with finding these phenomena and with their description. Mainly that means description of panning of the sound and of the Doppler effect. In the following practical part of the work, the previously described phenomena are applied into Matlab. In the created interface , it is then possible to select the parameters of the simulation, such as the trajectory of the motion of the virtual sound source, the number of the points of the trajectory, etc. Analyses of the generated signal are then made of the selected referential parameters of the simulation.
|
|
|
Variable segmentation for sound signal processing
Garai, Szabolcs ; Míča, Ivan (referee) ; Balík, Miroslav (advisor)
This paper describes the methods used mainly in the filtration of audio signals -- noise reduction. It realizes segmentation with variable lenght of segment, constant overlap add and segmentation with variable lenght of segment and overlap add. These methods are then compared with comon methods of segmentation in dependance of the lenght of segment and used window function by the thresholding method. For this purpose it uses the database of audio records. In the first part it desrcibes the technique of audio signal processing with comon method of segmentation. According to this method it continues in design of signal processing by variable segmentation method, in which it is needed to modify the shape of window function, which influences and attributes are explained in next chapter. In practical part it describes the implemented methods in MATLAB programming language with each steps of testing. It continues with chart of enclosed files and the evaluation of the results of hearing tests.
|
guest ::
