Conclusion
Summary: Conclusion of determining room inpulse responses and deconvolving the response out of a recorded signal.
Conclusion
After taking the impulse responses and comparing them to the theoretical values, we noticed large differences. These differences could be due to the non ideality of the rooms. In other words our rooms were not perfectly rectangular as the theoretical model assumed. Objects in the room and clipping were also neglected in theoretical model. The non-linear effect of clipping not only removed information from the signal, but did so in such a way that the lost information was unrecoverable by our Fourier analysis. With a better theoretical model, we could take responses from multiple sources and receivers quickly and get a better comparison to real world room responses. Commercial application of measuring multiple impulse responses such as measuring the response from each seat in an orchestra hall would require a robust theoretical model that accounted for objects in the room and for fact that most real world spaces are not perfectly rectangular. Surface materials played an important factor in the measurement of the impulse responses. Most of the rooms we considered had similar reflection coefficients, however the Will Rice room had wooden ceilings that had a drastic impact on the impulse response. The lower reflection coefficient led to more distortion in the signal and a loss of energy compared to the other surfaces
The deconvolution of the signal was intended to remove the room response on a recorded signal, but in the process the noise was amplified. Much of the noise was in the signal and could not be easily filtered out. The deconvolution did reproduce the original signal, however the quality was significantly worse than the recorded signal. With a better method of deconvolution clearer signals could be produced and ideally the original high quality recorded signal. This method would need to implement advanced filtering techniques. The clipping of our signal caused the deconvolution to remove part of the response that was already taken out, creating amplitude aliasing. Perfect deconvolution would be useful in creating a clean recorded signal, almost regardless of recording environment. Deconvolution has uses in non-acoustic signals as well. Given a signal and the impulse response of the environment in which it was taken, the original signal should be retrievable through deconvolution given that the system is linear and time-invariant. For specific cases, a true impulse response isn't needed. If the response for all possible frequencies is known, it can be used instead of the impulse response with the other frequencies filtered out to remove noise. The specific response can be deconvolved from a known input signal and the recorded output of the system.
Comments, questions, feedback, criticisms?
Send feedback
- E-mail the authors of the module, Conclusion
Footer
More about this content: Metadata | Version History
This work is licensed by Bryce Luna under a Creative Commons Attribution License (CC-BY 2.0).
Last edited by Bryce Luna on Dec 13, 2005 8:47 pm US/Central.