Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Electrical signal parameter measurement system
Reexamination Certificate
1999-08-18
2002-09-17
Hoff, Marc S. (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Electrical signal parameter measurement system
C381S059000
Reexamination Certificate
active
06453253
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to impulse response measuring methods that produce impulse responses of anechoic chambers from impulse responses measured in echo chambers.
This application is based on Patent Application No. Hei 10-250442 filed in Japan, the content of which is incorporated herein by reference.
2. Description of the Related Art
Impulse responses of electroacoustic transducers (or electroacoustic converters) such as speakers and microphones should be originally measured in anechoic chambers to avoid effects due to echo and reverberation. However, it is not easy to construct the echo chambers. For this reason, it is necessary for engineers (or scientists) to carry measuring objects to the place where the echo chamber is located. However, it is difficult to perform measurement on “immovable” embedded speakers, which cannot be moved with ease, in the echo chamber.
The aforementioned problem can be eliminated if an impulse response (hereinafter referred to as anechoic-chamber-equivalent impulse response), which is equivalent to an impulse response measured in the anechoic chamber, can be obtained from an impulse response actually measured in the echo chamber. The paper of Japanese Patent Application, Publication No. Hei 6-324689 discloses a method that produces the anechoic-chamber-equivalent impulse response from the impulse response of the echo chamber. This method produces the anechoic-chamber-equivalent impulse response by eliminating reverberation components from the impulse response actually measured in the echo chamber. This is actualized by extracting a part of a waveform of the impulse response prior to a first echo, such as a part of the waveform measured for two milli-seconds or so after generation of a test pulse.
There is another method that uses kepstrum analysis. In general, reverberation components exist in the impulse responses measured in the echo chambers. As a result, comb characteristics emerge in transfer characteristics. In the comb characteristic, peak dips emerge on the transfer function on its axis of frequency at equal intervals. Kepstrum process is a mathematical calculation method to analyze, detect and eliminate components of notable periods in the characteristic. Concretely speaking, the kepstrum process uses results of Fourier transform, which is performed on logarithmic power spectrum of the transfer characteristic.
The aforementioned method that extracts and uses the part of the waveform of the measured impulse response prior to the first echo is disadvantageous because the characteristics below the frequency of 500 Hz (½ milli-second in the time domain) become unclear, so impulse response cannot be accurately measured with respect to the measured object.
In addition, the method using the kepstrum analysis is computationally complicated and therefore has a drawback that the processing time is long.
SUMMARY OF THE INVENTION
It is an object of the invention to provide an impulse response measuring method that has a high accuracy in calculations of anechoic-chamber-equivalent impulse responses based on impulse responses measured in echo chambers.
An impulse response measuring method of this invention uses an echo chamber in which a speaker and a microphone are arranged. The speaker produces measured sounds, used for measurement of impulse response, which are detected by the microphone. The microphone produces sound signals corresponding to the measured sounds. Impulse response of the echo chamber is calculated based on the sound signals with respect to an axis of time. Then, the impulse response is subjected to Fourier transform to produce complex data consisting of real parts and imaginary parts with respect to an axis of frequency. The real parts and imaginary parts of the complex data are respectively subjected to smoothing processes to produce average data, which are combined together to form smoothed complex data. The smoothed complex data are subjected to inverse Fourier transform to produce anechoic-chamber-equivalent impulse response, which is a simulated impulse response of an anechoic chamber substantially corresponding to the impulse response actually measured in the echo chamber.
If the Fourier transform is directly performed on the impulse response of the echo chamber, which has a characteristic with respect to the axis of time, to produce a transfer function, a characteristic containing reverberation components emerges with respect to the axis of frequency. So, the transfer function has a so-called comb characteristic. This invention is characterized by eliminating or reducing such a comb characteristic by smoothing the complex data, which are obtained from the impulse response of the echo chamber. That is, the impulse response of the echo chamber is converted to the complex data with respect to the axis of frequency, so the complex data are smoothed and are then re-converted to restore the original characteristic with respect to the axis of time. As compared with the conventional method, this invention does not require limitation of a time range used for sampling of impulse response. Therefore, it is possible to maintain characteristics of low-frequency components, so it is possible to obtain anechoic-chamber-equivalent impulse response with accuracy.
The smoothing processes are performed with respect to a prescribed frequency range on the axis of frequency, in which a number of frequency points are set at equal intervals, each of which is set in response to a constant frequency difference or constant frequency ratio, for example. Herein, average data are produced respectively with respect to the real parts and imaginary parts of the complex data belonging to a prescribed bandwidth which is set before and after each one frequency point to include multiple frequency points. Thus, the average data are substituted for original data of each one frequency point. Such averaging calculations and substitution are performed with respect to each of the frequency points in the prescribed frequency range.
Mathematically, the smoothing processes independently performed on the real parts and imaginary parts of the complex data are equivalent to the averaging process of the complex data. However, in an aspect of a configuration of the processing circuit(s) and an architecture of the program(s), the aforementioned “independent” smoothing processes are useful because the system configuration can be simplified by repeatedly using the same processing structure or same processing routine. Incidentally, the aforementioned bandwidth can be set such that each one frequency point whose data are substituted substantially corresponds to a center frequency of the bandwidth.
In addition, it is possible to change a bandwidth located in proximity to a lowest frequency of the frequency range used for the calculations of the average data. That is, it is possible to narrow such a bandwidth whose lower-limit frequency is lower than the lowest frequency of the frequency range such that the lower-limit frequency of the bandwidth is fixed at the lowest frequency of the frequency range or a frequency point slightly higher than the lowest frequency of the frequency range, while each one frequency point whose data are substituted substantially coincides with a center frequency of the bandwidth. Further, it is possible to change a bandwidth located in proximity to a highest frequency of the frequency range. That is, it is possible to narrow such a bandwidth whose upper-limit frequency is higher than the highest frequency of the frequency range such that the upper-limit frequency of the bandwidth is fixed at the highest frequency of the frequency range or a frequency point slightly lower than the highest frequency of the frequency range, while each one frequency point whose data are substituted substantially coincides with a center frequency of the bandwidth. By narrowing the aforementioned bandwidths, it is possible to prevent “meaningless” data out of the frequency range from being used in the calculations f
Hoff Marc S.
Pillsbury & Winthrop LLP
Raymond Edward
Yamaha Corporation
LandOfFree
Impulse response measuring method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Impulse response measuring method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Impulse response measuring method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2890739