From the standpoint of the actual application of the condenser microphone, an evaluating factor may be suitable as the so-called, signal-to-noise ratio. The relation between the optimum condition for sensitivity ( Fig. 2 ) and the optimum condition for the signal-to-noise ratio ( Fig. 9 ) was sought after. As a result of the above, it has been made clear that the optimum condition for sensitivity can be obtained when the distance between the diaphragm and the back electrode is widened until the self-capactance of the microphone and stray capacitance are equal, however, the signal-to-noise ratio is a maximum under different conditions. The above is due to the fact that the noise in a condenser microphone circuit is dependent on the electrostatic capacitance and circuit resistance. If the noise in the amplifier can be disregarded, the signal-to-noise ratio improves when the self-capacitance is large or when the distance between the diaphragm and the back electrode is small and, since it is proportional to the square root of the circuit resistance, it is possible to design in accordance with the balance between the costs of the formation of the distance between the diaphragm and the back electrode and that of the resistor. When the noise of the amplifier cannot be ignored, the signal-to-noise ratio has, with respect to the diaphragm electrode distance, the characteristics of Fig. 9. When the distance between the diaphragm electrode and the back electrode d, back electode area S_b, circuit resistance R, stray capacity C_s and center frequency f_0 and noise level N_<( f_0 , 1/3 )> in a 1/3 octave band, the signal-to-noise ratio of a condenser microphone becomes the maximum value when the distance between the diaphragm and the back electrode is equal to ε_0 S_b / ( ( 9. 65 × 10^<-12> / N_( f_0 , 1/3) √< R f_0 > ) - C_s ). The above occurs when the amplifier noise and resistor noise are equal. From the data cited above when exterior specifications are given, it is possible to design a microphone based on the signal-to-noise ratio.
Due to the requirements in television broadcasting and other fields, it is necessary to miniaturize the condenser microphone. In this case, however, the designing standard, for the signal-to-noise ratio must be made clear. In this paper some designing method, and the permissible limit of the signal-to-noise ratio using the noise level of a broadcasting studio as reference and based on the condition that and increase of the above level is not detected during reception, are described. The signal sound pressure in the original sound field and the signal sound pressure at the time of reproduction are practically equal. Based on this, the sensation level of the noise ( Fig. 4a ) at the original sound field is obtained and the allowable limit of the signal-to-noise ratio was determined according to the conditions described above. ( Fig. 4b ) On the other hand, when the formula which express the signal-to-noise ratio of the condenser microphone and spectrum of the noise which forms the above are noted, since the noise of the condenser microphone has the same form as the noise formula obtained from the allowable limit, Fig. 5 and Fig. 6 have been obtained as the designing conditions of the microphone. The result of a preliminary calculation as a concrete example is shown in Fig. 7 and Fig. 8. The fair values which can be used in actual practice were obtained. Based on the above concept, three microphones were trial manufactured and each microphone had sufficiently high performance so that it may be used, with satisfaction, in broadcasting. ( Table 1 )
In this paper a study of wide-band high-reliability standard condenser microphone is described. Frequency band of the condenser microphone is limited by the first mode resonance frequency of the membrane and given by the following equation: f_1 = (α_1)/(2πa )√< T/ρ > where α_1 : eigen value for the first mode, a : radius of membrane, T : tension of membrane ρ: density of membrane To obtain the wide frequency band, it is necessary to use a membrane material having a large value of the ratio of ultimate tensile strength to density. As a result of the study about several kinds of materials, we found that titanium alloy( 2Al-2Mn-1Mo-95Ti ) is the most suitable one for the membrane material. This alloy is possible to roll up the thickness of 3μm, and after cold rolling an ultimate tensile strength of about 100 kg/mm^2 is obtainable. Density of the material is 4. 52g/cm^3. Furthermore, this material has a special feature that the instability due to the creep after tension is avoided by applying thermal aging 200°C for 5 hours similarly as in metallic titanium. This feature is shown in Figure3. Titanium alloy has sufficient values of Young's modules and hardness, so that it is also suitable as the body material of a condenser microphone. The structure of a newly designed standard condenser microphone named MR-112 is shoun in Figure 5. Tension of the membrance is impressed by a non-rotating ring, and structure contributes to the improvement of reliability. Pressure response of the microphone is shown in Figure8. Followings are the main characteristics of the condenser microphone of MR-112, i. e. , the response at 1 kHz is 54. 5 dB ( 0dB : 1 V/1 μbar ), the resonance frequency is 30 kHz, terminal electric capacitance is 22 pF, the equivalent air volume of membrane is 0. 01 cm^3, and the stability factor of membrane is 7. Value of the free field correction of the microphone used with a cylindrical head amplifier of semi-infinite length are shown in Figure 12. The condenser microphone of MR-112 has a high reliability as in MR-103, as well as wide band characteristics, then it is useful as a standard microphone, especially in the high frequency range.
The conventional measuring method of nonlinear distortion using one or two tones of a fixed level does not always provide an objective value of distortion which agrees with the subjective sound quality of speech or music signal passing through the apparatus or the system to be measured. To find a better agreement, the distortion produced in the actual performance of the system should be obtained, and the distortion components appearing in various frequency bands should be extracted separately. A measuring method that meets these requirements is proposed in this paper. Fig. 1 shows the principle. As a test signal, a wide-band signal whose spectrum is similar to that of speech or music is used. The test signal is applied to the system to be measured after removing the narrow band components with band-elimination filter. Distortion components that appear in the eliminated band are extracted with an appropriate band-pass filter having the same frequency as that of the band-elimination filter. In this way the distortion component appearing in a specified frequency band is extracted, and the whole spectrum of distortion is also obtained by using pairs of filters of various central frequencies. Perceptibility of nonlinear distortion for three kinds of typical program sound was obtained as a function of frequency of distortion components using a distortion circuit having an input-output characteristics as shown in Fig. 6. The results are shown in Fig. 11, from which it is clearly seen that the distortion appearing in the high frequency band is much noticeable. Results of this measurement and a few additional experiments were discussed from the viewpoint of masking, and it was shown that the perceptibility of distortion components in a specified frequency band is mainly determined by the signal level in the same and the neibouring frequency bands, and that the just perceptible level of distortion is almost constant when the level is given relative to the signal level of the same frequency band as far as a fixed program signal is concerned.