主题：[求助]关于开孔式音箱的近场测试

ERIC:

你今天也去听课啦?

个人理解应是相加,虽然倒相管所起的作用是倒响,但它是将低音扬声器背向辐射的声音反了一个相位,对正面而言是负负得正.

至于倒相管在箱体侧面或后面,需加一个距离修正(即相位修正),全部换算到远场测试,先相加再与远场测试的曲线拼接.MLSSA测试就是如此处理的,含无源辐射器.

不知LMS是否测试信号和规定不同?可实际验证一下,哪种说法更准确.

MLSSA:

Center the microphone and position it as close to the
diaphragm or dust cap as possible without touching. In the case of open ports, position
the microphone's front face flush with the enclosure surface.

Make the first measurement by executing a Go Once command. Use FFT Execute to
compute the near-field transfer function of the port or driver in question. Save the
transfer function to an FRQ data file by executing Transfer Save from the frequency
domain. Use Go Once from the frequency domain to repeat this procedure for all ports,
drivers or passive radiators. Each measurement must be saved to a separate FRQ file.
A good idea is to use filenames that start with the same prefix and use a suffix that
clearly identifies the port or driver in question later on when it is time to combine the
files.
Measure the diameter of each port, driver or passive radiator and record these data on
a piece of paper for future reference. The diameters of the drivers should be measured
to include about 1/3 of the surround on each side.
If any port or radiator is not circular, you can alternatively use the square root of its area
to compute its effective diameter. For example, assume a loudspeaker having an
elliptical 6" by 9" passive radiator. These are nominal dimensions but in practice, you
should actually measure the diameters involved including about 1/3 of the surround on
each side. The area of an ellipse is (πab)/4 where a and b are the minor and major
diameter respectively. This area of this passive radiator is then 42.41 sq inches and
thus its effective diameter is √(42.41) = 6.51 inches.
When radiators or ports are mounted in vastly different planes further corrections are
required as explained below.

To properly combine all the near-field measurements execute Library Operations
Frequency-files Average Weighted Complex. Enter the directory where the near-field
measurement files reside followed by the file prefix you selected or just the file prefix if
they reside in the current directory. MLSSA will list all the FRQ files beginning with the
specified prefix. Use Spacebar or the left mouse button to mark all the near-field FRQ
files you just measured. Press the Ctrl-Enter key combination or the right mouse button
to exit the filebox. MLSSA will next display each filename in turn and prompt you to
enter the weighting factor for each. Enter these as the diameters you had previously
recorded making sure each corresponds to the correct filename. Finally, enter an
output filename to save the composite near-field low frequency response.
The resulting anechoic low frequency response can either be spliced to a far-field
Transfer-Function or Sensitivity measurement using the Library Operations Frequencyfiles
spLice command or, displayed as an offset overlay on top of the far-field anechoic
response.
The Library Operations Frequency-files spLice command will prompt you for the splice
frequency in Hz. If you enter nothing and press the Enter key, MLSSA will select a
splice frequency determined by the true resolution of the high frequency far-field
measurement file. This command adjusts the decibel level and phase delay of the low
frequency near-field file to smoothly connect it to the high frequency far-field file. The
output file contains the spliced measurements, which can be loaded into the frequency
domain and displayed through Transfer Load.
In case a port or woofer is mounted in a vastly different plane than the others, you need
to account for the acoustical delay between it and them. This is most easily
accomplished in the frequency domain through the View Phase Delay command.
Suppose a passive radiator is mounted at the rear of the enclosure with the woofer in
front and, suppose further that the enclosure is 2 feet deep. This requires that a delay
of about 2 ms be added to the passive radiator output before combining it with the other
radiators. To do this, load the rear port's FRQ file and execute View Phase Delay and
enter -2 ms. Next save the result to the original file and answer 'yes' when asked to
apply the phase delay correction to the saved data. You can now combine the FRQ
files according to the procedure given above.
Here are some points to remember when making near-field measurements. Near-field
loudspeaker measurements correctly predict the true anechoic far-field response only in
the driver's piston range of operation below roughly f = 4290/d Hz where d is largest
dimension of the largest diaphragm or port in inches.
Another limitation of the near-field method is that it does not account for sound
diffracted from the enclosure edges. Diffraction will affect the desired far-field response
but will not affect the measured near-field response. While this is a theoretical
limitation, in practice, diffraction effects from typical enclosures are small below about
250 Hz because the wavelengths involved at such low frequencies are typically much
larger than the enclosure dimensions.

MLSSA附加说明:

Here are some points to remember when making near-field measurements. Near-field
loudspeaker measurements correctly predict the true anechoic far-field response only in
the driver's piston range of operation below roughly f = 4290/d Hz where d is largest
dimension of the largest diaphragm or port in inches.
Another limitation of the near-field method is that it does not account for sound
diffracted from the enclosure edges. Diffraction will affect the desired far-field response
but will not affect the measured near-field response. While this is a theoretical
limitation, in practice, diffraction effects from typical enclosures are small below about
250 Hz because the wavelengths involved at such low frequencies are typically much
larger than the enclosure dimensions.
In the case of unusually large enclosures, cognizance should be taken of possible
errors due to diffraction effects. Since diffraction tends to increase loudspeaker output
in the forward direction with increasing frequency, diffraction errors will typically result in
a near-field response that is lower than actual far-field response near the splice
frequency. The effects of diffraction are compensated near the splice frequency,
however, by the spLice command, which automatically adjusts the level of the near-field
measurement to match the level of the far-field measurement at the splice frequency.
However, since diffraction has less effect at frequencies below the splice frequency its
net effect, if present, will be to cause the bass roll-off rate of the loudspeaker to look a
little better than its actual far-field anechoic roll-off rate. In other words, a loudspeaker's
bass response will appear more extended than reality when measured with the near-

field method if the enclosure dimensions are large enough to cause significant
diffraction in the region at and below the selected splice frequency. Diffraction effects
can be minimized by measuring the far-field response of a large loudspeaker in a larger
room elevated on a stand in order to achieve a lower valid far-field frequency limit and
hence allow a lower splice frequency to the near-field measurement.

此主题相关图片如下：

是"叠加"--矢量(严格讲是复变量)相加.

假如有些系统只能测到声压再根据自己的算法算出最小相位(不是真实相位),得到的相位是相反时,那就只能相减运算了.LMS系统就是这样.

多单元低频近场测试,有一个问题就是LMS所说怎样将多单元相互干扰减少的问题,还有一个就是MLSSA提到的边角反射和衍射问题.

只是近似和模拟测量,有大消声室远场测试当然更权威和可信.也可考虑用半空间或其它方法测试.

还有一个简便和实用的远近场转换方法,即若想在200Hz接合,直接读出远近场在200Hz的差值,然后减掉就行.DAAS就可直接接合,它能自动计算自动减掉差值并接合好远近场曲线.

DAAS不能进行加减乘除等运算,也不能进行SPL的偏移处理,所以不能用多单元低频组合的方法.