In the previous discussion, the microphone channel was not correct. (See http://klog.icrr.u-tokyo.ac.jp/osl/?r=13292)
So I reproduce the same analysis (evaluation for acoustic components in accelerometer signal) for the correct channel (K1:PEM-SENSOR_BOOTH_IMC_BNC3)

(1) White noise 60-3000 Hz, DAC amplitude = 2000 counts
 - Background data : from GPS 1266985700, 100sec
 - Injection data :  from GPS 1266985880, 64sec

Projected acoustic noise is almost same with ACC signal. So the ACC signal is occupied by the sound and is is well evaluated. 

(2) White noise 60-3000 Hz, DAC amplitude = 5000 counts
 - Background data : from GPS 1266986785, 64sec
 - Injection data :  from GPS 1266986945, 100sec

We can see many overestimated peaks.They are near-miss of the resonant peaks. [See http://klog.icrr.u-tokyo.ac.jp/osl/?r=12302]