Editing Sound for ALPHA IC AM4E
1. Features
1. Sound type: Little boy voice (tell story) + Background music
2. Speaker Specification: Large magnetic speaker with 29mm diameter, 8Ω resistance, 0.25 W power, and plastic film.
3. Output type: PWM
4. Objective: Stand out the
little boy voice and filter the low frequency of background music.
2. Basic Operation Steps
1. Resample.
2. Use FFT filter to adjust.
3. Editing Sound
Run Cool Edit v2.1 and open a voice file (a1.wav). See the following figure 1.

Figure 1
Then click Analyze->Show Frequency Analysis in toolbar to analyze the frequency of a1.wav. See the figure 2.

Figure 2

Figure 3
In figure 3, you can see that the voice frequency is mainly distributed in the band above 200Hz. While the lower frequency band of background music causes that the speaker cannot respond to the lower band and it turns into the “Si-si” sound. Hence, in order to stand out the human voice, you should raise the frequency band ranging from 200hz to 500Hz. In this way, the human voice is highlighted and the un-responded part is filtered.
Speaker Feature
The used speaker is a large magnetic speaker with 29mm diameter, 8Ω resistance, 0.25 W power, and plastic film. The Frequency Response of this speaker is generally above 200Hz. Note that the speaker frequency responses for different factories are various. And you should filter the frequency band below 200Hz.
Using FFT Filter to Edit Sound
Then we use FFT Filter to edit sound in accordance with the sound analyzed result and the Frequency Response of speaker. The figure 4 shows how to open FFT Filter.

Figure 4
In figure 5, the frequency band ranging from 200Hz to 3KHz is raised properly and the band below 200Hz is filtered. Raising the band (200Hz to 1Khz) can stand out human voice. While to stand out the boy voice you need to further raise the band above 1Khz due to the wider frequency band of boy voice. You can see from the figure 5 that the high frequency band above 1Khz has been raised and the un-responded part is filter. In this way, the “Si-si” noise no longer occurs and the sound is edited.

Figure 5