Shenzhen Xuanda Electronics Co., Ltd.

Shenzhen Xuanda Electronics Co., Ltd.

How does the headphone achieve noise reduction

2021 06/17



The noise reduction design of headphones can be divided into active noise reduction and passive noise reduction


Passive noise reduction is to cover your ears with some sound-absorbing materials and block them tightly to reduce noise

we try coc nut, bow, cloth, cotton, not big good use.

However, the commonly used sound-absorbing materials have excellent effects in the high-frequency range, and have general effects in the low-frequency range, and passive noise reduction is not satisfactory. The emergence of active noise reduction makes up for the lack of passive noise reduction. Active noise reduction, also known as ANC (Active Noise Cancellation) noise reduction, its principle is very simple. The noise signal is collected through the microphone and sent to the DSP chip for data calculation, generating a signal equal in magnitude to the noise but opposite in phase, and the two signals overlap Then, as soon as it is cancelled, the noise is gone.



Is it really that simple?
It is important to know that different components of the headset will affect the frequency response and phase response. The sound curve collected by the microphone is different from the sound curve heard by the ear. If the curve collected by the microphone is directly inverted, the noise reduction requirement cannot be met. Yes, so how can we get the ideal ANC filter curve?
The answer is simulation experiment.


We use coaxial speakers to emit 20Hz-22kHz frequency sweeping sound waves as the noise source, and use the internal microphone of the human ear simulator to replace the human ear. And measure the sound waves reaching the human ear. In this way, we measure the noise attenuation caused by the earphone itself, which is recorded as A1(f). At the same time, the sound wave reaching the microphone is measured, denoted as A2 (f), and the last is the sound wave from the internal speaker of the headset to the human ear.


The frequency sweep signal from 20 Hz to 22 kHz is emitted by the speaker inside the earphone, and the signal is collected by the microphone in the artificial head used for the test, which is recorded as the function A3(f). With these three measured values, we can calculate the ideal value of the ANC filter according to the formula A_filter= A1(f)-( A2(f) + A3(f)), so as to achieve a more perfect noise reduction effect.


Such noise reduction methods are called feedforward noise reduction.
In reality, because the location of the noise source cannot be determined, and the difference between each person`s ears, these will affect the final value of the filtering curve. The noise is not complete.


You might say that if it is so complicated, it is not enough to directly collect the noise signal in the ear canal. That's right, feedback noise reduction is to place the microphone in the earphone close to the ear canal and collect the noise signal from the ear canal for processing. Feedback noise reduction does not need to care about the direction of the noise source, but it needs to pay more attention to the delay problem. Because the feedback-type noise reduction microphone is in front of the speaker, even if there is no signal processing time, it is always one beat late.


We know that the low-frequency sound period of 100Hz is 0.01s after conversion according to the formula T=1/f, and the period of the high-frequency signal of 5kHz is 0.0002s.
If there is a 0.0001 second delay in the processing process, the 100Hz inverted low frequency signal will be delayed by 1/100 cycles, which may not feel anything;


The 5kHz reverse high-frequency signal will be delayed by half a cycle, and it will be superimposed with the original signal to perfectly become an active noise enhancement!


Therefore, the lower the noise frequency, the slower the waveform changes, and the better the noise reduction effect. The feedforward noise reduction is because the microphone is outside, as long as the Ref mic receives the environmental noise + signal processing time <Error mic receives the environmental noise time , Then a relatively perfect noise reduction can be performed. The double-fed noise reduction combines the advantages of the two, and the noise reduction effect is good, but the cost and technical threshold are high. So, what is the effect of active noise reduction? Let's test it with noise-canceling headphones. A signal generator that can emit signals in the full frequency range is connected to the speaker to amplify and become a noise source.