The most popular ways to improve the audio quality

  • Detail

A variety of ways to improve the audio quality of mobile device speakers

in the past decade or so, mobile devices with micro speakers have shown explosive growth -- MP3, GPS systems, laptops and laptops, tablets, game consoles, toys, and so on. Due to the improvement of consumers' taste, the demand for higher quality sound playback of these devices is also growing, which brings more challenges to product manufacturers: making small, light and cheap speakers produce more pleasant and high-quality sound. This paper will discuss some methods adopted by portable electronic product manufacturers to achieve these goals

equalization circuit

an ideal loudspeaker (of any size) should have a "flat" frequency response -- that is, the loudspeaker can transmit sound to the surrounding space at the same volume level in the frequency range from 20Hz to 20kHz without peak or trough in amplitude. In practice, no speaker can do this, and the more efforts are made to flatten the response of the speaker, the higher the cost and complexity involved. Due to the low cost and simple structure of the speakers used in portable devices, they cannot contain the complex structure required to achieve a truly flat response. Therefore, there are very obvious output level changes in the audio frequency band - the most significant when the frequency is lower than hundreds of Hertz (Hz)

Figure 1: loudspeaker frequency response chart

in order to compensate for the imbalance, the spectral response of the loudspeaker can be measured and its characteristics can be found out, and then compensated by using equalization or filter circuits. The frequency response of these circuits can compensate for the imbalance of the loudspeaker response. That is, in some frequency bands where the speaker weakens the audio signal, the equalization circuit can enhance the signal proportionally. Accordingly, the equalization circuit can flatten the signal in the sound peak region of the speaker response. As a result, the speaker perceivable output achieves a high flatness

there are at least two disadvantages to achieve equalization. First, it increases the complexity of the system. The more unbalanced the speaker response, the more equalization schemes involved. DSP can be effectively used to realize the equalization curve, but at least it will bring some costs in silicon area and power consumption. Second, in the entire audio spectrum, even if the equalization circuit is added, the physical limitations of the speaker will prevent the flatness from being achieved. Single element transducers such as loudspeakers, whose effective drive unit diameter is usually 1 inch or less, cannot provide useful or identifiable audio energy in the entire audio frequency band (this is why 2-way and 3-way loudspeakers are widely used in home and car stereo devices and public address systems), which is particularly prominent in low-frequency areas. Such a small diaphragm can not effectively transmit low-frequency energy to the air. Trying to compensate by increasing the amplitude of low-frequency signal will cause the speaker to exceed its physical and thermal limits. Therefore, even if there is an equalization circuit, the low-frequency or low-frequency response in portable electronic devices is generally insufficient

synthetic bass enhancement

as mentioned above, the micro loudspeaker in the handheld device has insufficient ability to transmit low-frequency signals, so the bass part in the audio program data will be lost. The current practice is to integrate elements into the sound to make the low-frequency part appear to exist. This method generates overtones from the low-frequency that cannot be emitted by the small speakers (overtones exist in the range that the speakers can indeed provide), and inserts them into the audio stream to deceive the human auditory system. There are at least two human auditory principles that make this possible. One is called "fundamental frequency loss" or "residual tone", and the other is called "difference tone"

figure 2:fo is missing but "implied" by 2fo and 3fo

as for the difference note, the bass note is simulated by mixing an octave higher than the target note with a 5th tone higher than that octave (see Figure 2). For example, if you need to synthesize a C tone that is 3 octaves lower than the midrange C, you can mix a C tone that is 2 octaves lower than the midrange C with a G tone that is only higher than the latter. (228) the experimental results should be rounded down to 1MPa. This practice of using differential tones has been incorporated into wind instruments and electronic instruments for many years to avoid the need for long wind instruments or large speaker systems. As for the loss of fundamental frequency, a series of overtones naturally generated from instrument notes can "hint" the fundamental to the human ear, even if the fundamental is lost (see Figure 3)

figure 3 Fo is lost but implied by its harmonic signal

either of these two methods can be used to generate tones in the passband of small speakers, which can imply or synthesize notes lower than the frequency range that the speakers can actually transmit, thus enhancing the low-frequency response of the surface. In order to achieve this goal, the low-end audio spectrum should be separated from the main signal channel, and nonlinear processing should be applied to generate the above overtones. The resulting synthetic bass is reintroduced into the signal channel and fed back to the speakers. The disadvantages of this method include unpredictable results or audible synthetic signs, and nonlinear processing from complex program data or highly dynamic sound sources (such as pulse sound)

compression circuit

nowadays, the miniature speakers used in portable electronic products are not only limited by the frequency range, but also by the absolute loudness. The limitation of loudness involves not only the small size of the vibrating element that transmits energy coupling into the air, but also the maximum allowable movement or deviation of the element; It shall not exceed the physical limit or damage the suspension device. One of the ways to improve the average perceived loudness of sound without over expanding or damaging the loudspeaker is to use the compression technology. The compression circuit continuously monitors the instantaneous loudness of the audio signal, increases the gain of the mute channel, and more or less retains the audio data with high loudness. These processes are completed on a very fast basis, with fairly smooth compression characteristics according to the loudness envelope of the program source

figure 4 Dynamic range compression. (input, output, compressed output, uncompressed output, inflection point, maximum lifting amount)

Figure 4 shows that the loudness of the smooth channel has been greatly increased, while the maximum output remains unchanged (intersection of compressed and compressed curves) to prevent possible over driving of the system. The average surface loudness after compression is substantially higher than that of the uncompressed signal. The compression ratio above the inflection point is about 2:1 (that is, the 2dB change in the input signal only brings about the 1dB change in the compressed output signal). The compression ratio of the part below the inflection point is 1:1, and the maximum lifting amount is set, so the overall gain requirement of the compression circuit is reduced, and a considerable "lifting" is still given to the relatively flat signal

auditory enhancement (high harmonic enhancement)

decades ago, efficient studio equipment could "stimulate" the auditory effect of music. The goal was to add color to the sound characteristics only by increasing the gain (amplifying the treble) aligned with the high-end frequency level. As discussed in the synthetic bass enhancement section, some distortion of the raw material may cause the human hearing system to perceive pleasure in reality, which is used to enhance tone. In particular, we can import very soft even harmonics, that is, we can add a lot of "warmth" to the amplified music sound. In this respect, the role of electron tubes is well known

as for auditory enhancement, only the high end of the audio spectrum (such as 1kHz and above) is separated from the signal channel, generating even harmonics and included in the controlled total volume, and the resulting correction signal is combined into the audio stream again in a volume adjustable manner. This effect adds a "hissing" or "crystal" feature to many pleasant sounds, depending on the material being listened to. And because this effect occurs in the middle to high-end frequency range, the human ear is more sensitive in this range, and music programs seem to become louder

flexible clipping

many portable audio devices contain technologies that can prevent the audio amplifier from over driving or allow saturation clipping, which may damage the speaker and at least produce an annoying cracking sound. Even with this protection, however, the audio level will still exceed the output range of the amplifier. A method to mitigate the consequences of sound saturation is to use flexible clipping, which can sense when the output voltage of the amplifier is close to its limit (above and below the limit), and trim the waveform to prevent sharp peaks from serious breakdown. This technology reduces the high-frequency energy that may be generated by flat top or sharply cut output waveforms, reduces the annoying popping sound effect and reduces the excessive high-frequency energy that may be transmitted to speakers

figure 5 Flexible clipping

speaker protection

people make every effort to maximize the perceived loudness emitted by portable device speakers. Care must be taken to avoid damage to the speakers themselves. These small transducers can only withstand such a large limited volume. There are two main aspects of speaker protection - maximum film offset and maximum voice coil temperature

figure 6 Loudspeaker profile

Figure 6 shows a typical loudspeaker profile, which can clearly see the physical limit of film movement, especially in the downward direction. The audio signal shall not be too strong, otherwise it will cause the vibrating element to contact the fixed basin frame assembly, or cause excessive tension of the suspension material (ring or spring frame). In addition, the RMS value of the audio signal should not be too large, otherwise the voice coil will overheat. Overheating of voice coil will deform the circle of coil tube and cause friction with the edge of magnet or pole piece. Moreover, the high temperature in the voice coil will also lead to the deterioration of its electrical insulation performance, and finally lead to the short circuit of the wire turns of the voice coil, thus reducing the voice coil impedance and overloading the amplifier. Too high voice coil temperature will also heat the permanent magnet, which may lead to its demagnetization

techniques used to prevent speaker damage include: automatic gain control (AGC) for input signal amplitude and/or power supply voltage, dynamic range compression (as described above), hard clipping, flexible clipping, and amplifier output Over-test. The disadvantage of these technologies is that they are feedforward methods, which can not sense the actual speaker basin offset, voice coil temperature, or speaker impedance (which changes proportionally with temperature). More complex protection mechanisms such as thermal feedback are expected to be implemented in the future, but the current conventional method is one or more of the above-mentioned protection mechanisms

Copyright © 2011 JIN SHI