A "perfect" AC amplifier should have the following qualities in terms of frequency response indicators: it can maintain stable amplification for signals at any frequency and have the same driving ability for the corresponding load. Obviously, this is completely impossible at the current level of technology, so there are different "prefixes" for different amplifiers. For audio signal amplifiers (power amplifiers or small signal amplifiers), we should also add such a "prefix": signals that are within the frequency range audible to the human ear and "may" affect frequencies within that range. This range has obviously narrowed down a lot. We know that the audible frequency range of the human ear is about20～20KHz, that is to say, as long as the amplifier can meet the "standard" for signals within this frequency range.

 

In fact, research shows that although some signals above this frequency band and some signals below this frequency band are "inaudible", they will still have an impact on people's hearing. Therefore, this range needs to be expanded further. In the field of modern audio, this range is usually5～50KHz, some highly demanding amplifiers can even achieve0.1~hundredsKHz。

 

However, the above requirements seem to be much lower than "perfection" on the surface, but they are still "impossible tasks". At present, it is impossible for us to even meet such requirements. Therefore, there is the indicator of "frequency response". (Postscript: Indicators themselves represent "imperfection". If everything is "perfect", indicators will have no reason to exist.)

 

Amplifiers have two types of distortion: linear distortion and nonlinear distortion. We usually call the latter "distortion" and express the former in other ways. We already know what nonlinear distortion is like. Linear distortion refers to the "error" in frequency and phase, that is, frequency distortion and phase distortion.

 

 

Frequency distortion is a kind of "linear distortion", which means that when this distortion occurs, the output signal waveform and the input waveform of the amplifier are still "similar" and will not cause the amplifier to "deform" the signal to be processed. A simple frequency distortion can be seen as the amplifier amplifies different signals at different frequencies, for example,1Ten times amplifier, yes1KHzThe amplification factor of the signal is10Times, and for10KHzthe possible amplification factor of the AC signal becomes9.99Times, so we can say that this amplifier has frequency distortion. In electroacoustics, we call this phenomenon "unevenness of the frequency response curve", and we will talk about the "curve" here later.

 

For an amplifier, there are many reasons for frequency distortion. The inherent characteristics of the multi-amplifier and the multi-amplifier will affect this parameter, and even distortion will intervene (this is caused by the measurement method, which will be discussed later). In general, there are some situations that can cause frequency distortion:

 

1, determined by the natural frequency characteristics of the components, this is the most fundamental reason, and some of the latter reasons actually originate here.

2, the open-loop characteristics of the amplifier using negative feedback technology and the frequency response characteristics of the negative feedback circuit itself.

3, the nonlinear distortion of the amplifier introduces "measurement error" to the measurement method;

4The circuit design of the amplifier leads to non-idealization of the transmission characteristics;

5, the installation and manufacturing processes are imperfect, and the introduction of external AC interference signals leads to uneven frequency response.

 

Speaking of this, we will find that there are many reasons here that seem to be related to the "measurement method", so it is necessary to mention how the frequency response is measured and calibrated.

 

 

Any "indicators" that can be written in instructions must be measured with the help of instruments. These indicators must have a common feature, that is,"repeatability", which means that as long as you use the same equipment, you can repeatedly obtain the same goods. Similar measurement results. We call this type of indicator "objective indicator", and frequency response certainly belongs to this category.

 

The frequency response measurement method is simple:

A standard signal generator is connected to the input of the amplifier. This signal generator can generate a standard sine wave signal, and the frequency of the output signal of this generator can be adjusted to change without changing the amplitude. Connect a standard purely resistive load to the output of the amplifier and an AC level meter. By reading the data from the level meter, the frequency response characteristics of the amplifier can be measured.

 

During measurement, in order to ensure the reliability and accuracy of the test results, it is necessary to select as many different frequencies as possible within the test frequency range. The "logarithmic sampling method" is usually used, that is, from a standard frequency (for example1KHz) Start, follow2The multiple relationship takes points up and down, for example2K、4K、8K……,500、250、125、62.5……, if you think this interval is too large, you can reduce it by multiple, for example√2，√2/2Wait. Compare the output levels of these corresponding frequencies (indB) Just record it and conduct statistical calculations.

 

Here, we may overlook a question, which is, can the amplification factor of this amplifier be adjusted? How much should the output power of the amplifier be used? It's not that I want to keep you in suspense, but that the "mystery" here is very big. Due to the imperfect characteristics of the amplifier, the frequency response characteristics of the amplifier will change under different working conditions. This is called "test conditions". We often find that two amplifiers of completely different quality "seem to have no difference" in frequency response indicators. Is the amplifier with poor quality "lying"? No, it is because the test conditions are fundamentally different.

 

The frequency response of an amplifier is different at different output powers. Generally, the higher the output power, the worse its frequency response index will be. A more responsible indicator label should refer to "the indicator measured at the maximum undistorted power of the amplifier." In order to avoid the degradation of amplifier characteristics under high power output, some manufacturers make the indicator "look good", often adopt a "standard test method", that is, the test is carried out under the conditions of a given amplifier amplification (gain), which is usually1。

 

Obviously, most amplifiers are used to "amplify", so this test method is not actually comprehensive, but this test is still considered "correct""for commercial purposes and as permitted by test standards." In this way, we should pay attention to that when looking at indicators, we should not only care about those numbers, but should be linked to the test conditions. Indicators without test conditions are meaningless.

 

The standard frequency response labeling method isXHz~YHz±ZdB, hereXRefers to the low end frequency,YRefers to the high-end frequency, which is the range of test frequencies,ZRepresents the difference in amplifier amplification over this frequency range.

 

Unfortunately, the frequency response characteristics of this amplifier cannot be fully understood by looking at the Ziega indicator alone, so the manufacturer gave another expression－Frequency response curve.

 

 

The frequency response curve is that in the above-mentioned test circuit, the frequency of the output signal of the signal generator continuously changes (i.e., commonly referred to as "frequency sweep") and keeps the amplitude unchanged. At the output end, the amplifier is recorded at the output end through an oscilloscope or some other recorder. Record the output level corresponding to this continuous change, and you can draw a curve of the level corresponding to the frequency on a coordinate. The ordinate of this coordinate is the level and the abscissa is the frequency. The unit of ordinate isdB, the horizontal target unit isHz(orKHz）。For convenience of recording, the scale for the abscissa is logarithmic, and the ordinate is linear.

 

We can look at the frequency response curves of different equipment provided by various manufacturers, and we will find that even two equipment with exactly the same frequency response indicators have very different frequency response curves. Here, we will not discuss the impact of different frequency response curves on sound quality for the time being, but only look at the important characteristics of the frequency response curve that need to be paid attention to. Two characteristics should be paid attention to here: peace and straight. Flat refers to the maximum difference in the frequency response of an amplifier within the operating frequency range. What we need to pay attention to here is the "working frequency". For audio equipment, we should care about20～20KHzIf the requirements in this paragraph are very high, the scope can be expanded to5～40KHz, this is enough.

 

1When looking at the frequency response curve, don't be confused by the "smoothness" or "rugged" of the curve. First, look at the ruler of the coordinates. Changing the unit of the ruler will make the curve look very different. If you increase the scale10Times, what you are probably seeing is almost a perfect straight line.

 

2"Straight" is another very important feature of the frequency response curve, which refers to the undulating characteristics of the frequency response curve. In a sense, we should pay more attention to "straight" than flat. This is not to say that straight than flat has a greater impact on sound quality, but because the unevenness of the frequency response curve often implies that there are problems with some other characteristics of this device, such as excessive high-frequency ringing volts, which often indicates that the amplifier's open-loop characteristics are poor, and the negative feedback depth is inappropriate, which is usually accompanied by serious transient distortion.

 

We generally believe that the smoother and straighter the frequency response characteristics of an amplifier, the better, so that the amplifier will have less impact on the signal. By looking at the curve, we would think4than5Be good.

 

3Here, what we should also note is that although we need to focus on inspections5～40KHzThis frequency band, but for different equipment, the frequency bands we assess are actually not exactly the same. For example, for slightly and headphones, this frequency band is sufficient, but for some "active equipment"(such asCDphonograph, amplifier), we may need to assess a wider frequency band. This is because for these equipment, although we cannot hear the sounds in these frequency bands, the performance of these frequency bands can reveal some of the inherent qualities of this equipment. For example, for an amplifier, if its frequency response index can be as high as300KHz,And the depth of negative feedback is appropriate, which shows that this amplifier has excellent open-loop performance, which must be reflected in the sense of hearing. In this sense,"we can hear the performance of these frequency bands."

 

 

Frequency distortion has a huge impact on sound quality. In many cases, it will completely affect a person's evaluation of sound quality. Since frequency response has too many factors that affect subjective sound quality evaluation, it is impossible to list them all here. I will just select some aspects that I think have the greatest impact.

 

1, influence on the timbre performance of musical instruments

 

In a broad sense, timbre is also an integral part of sound quality. We know that different musical instruments have different sound characteristics. The interaction of fundamental tones, overtones, and resonances constitutes the timbre characteristics of an instrument. The timbre is the frequency and proportional relationship of these fundamental tones, overtones, and resonances. If a system is not straight enough in terms of frequency response, it may cause the proportion of various components in the timbre to change. Some overtones may be enhanced, while other overtones may be weakened or even difficult to hear, which changes the timbre characteristics of the instrument.

 

Since we often don't have the opportunity to compare the sound of the original instrument, this change is not extremely important. However, since whether the instrument is "good" or not is almost synonymous with timbre, excessive damage to the timbre characteristics may cause the sound of this instrument becomes unpleasant, so it is best not to change the timbre characteristics for people with high requirements. Since the frequency response will have an impact on the tone, some equipment designers will cleverly use this phenomenon to make up for the shortcomings of recording. For sound engineers, this kind of adjustment is also a "common occurrence" because they cannot "invite" those "famous piano" for every record.

 

2Impact on sound field and positioning

 

The sound field is a very complex electro-acoustic phenomenon, and the frequency response characteristics will also affect the performance of the sound field to some extent. Due to the influence of frequency response, some sound details related to the performance of the sound field will be weakened or strengthened, which will lead to the so-called "distortion" of the sound field. This is a very subtle influence that cannot be fully explained in this limited text and discussed later.

 

For positioning, the situation is also very complicated, especially for instruments with a wide frequency range, which has a greater impact. This is easy to understand. The sense of distance is closely related to the size of the sound. If the frequency response is not straight, the instrument will feel farther or closer when emitting a sound of a certain frequency than when emitting other sounds. In this way, we will feel that the instrument has been elongated longitudinally and its shape has changed. When the unevenness of the frequency response is serious, we will feel the instrument shaking back and forth.

 

3, impact on the overall timbre

 

This topic can be very old, so I won't talk about it anymore. The coolness and warmth of equipment, the density and intensity of sound mainly come from this (of course, there are also other factors that will be discussed later).

 

 

We should pay enough attention to the manufacturer's frequency response indicators. But we should also remember that this indicator is not "labeled" as high as possible. Because our ears have some of their own characteristics, we need to have a clear understanding of frequency response.

 

1The frequency response indicators we need should be for the entire system, not a single equipment. The flat frequency response of a single device does not mean that we will definitely hear a "flat" sound. It also depends on the conditions of other devices in the system.

 

2Even when the frequency response of all equipment in the system is straight, we may not be able to hear a straight sound. This is because our ears themselves are not "straight". We know that the sensitivity of the human ear to high frequencies changes over life,20Reaching its peak around the year of age,35Around the age of age, the decline began to occur, and60More than half of the loss will be lost around the age of age, which is also related to physical health and genetics. Therefore, when we consider straightness, we must consider ears together. In this regard, there seems to be a tacit agreement in the industry. This part is mainly completed by the slightly, headphone manufacturer and sound engineer.

 

3. Our ability to distinguish frequency response fluctuations is limited. Experiments have shown that0.2dBIt is the limit of a very small number of people (less than one in hundreds of thousands of people), and the vast majority of people are1~3dBbetween. In other words, less than1dBIt is almost meaningless to have a non-straight frequency response. If you abandon some other elements in order to pursue an excessively straight frequency response, it will be worth the loss. This principle is the same for other indicators.

 

4As mentioned earlier, we cannot ignore certain frequency bands just because we cannot hear them, because those things may imply some other characteristics of the equipment.

 

5Any indicator must be viewed in conjunction with other indicators, and cannot be viewed in isolation.