Technical testing of speakers and subjective sound quality evaluation

In the sound system, the speaker has the simplest structure, but plays the most important role. The speaker is the final link in the sound system and is the key device that converts electrical signals into sound signals. Since the speaker has few structural components, any deficiency will affect the entire playback effect. However, the evaluation of speakers is generally based on subjective listening - the quality of the final music playback. Therefore, the following first problem arises: this is the problem of unifying the technical testing of speakers and the subjective sound quality evaluation.

1. About the measurement and listening of speakers

Friends who have a certain amount of HiFi experience will have a consensus: there is a gap between the technical indicators of the speakers and the subjective listening experience.

Why is this the case? We need to start from the beginning and talk about the main technical indicators of current speaker testing.

Among the technical indicators of speakers, the most important ones are: frequency range, power handling, sensitivity, phase characteristics and transient characteristics.

Frequency range, also known as frequency response, refers to the range of reproduction from bass to treble in a speaker. Sometimes, the frequency response ranges of a pair of small consumer speakers and a pair of large monitor speakers may be the same; but the subjective listening experience is completely different. Why is this? Is it because the test is inaccurate or someone deliberately falsifies it? Under normal circumstances, the above speculations are not correct, but there are other reasons.

Reason 1: The gap between current test standards and actual applications

The current frequency response test standard still uses the 1 watt per meter standard from many years ago. When the audio test standard was formulated many years ago, the speaker frequency response curve at 1 watt per meter often represented a full-power test, which was close to the actual use. The speakers of that era were basically low-power and high-sensitivity. For the input 1W test power, it basically represents the actual working state of the low-power speaker. In recent years, with the development of speaker manufacturing technology, most speakers have become high-power and low-sensitivity. This makes the current frequency response test increasingly far away from actual use.

At present, most speakers (mainly home speakers) have a power handling capacity of more than 50W, and some even reach 200W, while the sensitivity can only reach 82-86dB.

In this case, if the 1-watt frequency response test is still used, the test will not match the actual use. The 1-watt test signal is only equivalent to 1/50 to 1/200 of the power of most speakers, while the actual power of these speakers is generally 15 to 50W, and even reaches full power at peak, which leads to the problem that the test itself deviates from the actual situation.

An accurate and feasible measurement method should try to simulate the actual working state of the equipment being tested, so that the test results will be most effective.

Therefore, when testing the frequency response of a speaker, if a half-power frequency response test can be added, then this half-power frequency response curve will have the most realistic impact on the actual operation of the speaker.

For example, a small speaker with a nominal power of 200W, when tested with a 100W half-power test signal, you will see serious degradation in the low frequency and a significant increase in distortion.

Why is this so? This is because the bass speaker unit is restricted by the equivalent volume of the speaker and is affected by factors such as the bass reflex tube, as well as the increase in the resonance of the box itself, which will inevitably produce the above results.

Testing the speakers with a low-power signal of 1W is equivalent to an unprofessional person singing opera in a low voice. It is not the actual working condition. The real actual working condition is to sing loudly on the stage. Therefore, the beauty of a small voice and the performance on the stage are completely different concepts and are two different things.

Therefore, adding half-power frequency response tests to modern speakers not only has very practical guiding significance, but also has the effect of exposing counterfeits - it can completely expose the true colors of those small speakers with a nominal power of 2000W.

Reason 2: Current testing standards are too loose

There is another reason for the inconsistency between subjective listening and frequency response testing, which is that the current testing indicators are too loose and not strict enough.

The current speaker frequency response test uses -3dB as the standard. In other words, the frequency at which the sound emitted from the high-frequency end and the low-frequency end is attenuated by half is used as the final frequency for measurement.

As the manufacturing technology of tweeters continues to mature, a mid-range or higher tweeter can easily produce 20kHz high frequencies. The world's most advanced silk film soft dome tweeters have reached a high frequency of 40 kHz, while the upper limit of the high frequency of metal film tweeters has reached 80 kHz. In addition, the tweeter is less affected by the speaker cabinet, so the new mid-range or higher speakers will not have high frequency problems. The factors that have a greater impact on high-frequency reproduction are limited to the choice of the low-end crossover point of the tweeter, whether the connection with the midrange unit is excellent, and whether there is enough power margin.

For woofers, the situation is much more complicated. Although the manufacturing technology of woofers has also made great progress, especially the continuous emergence of new diaphragm materials, the technical indicators of woofers are already good. However, the gap between the test indicators of woofers and the actual indicators after they are installed on the speakers is very obvious. Sometimes, an excellent woofer, installed on a speaker box with many problems, will completely lose its original advantages.

For example, a poorly designed cabinet will produce serious mid- and low-frequency resonances, which will not only cause sound coloration when the speaker is playing back, but will also cause the test curve to deteriorate.

The test indicators of large monitor speakers are relatively "accurate". In fact, the word "accurate" here does not mean correct, but it means that the test indicators are closer to subjective listening. Since the current 1 watt 1 meter test method has shortcomings, which speakers within the current test standard range have the closest technical tests to subjective listening results? The answer is as follows: these are large monitor speakers and large civilian speakers of relatively good quality.

Why is the result a large speaker? There are three reasons. First, the volume of the cabinet of a large speaker is large, so the low-frequency response in actual work is less restricted by the volume of the cabinet. Second, the power handling power of a large speaker is large, and the power margin is large, so the possibility of overload distortion during work is relatively small. Especially for large monitoring speakers, the power margin is very large, and there is no possibility of overload distortion at all. Third, the woofer of a large speaker is relatively large. When producing the same bass, the stroke of the unit is very small, and the distortion is relatively small.