Perhaps everyone was interested in the frequency response graph of their brand new speakers or headphones. But what does it really give? How to read it correctly and how to use it? What influences it, how to measure it yourself, and is it necessary to achieve a perfectly flat frequency response? More on this below.
What is frequency response, how is it measured
The frequency response in audio equipment indicates what volume the device will give at each of the frequencies of the audible spectrum – from 20 Hz to 20 kHz. In English, the term frequency response is used. Visually, it is presented in the form of a graph, where the audible frequency spectrum in hertz is located on the X-axis, and the loudness in decibels on the Y-axis. For example, the picture above shows the frequency response graph of the Sony SS-CS5 speakers.
To measure the frequency response, you need to “feed” the device a set of sinusoids of the same volume across the entire spectrum. There are special generators of such test signals, however, you can even find them in the recording on YouTube . If you run this signal through a speaker located in a room with a flat frequency response, and record the result with a microphone (do I need to say that the entire audio path for recording must have a flat frequency response?), Then you get the same graph. It sounds scary, but in reality everything is not so complicated – for household purposes, the simplest measuring microphone, placed close to the speaker and connected to an audio card, is enough. If there are several speakers in a column, each of them is measured separately.
High fidelity acoustics manufacturers measure frequency response in anechoic chambers using microphones mounted on a robotic arm. It allows you to measure the frequency response at different distances and in different planes in seconds.
The frequency response of the headphones is measured with a dummy, similar to that used to record binaural audio . It is a model of a human head with auricles of average shape, in which microphones are located.
The frequency response of an amplifier or receiver can be measured either using a speaker with a perfectly flat frequency response, or bypassing it – using a dummy load that is plugged directly into the audio card (the amplifier will burn out without load).
What is the spread of the frequency response is considered perceptible by ear
The resulting graph will tell you about the timbre balance of the audio device. As a rule, household loudspeakers do not have a flat frequency response. Humps and peaks in some areas will color the sound.
How audible will it be? If we look at the example of speakers, then only studio monitors will have a relatively flat frequency response. Relatively – because their deviation from a straight line can fluctuate within 2-5 dB in one direction or another. For example, the picture above shows the frequency response of the Adam T5V monitors.
Active Monitor 1.0 ADAM Audio Т5V 18 499 *
Based on this, we can conclude that humps and dips in the frequency response of more than 2–5 dB become clearly audible. But here you need to understand one important thing – this is anti-aliasing. Usually, even for studio monitors, the frequency response graph is more like a cardiogram with small but frequent hills and depressions rather than a straight, even line. However, the human ear does not hear in enough detail to distinguish these irregularities. Therefore, programs that measure frequency response have an anti-aliasing function to bring the graph to the form as perceived by the human ear. It is indicated in fractions of an octave – for example, smoothing of 1 \ 12 octave means that the graph will be in steps of one note.
Smoothing of two frequency response graphs in 1/12 octave (left) and 1/3 octave (right)
Of course, here audio manufacturers see an excellent loophole for tweaks – if you smooth too much, for example, to a whole octave, then the schedule will be deceptively flat. Even if the frequency response graph is attached to the columns, it is very rarely indicated which smoothing was used. Therefore, enthusiasts often measure the frequency response of the most popular audio devices and share the results on forums and social networks.
How the nature of the sound depends on the uniformity of the frequency response
To talk in more detail about the nature of the sound, you need to figure out what the different frequency ranges are responsible for on the frequency response graph. The entire audible spectrum can be divided into low, medium and high frequencies. In turn, each of these ranges (rather conditionally) can also be divided into three parts:
Sub bass: 20-50 Hz. These frequencies are captured not only by the ear, but also by the whole body – there is so much energy in them. To hear sub-bass, you need a subwoofer – household speakers don’t play that low.
Mid-bass: 50-100 Hz. This is where the main bass and kick tones are located. Mid-bass gives basic pressure and pitch.
Upper bass: 100-200 Hz. The upper harmonics of the bass and kick of the kick are located here, these frequencies are responsible for the readability of these instruments.
Low mids: 200-400 Hz. There is a punch of overloaded guitars in rock and synthesizers in electronics, basic harmonics of male vocals, kick of a snare drum.
Average “box” frequencies: 400-800 Hz. They are called boxed because they sound very cardboard, like sticking your head in a wooden box and talking. Often they are cut with an equalizer because of the unpleasant sound. However, the fundamental harmonics of female vocals and some high instruments – violins, flutes, etc. can also be located here.
Upper mids: 800-2000 Hz. This range is responsible for the richness and readability of most instruments and vocals; here lies the grain of electric guitars and the fundamental harmonics of synthesizers.
Present: 2-5 kHz. These frequencies correspond to how close the instruments will be to the listener. If you subtract them, then the track becomes wadded, as if a blanket was thrown over the speaker.
Brightness: 5-10 kHz. Brightness is a comprehensive description of this range. Without it, the music will sound dull, like an old radio.
Air: anything above 10 kHz. Responsible for the openness and transparency of the sound. Especially for cymbals, drum kits, high female vocals, violins and other high instruments.
This information will help determine the nature of the sound of acoustics by its frequency response. For example:
A hump around 100 Hz is common – this makes the bass more punchy.
No less often you can find a hill in the 5 kHz region, so that vocals and instruments sound closer to the listener and better read.
Raising the treble around 10 kHz makes the sound more open and airy.
If the device has a drop in the range of 3-6 kHz, then it will not suit the fans of hard rock, because guitars and rough male vocals will sound jammed and illegible.
If there is a blockage on the graph in the region of 400-800 Hz – for most genres of music this is okay, but for some jazz with female vocals it can be critical.
The short answer is for what. For example, to mix music, a sound engineer needs a perfectly flat frequency response of speakers and headphones. Mixing music on speakers with a frequency response curve is not only difficult, but also pointless – you can achieve good sounding specifically on these speakers, but on any other acoustics, the mix will crumble and will sound completely different from intended.
However, for most listeners, speakers with a flat frequency response will sound insipid, boring and impersonal. Usually you want it to rock harder, so that the instruments seem to jump out of the speakers, so that the top is soft and silky, the middle is fat and voluminous, the bass is deep and elastic. The speaker manufacturers also understand this, so they deliberately tint the sound of their audio equipment (as in the graph above, humps are noticeable at low, medium and high frequencies).
In stores: in 2 stores
The same is true for professional audio equipment. For example, if you connect a guitar amplifier to a regular speaker, the sound will be like a bee hive next to a sawmill. However, the speaker in the guitar cabinet has a very high frequency response and acts as a filter, removing all the unpleasant sawing frequencies and leaving only a pleasant growl to the ear. It’s the same with microphones – many iconic models have a frequency response curve. For example, the Shure SM58 has humps in the 2-8 Hz region to emphasize vocal frequencies.
What affects the frequency response besides the parameters of the speaker system itself
A lot of things affect the frequency response. Headphones illustrate this especially clearly. The frequency response of headphones is never a straight line at all. For example, the frequency response of the default studio headphones for mixing music Sennheiser HD600, in theory, should be flat, but the measurements show significant dips and humps.
Why is that? First, music right near the ear is perceived differently than music in the distance. The same in-ear headphones often have a drop in the high-frequency region, because there are no obstacles in the sound path that provide a natural reduction in high frequencies. As a result, they have to be reduced artificially.
Secondly, our head, auricle and the auditory canal have their own frequency response and resonances, which are exacerbated in the closed volume formed by the ear and headphones.
Research in this area has led to the creation of the Harman Curve, the headphone frequency response graph that listeners find most enjoyable. This graph has been updated several times and is different for overhead (OE) and in-ear (IE) devices.
Not all manufacturers take the Harman curve into account when creating headphones, but the frequency response of many top-end models strongly resembles it. Perhaps the most famous among them will be the AKG K361 and K371.
Bluetooth headset AKG K371-BT black
You also need to tinker with the speakers to hear an even response across all frequencies. Yes, studio monitors are quite flat across the spectrum, and the influence of the auricle is not so significant, but all raspberries spoil the room.
Frequency response graph of the same column: green – outdoors (without the influence of the room) and red – indoors (more pronounced irregularity is noticeable)
For example, a small room will hum at low frequencies due to indoor modes – standing bass waves that especially like to accumulate in corners. The dimensions of the room, the materials of the walls and finishes, furniture, the listening point and the location of the speakers themselves all contribute to the formation of the frequency response.
How to achieve greater frequency response uniformity
Even most recording studios built from the ground up have uneven frequency response, which are corrected with acoustic panels, bass traps, diffusers and absorbers of various designs. Some of the easiest ways to improve your room sound can be found in this article , and here are more professional techniques. But one should not expect much – even cubic meters of mineral wool can only slightly smooth out humps and dips, but they will not radically change the situation.
If all else fails, then you have to correct the matter programmatically – with the help of equalizers. In professional studios, these are separate devices with a digital interface that are placed in circuits in front of the monitors. For home studios, there are VST plugins with the same function.
For household purposes, you can use the software equalizers of smartphones, players and PCs. The main thing to remember about EQ is that it is better to decrease than to increase.
An equalizer should be treated like a set of bad amps – adding volume at a certain frequency will add distortion.
This is true even for very accurate and transparent studio equalizers, let alone the default ones in a smartphone.
For example, if there is not enough bass in the headphones, this can be interpreted as an excess of mid and high frequencies. It is better to turn them down a little, and then add the overall volume, than to add bass directly – this way you get a cleaner and more pleasant sound to the ear. And only if it is not possible to achieve the desired by decreasing, then it is worth carefully adding the missing frequencies. There are exceptions to this rule – for example, when it is necessary to compensate for a specific and pronounced dip in a narrow frequency range. But even in this case, it is worthwhile to act carefully and add 1–2 dB.
The frequency response graph allows you to understand the overall timbre balance of the selected device and can become a starting point for correcting it with an equalizer. However, it would be incorrect to compare the frequency response graph from one manufacturer with the graph of another manufacturer, because everyone measures the frequency response in their own way and never tells how exactly. Measurements from audio enthusiasts and resources can be of great benefit – they generally don’t hesitate to talk about the measurement method. Also on YouTube, comparisons of recordings of the sound of speakers and headphones are popular , with the help of which you can get a general idea of the timbre balance and the nature of the sound.
However, the best measuring instrument for the average consumer will be his own ears. As mentioned above, they, like artistic tastes, are different for everyone. Before buying expensive speakers or headphones, you should listen to them in the showroom, or find the owner of the same device on the Internet.