Great Equalization

     as the Use of an Equalizer

Blog posts here have mentioned that some speaker problems can be solved with an equalizer. Let’s take a closer look at what kind of equalizer is needed and how to use it. Firstly, a "parametric equalizer" is needed. This means a simple thing: for each adjustment, you can choose the frequency, the amount of change, and the width of the frequency band to be adjusted. Most of these also show the result graphically, so you don't have to worry about finding the right numbers right away. The need for this is evident when, for example, you need to change the signal strength at 320 Hz. In such a case, a slider equalizer with only 250 Hz and 500 Hz options is of little use.

Software

The most affordable option is to use an equalizer on a computer or mobile phone, where free high-quality software is available. The power of these devices is more than sufficient for calculations and remains completely unnoticed even when using the device for other purposes.

There are many applications for mobile phones, but the best ones are often attached to music players. For example, for playing music files stored on an Android phone or a home hard drive, a good choice is the Onkyo HF Player  (https://play.google.com/store/apps/details?id=com.onkyo.jp.musicplayer) with a proper equalizer.

On Windows computers, things are a bit more complicated, requiring two applications. The first one applies filters to the selected outputs – Equalizer APO (https://sourceforge.net/projects/equalizerapo/). This is also the core of many other equalizing applications, so you should ensure that no other such application is currently in use before installing it. To use this application, you also need a user interface, one of the most convenient being AQUA (https://github.com/h39s/AQUA). Once the filter is ready and saved, Equalizer APO will apply it by default at the every startup of the computer, without needing to load or start anything manually.

It's important that applications allow you to save filters. This is especially crucial when music can be played through multiple speaker systems, most often with a mobile phone. A computer is often more associated with a single set of equipment, and it's convenient to use saved settings for precise audio tuning. Not every change improves things, and it's good to be able to revert to a previous version.

This is the easy part, as the complex work has already been done by programmers. Now begins the complex part for us—what and how to adjust. If you have musical hearing, education, and golden ears, life is a bit simpler. You listen and make adjustments. Unfortunately, most situations are like in a top restaurant: yes, I can tell if the food is delicious. That doesn’t mean I could make it myself. Hence, measurement is helpful.

Measurement

Most people don't have access to a calibrated microphone. However, this might not be entirely true. For instance, if you have recently purchased a home theater receiver or a more advanced soundbar, it might come with a microphone suitable for basic measurements. Using REW software (https://www.roomeqwizard.com/) and this microphone, you can already measure several aspects. The learning process is, of course, lengthy, and initially, it might seem that the measurements are nonsensical, but there are many online guides available. With a bit of reading, a few videos, and some measurements, you can get useful information soon.

It's even easier with a mobile phone, as there are many suitable apps available. The simplest one is Spectroid (https://play.google.com/store/apps/details?id=org.intoorbit.spectrum). For measurements, you can use pink noise or a sweep signal (both of which can be recorded as audio files using REW).

For both solutions, it's important to note that a microphone's sensitivity varies across different frequency ranges. Unfortunately, calibration files for AV receiver or mobile phone microphones are not readily available online. Typically, the response has increased sensitivity around the 8 kHz region and reduced sensitivity at very low and/or high frequencies. Therefore, these uncalibrated measurements should be combined with listening, and it might not be a good idea to aggressively boost or cut large frequency ranges based on measurements only. However, they can quite accurately indicate where the bass is booming, the highs are hissing, the clear guitar sound is missing, or an instrument seems dull. Quick measurements and experimenting with the equalizer will soon clarify which frequency and how broad a range needs adjustment. A thorough guide on measurement, especially in-room measurement, would likely require about ten pages, which is beyond the scope of this post. Therefore, the strong recommendation is to read a bit and watch a couple of videos.

Objective

Although the goal seems simple (a good sound image), it's also important to set the right objectives. Some attention should be put to three aspects when setting your goals.

Bass Proportion. There is a significant difference between experienced listeners (musicians, audio engineers, DJs) and inexperienced listeners in terms of bass preference. Inexperienced listeners tend to prefer over-amplified bass (on average, though each listener is different). Experienced listeners prefer a flat bass profile that matches the midrange volume. The main issue with strong bass is masking. To understand this phenomenon, try playing some talk show with over amplified bass on a really good sound system. Many male voices will become unintelligible because the boosted bass diminishes the ears' ability to hear other frequencies. Applied to music, strong bass similarly reduces the ability to listen to the music's details and quality. Often, bass is boosted because smaller sound systems may struggle to reproduce lower bass frequencies. Unfortunately, they do not compensate for each other. Lower bass frequencies are still audible at lower volumes from good systems, and boosting bass in the frequency range accessible to a small speaker does not replace missing range.

Conservation of Energy. The second aspect is the most important law of physics for speaker design and setup. One might expect this to be an exciting formula about sound propagation. However, it simply states that energy cannot be created or destroyed, only transformed from one form to another. When setting up a speaker's sound profile, we don't need to examine the efficiency of the drivers or the internal design of the speaker. We just need to remember that if an audio signal is sent to the speaker, it will be reproduced by any normal speaker. If there are dips or peaks in the sound profile, it's not because the signal itself is lacking or excessive but because the signal is being applied in an undesirable way. The most common issues are directivity problems (sound radiates unevenly in different directions, though uniformly as a whole) and enclosure or room problems (frequencies do not attenuate equally) and possibly crossover problems. It is always worth considering what could be the added energy final outcome, when you want to boost certain frequencies within the speaker's range. Additional amplification at the lower and sometimes the upper end is more natural, but here caution is needed not to direct added energy to distortion.

Ratio of Room Sound to Direct Sound. To understand the role of these ratios, we can refer to the speaker measurement standard ANSI/CTA-2034-A, which states that the sound at the listening position consists of 12% direct sound, 44% early reflections, and 44% total speaker sound energy. Therefore, if we only measure the direct sound and set it to a flat response, the speaker's sound image may become significantly distorted. It is better to optimize for early reflections, i.e., at angles 30-60 degrees off-axis, both horizontal and vertical. Alternatively, you can optimize only for the listening position sound, as most automatic correction systems do. However, such optimally tuned setup often sounds unnatural. A compromise is to focus on the speaker and its overall balance of sound in different directions above 300 Hz, and on the listening position to reduce bass resonances below 300 Hz. A third, more convenient alternative is, if Spinorama measurements are available: you can use the graphs to make corrections without any measurements. This has been tested with TiFi Argos and Kraken and works excellently.

This can also be taken as an example, as it includes both the initial measurements and the filter with the graph. So far, the notation of bandwidth in a parametric equalizer has not been addressed. The usual notation is Q (Quality) with a range of 0.5–20. The smaller the number, the wider the bandwidth.

Example

Kraken needs assistance with lower frequencies typical for small subwoofers.

• Filter: Frequency 35 Hz, change +6 dB, width (Q) 3.
• Filter: Frequency 115 Hz, change -4 dB, width (Q) 1.2.

With these two filters, the bass is nicely adjusted within Kraken's capabilities. The only concern is the listening room, which has a strong resonance at 121 Hz.

• Filter: Frequency 121 Hz, change -7 dB, width (Q) 10.

With this, the Kraken is in order.

Next, the Argos speakers.

As explained in the speaker description, the lower frequency range needs additional boosting.

• Filter: Frequency 400 Hz, change +6 dB, width (Q) 2.

There is a small camel hump at the beginning of the high frequencies, which is easy to remove.

• Filter: Frequency 1747 Hz, change -2 dB, width (Q) 1.36.

A similar hump is also at the upper end of the high frequencies, which is also easy to remove.

• Filter: Frequency 9090 Hz, change -1.6 dB, width (Q) 2.9.

For peace of mind, you can add boost in the last octave's second half (an almost non-existent frequency range in real music).

• Filter: Frequency 20,000 Hz, change +8 dB, width (Q) 2.

The Argos graphs show a slight volume fluctuation around 5–6 kHz, where the ribbon tweeter takes more of the sound presentation from the main driver. A close look at the graphs reveals that there is also fluctuation in the directivity profiles at the same point. Therefore, it is not reasonable to adjust anything here unless something is directly bothering you. For instance, a volume increase around 6.5 kHz could cause sibilance, but this is not an issue here and no intervention is needed.

The resulting filters are shown graphically in the picture below.

It is not directly visible from the picture, but the entire signal was first given a -6 dB correction, and only then were the +/- adjustments made. This ensures that the equalizer application is possible in significant frequency ranges, even when the volume is maximized.

Moderation is a Virtue. As with such corrections, the main virtue is moderation. Correct half of what you think is necessary, and the result will be better. As seen in the Argos example, corrections are minimal in the sensitive range (2–5 kHz). Large corrections are needed at the beginning and end of the speaker's frequency range. One must also pay attention to the speaker's capabilities. For example, adding 25 Hz to the Kraken would be useless as small drivers wouldn't reproduce such low bass anyway. Also, if you have a computer speaker branded by some computer manufacturer, you should first check the frequency range they are capable of reproducing. This is another reason to be moderate with corrections because if a speaker cannot actually reproduce the desired frequency range, forcing it to do so can create strong harmonics and other distortions. Added harmonics can even be stronger than the fundamental sound, which could make the sound worse than before your correction. Corrections can be made to speakers that have the necessary capability in the frequency range. Speakers are the ones to reproduce the sound, and the equalizer does not change that in any way.