Revolutionizing Bass Shaker Equalization

If you are into high-end home theater, check out our Display and Audio Calibration Guides to maximize your experience.

Intro

When setting up my Dayton Audio BST-300EX, I noticed that typical EQ methods for bass shakers either rely on an all subjective approach or involve complicated procedures that sometimes lead to having to build your own measurement device. While not inherently wrong, these approaches are hard to learn, hard to complete, or inaccurate.

The largest number of these approaches involve sitting in the listening position, playing different frequencies, and adjusting them until they subjectively “feel right” or seem the same. However, relying solely on subjective feelings can be imprecise and difficult, and it is challenging to ensure accurate equalization without objective data. Plus, how can you be sure you are creating the right filters without unintended consequences?

Not being comfortable with any of the existing methods, I started brainstorming for an alternative solution.

Theory

Since we are creating vibrations in furniture and not creating sound, using a microphone will not work as a tool to measure the effect. This is because microphones are designed to detect vibrations in the air (sound), not in solid objects. Therefore, we require the equivalent of a microphone, but designed for solid objects instead.

Luckily such a device exists, it is called an accelerometer. An accelerometer is a sensor that measures the acceleration of an object to which it is attached. Acceleration refers to any change in velocity, whether that involves speeding up or slowing down. The accelerometer detects these changes in motion and provides corresponding electrical signals that can be interpreted to determine the acceleration. When an object is vibrating, it is changing direction at the frequency at which that vibration is occurring, and therefore, changing the acceleration. That being the case, we can use an accelerometer to measure the vibration in the furniture from the bass shaker.

With some clever math, which is way beyond the scope of this article but you can learn more about it in this video, we can change the acceleration data into frequency and magnitude, which we can then use to set up our EQ filters.

While the bass shakers may cause the furniture to produce audible sound above 20Hz, once the volume is properly adjusted, the sound level should be sufficiently low to avoid interference with your subwoofer equalization. Therefore, we will disregard any audible sound produced by the bass shakers during this process.

Image depicting how vibration can be converted to magnitude and frequency, albeit for a different application than what we are doing.

That is the idea: Use an accelerometer to measure the strength of vibrations, then convert that data into a response curve. This curve becomes the basis for setting up EQ filters like any other speaker. Moreover, accomplish all of this without having to build or use complicated equipment. The crucial aspect here is simplicity and accessibility, unlike other approaches that entail a steep learning curve and demand custom hardware and software.

Required Materials

An Accelerometer

Thanks to modern technology, most of us are walking around with accelerometers in our pockets, our smart phones. Smart phones have an MEM accelerometer in them, and there are many apps that allow us to see the data from them.

Not just any accelerometer app will work though. We are in search of one equipped with a graph called Spectrum displaying “frequency” on the x-axis and “magnitude” on the y-axis, similar to the illustration to the right.

For Android phones, the app that I use is called Accelerometer Meter.

For IOS phones, feel free to share your ideas in the comments below. One example is the Physics Toolbox Accelerometer, but I don’t have an iOS device to test this.

Google Sheets

For recording and analyzing the data, I have developed a Google Sheets file. This file helps with the conversion of recorded accelerometer data into a practical frequency response, which can then be exported to REW (Room EQ Wizard) or any other audio software.

You can find the file and save it to your own personal Google drive: Here.

A DSP

For this procedure, I will be utilizing an external DSP, specifically a MiniDSP 2x4HD. To successfully execute this process, it is essential to have a similar device, or alternatively, another means to adjust PEQ filters, volume, and delay for the bass shakers.

Please be aware that if you are using a different device, your method may vary slightly from mine, but the underlying concept remains the same.

An Audio Measurement Software

Once we have transformed the vibrational data into a frequency response, we will require a program to generate EQ filters. The one that I will be using is REW (Room EQ Wizard) which can be found here, but in theory an audio software that can adjust EQ should work.

Setup

The setup for this process is straightforward, but there are some preliminary steps that need to be completed before we can begin collecting response data.

Phone (Accelerometer)

It is best to lay your phone flat on the seat, without sitting in the seat, that you want to measure as adding any material in-between the furniture and the phone will affect the results. See picture below for reference:

Note: Some seats will not have a flat seating surface. Although not optimal, this should not have too much affect on the results as we will just use the highest peak axis when taking our measurements. Do not add any object under the device as it will affect the measurements.

App Settings:

As mentioned earlier, it is crucial to use the Spectrum graph, displaying frequency on the x-axis and magnitude on the y-axis. Additionally, ensure that the app’s settings are configured to measure using the accelerometer not linear acceleration. Verify that the accelerometer option is turned on, typically accessible in the app’s settings.

Ensure the graph is set to “Start” and “Auto-Scale.”

To determine the optimal axis for measurements, play a pure test tone, preferably at 40Hz, exclusively through the shakers. Identify the line with the highest peak and deactivate the other lines. For example, in the image above, the y measurement (green line) has the highest peak, therefore, only that line should be visible.

Audio Setup

To set up your audio measurement software for this process, follow the same steps as you would when measuring the response of speakers, excluding the use of the microphone.

Be sure that the output from your audio setup is free of distortion as this will greatly affect the consistency of the results. For example, when using ASIO audio drivers, sometimes it causes speakers to crackle, this is bad and we do not want this. To alleviate this problem, be sure you have the latest Beta version of REW and use the JAVA drivers instead.

For those with a dedicated bass shaker output ensure that the output is set to its default before proceeding.

If your processor lacks a dedicated bass shaker output, the setup becomes a bit more complex. It is beneficial but optional to cancel out any effects that existing EQ may already have. Only if the filters are known, you can input the opposite values into the new EQ, ensuring you leave at least 5 filters open for new EQ adjustments.

In my setup, lacking a dedicated bass shaker EQ, I use a MiniDSP 2x4HD, which I also use for the rest of my subs as well. Since I only use a couple of shared filters, I decided not to cancel them out. However, if I was to redo my EQ for my subs, I will need to redo my bass shaker EQ as well. Please keep this in mind as you decide what filters will be on during your measurements.

Crossover and Cutoff

Next, I would advise setting a crossover (low pass filter) and cutoff point (high pass filter) for your bass shakers before measuring. This is to ensure that the bass shakers will not be damaged from testing. However, it is worth considering that after applying EQ, there is a possibility that the crossover and cutoff may require further adjustments. Keep this in mind as you go through the process.

To determine the cutoff point, begin by examining the minimum frequency of the shaker specified by the manufacturer, which in this case is 12Hz. This is a good place to start, and you can modify from there. Regarding the slope, I set mine at 48dB. This choice is intentional, aiming for a flat response until the cutoff point, and then a steep drop afterward. For my BST-300EX, I have configured them with a cutoff at 10Hz and a slope of 48dB to maintain a flat curve.

Image showing where to set the crossover and cutoff point in the MiniDSP software.

Choosing the crossover point is subjective and can vary for each system. A practical method is to sit in the seat, play a bass sweep, and identify the highest frequency before it starts getting too distracting. Set the crossover point approximately 5-10Hz below that frequency. In my system, 70Hz was the highest I could go before it became distracting, so I set the crossover at 60Hz with a slope of 24dB.

Setting Bass Shaker Gain

Chances are, when you initially activate the bass shakers, they might be set to either a subtle or a violent intensity. This initial setting could potentially skew the accuracy of the measurements if it is too extreme in either direction. Therefore, it is crucial to find the right balance between these extremes.

I have found that achieving the best results in this process involves adjusting the bass shaker level to a few decibels above your preferred setting. To do this, play some of your favorite bass-heavy content, adjust the bass shaker gain until it meets your satisfaction, and then raise it by two or three decibels.

Data Collection

Data collection in this process is not overly complex, but it does require a considerable amount of manual data recording and time.

If you are using dynamic EQ or any other loudness compensation, it is recommended that you turn it off during measurements as it can affect the results. This will mean that it is possible that your bass shakers could receive the same loudness compensation as your other subs depending on your setup. Be mindful of this as you go through this process.

Step 1: Setup

Before we take measurements, we need to input our wanted measurement range and the volume we will be testing at into the spreadsheet as it will be used to help guide you through the process.

Enter this data on the Sheet labeled “Step 1: Setup.”

Step 2: Measuring Response

With the phone correctly positioned and recording, generate a pure tone at the desired frequency for measurement and input the value into the table on the sheet labeled “Step 2: Measuring Response.” If using REW, this is done using the generator function.

For optimal accuracy using the recommended app, it is advised to play the signal for at least 3 seconds to allow for stabilization. Afterward, click the “Stop” button on the app to cease recording. While minimizing movement of the phone, zoom in on the maximum value, and document that specific value.

Using the image to the right as an example, here we would take the value 1150 for the 40hz tone that was played.

Repeat the same process for all wanted frequencies. We suggest keeping the difference between measurements 5Hz and under. Going wider will not give you a clear picture of the response.

Step 3: Adjusting Response

Now that we have our starting point, we can work on adjusting our response. This is where we are going to determine a magnitude to equalize all frequencies to, and then figure out how much change in volume there needs to be.

First, we need to set the Target Value for our response. This is the magnitude on the accelerometer we will be aiming for with each frequency. On the sheet labeled “Step 3: Adjusting Response,” you have the option to set the target magnitude based on any value within the range of the initial response. I suggest going with a value that is near the average magnitude as a good starting point. If you find that value is too hard to hit for all frequencies, you can always adjust it. Just remember this might require to remeasure your changes in volume.

After setting the target in the spreadsheet, we will now adjust the volume for each frequency with the goal of hitting the target magnitude on the accelerometer. To complete this, play a pure tone with the same setup used to captured the response. However this time instead of just recording the magnitude, keep the tone playing and adjust the volume of the signal until it aligns with the target magnitude.

Using the image to the above right as an example, we can see that the magnitude for 40 Hz was brought down to 450. This was done by lowering the volume by 7dB.

Record the newly achieved magnitude value, along with the corresponding volume adjustment necessary to attain it. Repeat this procedure for each individual frequency in your list.

Please note that this does not have to be perfect; there is some wiggle room in how precise you need to be, so use your judgment. I found that anywhere within a magnitude of 50 was good enough. It may be hard for some frequencies, especially below 20hz to hit the target value. Use you best judgment, on how close you can get without damaging your equipment.

This is the longest and most tenuous part of the process, but the goal here is to bypass all of the other factors that affect how the system reacts to vibration by determining how much change in volume is needed to reach the target for each frequency. I have found that not all frequencies act the same when reducing/increasing volume.

Step 4: Exporting to REW

After making the necessary adjustments to achieve a flat response, the next step is to export the changes in volume as a response curve in dB. To do this, simply copy the text within the red box on the sheet labeled “Step 4: REW Export” and paste it into a .txt file.

Save the file to retain the exported response curve data.

Step 5: Creating Filters

Now that we have the exported response from our shakers, we can treat this like we would for any other speaker equalization.

We need to import the frequency response into REW, and then use the EQ function to create our filters. The way I found that works best for this process is to use the Auto EQ function and then use the manual filters to adjust the curve to your liking.

Image showing the EQ tool inside of REW with filters created to flatten the bass shaker response.

It is worth noting that we do not have to be overly concerned about the negative acoustic effects of boosting frequencies or narrow q bands because we are not producing sound. As long as we stay within the mechanical limits of the shaker, boosting and narrow q bands should not be a problem. However, due to the negative acoustic effects, most auto EQ programs may not allow for much boosting. This just means that we have to either create any boost filters ourselves in the manual EQ section of REW or drop the target level in the auto EQ target level box and then go back and raise the volume in the DSP that same amount. In the picture above, filters 1-5 were auto generated but since REW does not like to boost, I had to manually add filter 6.

The final step will be to input the filters into whichever DSP device that you are using.

Final Adjustments EQ

After inputting the EQ filters into the DSP, it is recommended to perform a final check. Optionally, but advisable, repeat the same process as in Step 2, but this time using the sheet labeled “Step 5: Verification” to obtain objective results. If you are satisfied with these results, proceed to the subjective testing phase, where you can make subtle adjustments to the EQ curve based on your personal preferences.

I discovered the need for a few minor adjustments to the EQ to better align with my personal feelings. Fortunately, the process was straightforward and uncomplicated because I already had filters in place. I simply made slight adjustments based on how the vibrations felt to my body. I mainly had to add more aggressive filters below 20hz resulting in the final curve (green) below:

Measured response after final EQ adjustments.

Adjusting Dynamic Range Compression and Delay

These topics are covered in more detailed in our article Dayton Audio BST-300EX, but I will summarize it here.

Adjusting bass shaker volume presents challenges because of uneven bass distribution in content, especially below 30Hz. Relying solely on demo videos is ineffective since the shakers may not activate at all for lower bass content. Striking a balance is tricky as some content barely activates the shakers, while others are overly intense. Boosting shaker gain for less active content is a straightforward solution, but it carries the risk of distortion with louder material.

To tackle this, DSPs like MiniDSP 2×4 HD offer a compressor tool. This tool reduces gain when the signal hits a set level, balancing shaker output and allowing engagement at lower input levels without overpowering at higher volumes.

My settings for the Compressor.

To set the delay, I followed Roland’s bass shaker article. First, I measured the distance from the bass shaker to the closest subwoofer and used that as a starting point. Then, I listened to a repeating bass sound until it matched the timing of the subs. In my system, a 1-millisecond delay adjustment was all it took to get them in sync.

A Work In Progress

The goal of this process is to improve the subjective approaches while not making it overly complicated where you have to be an expert to complete it. On top of that, I hope that this starts a larger conversation on the topic of how to EQ bass shakers as there is not much out there about the implementation process or its effectiveness. The method outlined above is not perfect; it is my initial attempt at addressing this problem.

I welcome your input on the process and any suggestions you may have. I hope that over time, this tool will evolve and become valuable for home theater enthusiasts seeking to install bass shakers, as I think they are an unappreciated upgrade to home theaters that many more should have. Your feedback is crucial in refining and improving this approach.

Please feel free to leave any suggestions in the comments below or email us:

Future Improvements

Areas at which I think that this can be improved on or need more experimentation are as follows:

1. Better data acquisition system
   • Better app
   • A more automated/more precise way of measuring
2. An Auto EQ program that is easy to use that does not care about boosting, as REW does not like to boost any frequencies.
3. Improved methods for handling measurements under 20Hz. From my observations, measuring under 20Hz was hard due to factors such as the shakers specs, the presence of Audyssey drop after 20Hz or variations in measurement devices. Improvements in this frequency range would be particularly valuable.
4. Work on optimal Compressor settings for the best attack and release time.
5. Experiment: Is a flat curve optimal for bass shakers, or would emphasizing specific frequencies or adopting a sloped curve be more effective?
6. Experiment: Does Loudness Compensation have a positive or negative affect on the bass shaker experience?
7. Experiment: How does the weight of a person affect the feeling of the shakers? Should we be EQ’ing with weight in the seat?

Thank you for reading. If you are into high-end home theatre, do not forget to check out our Display and Audio Calibration Guides to maximize your experience.