“Let our rigorous testing and reviews be your guidelines to A/V equipment – not marketing slogans”
Facebook Youtube Twitter instagram pinterest

The Powered Subwoofer Tests


 1.   Basic Frequency Response: After calibrating the test equipment the subwoofer is placed in the testing area, the microphone is placed at a 2 meter distance from the closest edge of the subwoofer and basic response measurements are taken over a range of 10-200Hz. An effort is made to adjust the subwoofer to achieve the most natural unprocessed response by defeating the crossover if available, or setting it to its maximum position and defeating any on board EQ if that option is available.  If the subwoofer is anything other than a simple front firing model an effort is made prior to this point to determine the best subwoofer orientation for the tests. This is the basic configuration that the subwoofer is left in for all tests afterward except for the testing of the subwoofer controls effect on its response and in cases where there are multiple subwoofer configurations or manufacturers suggested settings. In some cases a measurement with an extended upper range will be taken if the sub exhibits good response up to 200Hz. This test gives a look at the basic range the sub is suitable to cover and also some clues to the actual alignment of the subwoofer enclosure and driver.

basic frequency response.jpg

Example: Basic Frequency Response of a Subwoofer

The graph above is a fairly typical frequency response of an unprocessed sealed subwoofer showing a slow and gradual 12dB/octave roll off into the low frequencies. A typical bass reflex, horn, or band-pass alignment will often exhibit deeper natural extension before the response corner, but will have a sharper rate of roll off from there, typically 36dB/octave or even faster after a high pass, or rumble filter is applied to the natural roll off. You could say that a perfect flat line from one end of the graph would be the goal but that is probably unrealistic and in some cases not even ideal for most system/room combinations. The important thing to look for here is that the subwoofer exhibits a smooth overall response with no radical changes in the response throughout its intended bandwidth: Spikes, dips, notches, or other response artifacts, may indicate problems with resonances, or ringing. Slow, gradual changes in response that are small in magnitude are usually less detrimental and more easily compensated for.

2.   Effect of Subwoofer Settings on the Response: After the basic frequency response settings and orientation to be used for the tests are determined, the effects of the various subwoofer controls on the response will be measured. EQ boost will be engaged or disengaged, if there is a parametric EQ a measurement will be taken with boost at minimum, maximum and middle settings and also with low, middle and high Q at the lowest, highest and middle center frequencies allowed. A response measurement will be taken with the crossover set to its lowest, middle and highest frequencies. If there is an LFE input with crossover bypass that will be checked also. Any other available settings such as room compensation or damping settings will also be checked. This shows not only how the controls interact but how flexible the unit is and gives a clue as to how easy it will be to integrate well into a random room or system.

response settings comparisol.jpg

 Example: Effect of a Subwoofer’s Different Crossover Settings on its Response

Here in the above example we see how various settings of this particular subwoofers low pass crossover affects its response. The purpose of examining how a subwoofer’s controls interact is to get an idea of how much utility and flexibility they will have to help in integrating the subwoofer into a random system, or room.

3.   Long term power compression testing: Using a 50Hz sine wave the DUT’s output is adjusted until it produces 90dB at 2 meters. This is the basic starting level for high power testing. After the level is set an ascending sine wave sweep covering 10-120Hz  and of 24 second duration is sent through the DUT starting at the base 90dB at 50Hz level. The upper cut-off of 120Hz was chosen as this corresponds with the LFE roll off in modern movie soundtracks. The reason that I use this full range even though many subs have no appreciable output at the lowest frequencies below 20Hz is because today’s blockbuster movies often have ridiculously hot bass content that many times extends well below 20Hz and even 10Hz. If the subwoofer designer has done their homework they will have built in protection or some other means for their subwoofer to cope with these occasional monstrous ultra low frequency signals showing up at the inputs, if they have not then this will be exposed very quickly with a sweep of this nature. After setting the basic 90dB at 50Hz sweep level, the sweep is repeated in rapid succession, each time with the level increased by 5dB, until the subwoofer ceases to increase in output or is driven into limiting, over-excursion, distortion or other obvious duress. Many times an increase of a full 5dB on the last sweep is far too much to try so an increase of 2 or 3dB is tried instead. After the final highest power sweep the DUT is returned to the drive level that produced the initial 90dB at 50Hz base level and another sweep is run. The tabulated results of this test show whether the output from the sub is increasing in a linear fashion. In other words with a 5dB input increase the sub should increase its output by 5dB. In the real world thermal heating effects, port compression, driver excursion limiting, amplifier clipping or limiting and driver impedance rise due to thermal heating of the voice coil causes the subwoofers to compress the input signal when driven hard. Additionally these effects vary throughout the frequency range as the subwoofer impedance, excursion, vent or passive radiator loading and demands on the amplifier all change.  A better sub will exhibit less compression overall. The final repeat of the 90dB sweep after all of the other sweeps indicates how the heating of the driver motor and voice coil are affecting even the low volume basic response now. Less change is better here too.

power compression sweeps.jpg

 Example: Power Compression Sweep measurements of a Subwoofer

What the graph above is showing is how the subwoofers output level and response shape changes with a long term, high duty cycle signal applied at ever increasing level. The topmost trace in the graph is the subwoofers maximum long term output. Ideally you would see a linear increase of output that perfectly tracks the increase in the input signal level and a response shape that looks exactly like the low volume ones except louder, of course.


Example: Power Compression Chart

What is being shown in this graph above is derived from the power compression sweep measurements shown in the last graph, but in this case only the amount of signal compression by the subwoofer is shown. Across the bottom of the graph at 0 is the base SPL from the 90dB sweep level. Each 5dB increase in level for the next sweep measurement should show a perfect 5 dB increase, however in the real world this is not the case. This subwoofer was run up all of the way to an 115dB sweep level which is an increase of 25dB over the base 90dB level. If the subwoofer exhibited no compression at all you would see a nice flat line at the 25dB marking. Instead what we see is that the sub is very heavily compressing over the entire bandwidth by an average of about 5dB from roughly 18Hz to 70Hz and compressing by almost 7dB from 80Hz to 120Hz. Showing the data in this manner it is easier to see where the subwoofer is compressing the most and by how much. It is also easier to compare this data between different units with this type of chart.

4.   Maximum long term output: This is derived from the highest sweep level attained during the long term power compression testing. It is not a new test but simply a different way of presenting and focusing some of the data gathered for the highest level output sweep that the DUT will handle. This is the highest long term sustained level attainable by the DUT within the entire intended bandwidth, considered to be 10-120Hz, before any limiting condition is met in any part of the frequency range. Limiting conditions include amplifier power or excess any of the following: excursion, noise, compression, or distortion. Obviously not all subwoofers can cover that entire range effectively. However if competently designed they should be filtered and limited so as not to damage themselves if a signal containing high level energy below or above their designed range of operation is presented. Basically stated this is the maximum level attainable in a long term manner, at which point the subwoofer’s behavior in some part of the frequency range has become a limiting factor. The higher the maximum output attainable the better the dynamic tracking of the subwoofer and the more suitable it will be for high output duties, or large, open spaces.


 Example: Maximum Long Term Output Comparison Chart

In the example graph above the maximum level reached by the subwoofer during the power compression sweep tests is charted in green and compared against the all time highest levels recorded, shown in red and the all time lowest maximum levels attained, shown in the black trace. Note that the all time maximum and minimum traces are composites from multiple units. There may be 4 or 5 different subwoofer units responsible for either the minimum or maximum trace.

5.   Distortion measurements: THD and distortion by component measurements are taken using 33 second long ascending sine wave sweeps at the same drive levels as used for the long term power compression sweeps. Lower THD is better but also the harmonic make-up of the THD can be a factor. A high level of second harmonic distortion is generally thought to be more benign to the ear and in some cases even pleasing. A high level of a 3rd harmonic distortion is usually held to be much more audible and displeasing to the ear. Additionally it is generally thought that the higher the harmonic the more easily noticed. Note that environmental noise makes getting uncorrupted outdoor distortion measurements in the low bass range at the lower volume levels difficult. Also most subs have reasonably low distortion until driven to a pretty significant percentage of their capabilities. For these 2 reasons often only the 3 or so highest level distortion measurements will be presented. This is usually enough to establish the trend in the subwoofers distortion performance.


Example: Subwoofer THD Measurement

In the above THD chart example each color corresponds with a sweep level that was used during the power compression tests. As can be seen the amount of distortion will rise with increased level and exhibits a certain trend to it. This is typical behavior. A moderate spike in level near 55Hz is seen and the distortion quickly gets out of hand at low frequencies. This particular data is from a small ported subwoofer that is tuned relatively high, which limits it to being useful above 30Hz.

component distortion.JPG 

Example: Subwoofer Distortion by Harmonic Component

Above is the distortion data for the same ported subwoofer from the 110dB sweep. This is the same data that was shown in purple on the earlier THD graph, but broken up into the harmonic components. This subwoofer exhibits high 2nd order distortion and moderate 3rd harmonic. The 4th harmonic is quite low. Even though the overall distortion is high its harmonic make-up means that it may not be as objectionable as a subwoofer with a much higher 3rd or 4th component.

6.   CEA-2010 Maximum short term clean output testing: This test involves running short 6.5 cycle duration, shaped signal bursts centered at 1/3rd octave intervals of 20, 25, 31.5, 40, 50 and 63Hz through the DUT and monitoring it’s peak output level and harmonic distortion of the signal against a prescribed stair stepped threshold in order to determine the DUT’s maximum peak useable output at that center frequency. Each center frequency is sent to the DUT and its output level is then increased while the distortion is monitored until either: the DUT stops gaining in output level, or the prescribed stair step distortion threshold for any harmonic is exceeded and the data is recorded. Additionally I have added further test center frequencies at both ends of the spectrum (10, 12.5, 16, 80, 100 and 125Hz) to better represent more of the typical full bass bandwidth and also prevents manufacturers from focusing on a narrow band of frequencies at the exclusion of others to try and better their results in this test. As long as the systems are capable of reproducing meaningful output at a particular center frequency the results for that band will be reported.  What you should look for in this test is the highest output levels over a broad range as that will indicate a subwoofer with very high dynamic capabilities and a lot of headroom. This is also a very useful test for indicating a subwoofer's useful extension and can and can provide good indicators of how suitable a subwoofer might be for loud playback levels and larger spaces. These particular tests as conducted are at 2 meters and in some cases back calculated to 2 meters. 3dB is subtracted from the CEA2010 peak readings to approximate RMS output.  If comparing to 1 meter RMS approximated CEA-2010 data you would add 6dB to 2 meter results or vice versa. For every doubling of distance the apparent SPL level will drop by 6dB. For each halving of the distance it will increase by 6dB. If comparing RMS  to peak readings you would add 3dB to the RMS one. It is all the same data, you just need to know what is being presented and how it was collected in order to compare it fairly.


Example: CEA-2010 Maximum Sort Term Subwoofer Output

The above example chart shows the maximum clean output recorded from the subwoofer under test with the CEA-2010 program at every center frequency within the entire 10-125Hz bandwidth of testing, in one plot. The data for the subwoofer currently under test is shown in green. The all-time maximum output recorded in each band is shown in red, and the all-time minimum is shown in black. Note that the maximum and minimum traces are a composite from multiple subwoofers.

CEA2010 signal.JPG    

Example: CEA-2010 Subwoofer Measurement

Above is an example of the analysis done by the CEA-2010 program on a subwoofer's output when stimulated with the shaped burst signal. This particular case is a 20Hz test and indicates that the 3rd harmonic is the limiting factor for this subwoofer. At any time if the output exceeds the threshold indicated by the red line the unit is considered as failing.

7.   Time domain tests: These are derived from previous measurements and can be presented as waterfall and group delay data among others. These types of measurements give insight into how much delay, ringing, or what rate of decay the DUT may have. Sometimes a subwoofer may have significant resonance at a specific frequency range that would otherwise not be obvious from looking at a basic frequency response graph, but may be audible. These tests are a way to look for these types of issues in the time domain. Ideally when looking at a group delay graph you would see a very small value that smoothly increased with a very shallow profile towards the lowest frequencies. With a spectrogram or waterfall plot what you will see ideally is a very rapid drop off of the signal level indicating that the DUT has a rapid decay and does not ring or hold onto any frequency longer than it should. If there is a range where the signal continues on with a shallow drop off that indicates a problem with the DUT holding on to that frequency longer than it should.


Example: Subwoofer Waterfall

This graph is showing the signal level and decay rate of the subwoofer across its frequency range. The graph above has a time scale indicating milliseconds on the right, a frequency scale across the bottom and a scale indicating sound pressure level on the left.  In this case it is showing the decay of a vented subwoofer which in general shows a nice and quick decay for most of the frequency range, with the SPL dropping off by 25 to 30dB very quickly. At the bottom near 20Hz the subwoofer has a longer more gradual decay. This is where the vent tuning is and this longer rate of decay near the vent tuning is a common trait for vented subwoofers to exhibit. Ideally what you would see is a very rapid drop off of energy across the board in this type of chart.

 group delay.jpg

Example: Group Delay

The type of graph shown above is a way to examine the phase and response characteristics to calculate the delay of the subwoofer imparted to the input signal. Lower values are better here indicating less delay to the sound. There is some subjective opinion to how much delay is acceptable or audible in the bass range and this also varies with frequency. The lower and longer wavelength that the frequency is, the more delay is held to be acceptable generally. It is usually said that as long as group delay is below 1 cycle, or sometimes 1.5 cycles it should not be of concern or easily detectable. In the above chart the blue line indicates 1 cycle and the red line indicates 1.5 cycles. In the example the subwoofer is a roughly 25Hz tuned vented subwoofer. Note that the group delay increases steadily towards the lower frequencies and has a peak near to the subwoofer's vent tuning, that exceeds 1 cycle, but still stays under 1.5 cycles. This is not uncharacteristic for a vented subwoofer. They usually exhibit increased values near the vent tuning.

Final Test Notes: Things to Remember

1.   All SPL measurements are referenced to 2 meters unless specifically stated otherwise.

2.  All data is either unsmoothed or has minimal smoothing applied. If you are used to seeing a lot of smoothing on measurements, and you probably are, then unsmoothed or lightly smoothed data will look rather rough to you. Excessive smoothing hides blemishes or features and diminishes their magnitude. In other words it can make things appear much better than they are.

3.   The purpose of testing and measuring in this manner is to provide as level of a playing field as possible for each unit and to provide relative data for each subwoofer that can be compared to each other and even to other data sets in some limited ways.

4.  This article will serve as the demarcation point for all subwoofer tabulated test results.  Check back regularly for the most recent copy of our downloadable Excel spreadsheet for all subwoofers tested under this new test protocol.

Additional Information

Further information on many subjects related to the testing of subwoofers such as: CEA-2010, ground plane testing and subwoofer orientation is available by following the links provided below.

Audioholics 2010 Subwoofer Shootout Measurements Overview — Reviews and News from Audioholics

Subwoofer Measurement Tactics: A Brief, Topical Overview & Method Comparison — Reviews and News from Audioholics

Audioholics Subwoofer Measurement Standard Part I — Reviews and News from Audioholics

Audio Measurements: The Useful vs. the Bogus in Consumer Audio — Reviews and News from Audioholics     


Also check out our Subwoofer Room Size Rating Protocol


Confused about what AV Gear to buy or how to set it up? Join our Exclusive Audioholics E-Book Membership Program!

Recent Forum Posts:

gene posts on June 25, 2014 20:24
its phillip posts on July 21, 2011 17:46
Yeah, there's a compact, plus, and ex. I'm hoping he's testing the ex
shadyJ posts on July 21, 2011 17:03
Which Outlaw sub are you testing? There are three LFM-1 subs. I would also be interested in knowing if you intend to test them with their normal orientation, driver and ports facing downward, or will the driver and ports be facing the mic. Hopefully you guys can do both, I would love to see the difference in the measurements.
its phillip posts on July 21, 2011 13:50
That would probably be a smart investment
Ricci posts on July 21, 2011 13:02
Looks like a busy August for me. Its a good cross section of subwoofer types too. There is sealed, passive radiator, vented and also a known performer represented. I just hope the weather cools down a little. We have had a brutal heat wave in the Ohio valley. 95-100 everyday and high humidity with little wind at all. Heat index of 114 today. Might have to invest in a campers gazebo.
Post Reply