In this post I will share with you just a case study to demonstrate the
amount of effort it takes to get a subwoofer working optimally in a
listening environment. I'll share the measurements, steps and results
in my own listening room.
We will cover equalization only, which is essentially integrating the
subwoofer into the room but not integrating it with your other
speakers. That requires even more effort we may cover in later posts.
This blog post covers perhaps one-half of the work needed to get to finished. If it seems too hard, I strongly encourage you to look into automatic room correction (ARC) systems such as from Anthem or self-calibrating subs from JL Audio.
One thing I want to point out in this post is that we are using just one subwoofer, in a good but not great room and moderate room treatment. It is possible to have really great bass performance this way.The Room
The room is approximately 12' x 20' with the audio system and TV arranged along the narrow wall. Two GIK Acoustics soffit traps are used in the corners behind the single subwoofer which is placed in one corner. See the picture, below. Other GIK Acoustic panels are also used, but they provide minimal bass interactions so I won't go over them here. The corner's at the opposite end of the room are open to the rest of the house. One to the upstairs and one to the kitchen and dining area.
The Subwoofer
A Hsu Acoustics V-15 Mk II is used which has only basic controls.
The Equalizer
I'm using a miniDSP which I believe has been discontinued as mine has XLR in/outs. The closet today is the miniDSP 2x4 or 2x4 HD.
I'm also using the miniDSP 2 Way Crossover Advanced software with it but really any DSP based EQ platform should allow you to do this yourself.
The Results
The red line indicates the original response, the blue line the corrected response. This took 5 bands of EQ to get right. There are two major differences in the responses. First note that in the original the peak output is around 80 Hz. We have moved it to 20 Hz. Next, notice the valley around 38 Hz. We actually corrected most of it! That should be exciting because this goes against the common audiophile myth that we cannot fix nulls. You can if you use bass traps appropriately. The GIK Soffit traps are some of the few affordable products I know of which work down here well enough to do this with only two.
The other attribute I want you to notice is that from 20 to 100 Hz the output is NOT flat. In fact it descends about 2 dB per octave. Experience has shown this sounds far better than flat. JL Audio and other good ARC systems take advantage of the same type of curve, somewhere between 1.5 and 2.5dB / octave.
The Correction Curve
Lets look at the final effect of the DSP correction. How we got here is not obvious so we will cover the individual filters in the next section. Broadly though, there's an obvious correction for a null at 37 Hz, while we suppress the signals above 40 Hz. This should make sense since we are trying to shift the maximum output down.
It's important to remember that when using an EQ it is better to cut the high frequencies than boost the bass. This is because you can create overload situations where the electronics clip and cause audible distortion. Theoretically the curve could be shifted upwards so that 70 Hz is the reference or 0 dB level and everything to the left is boosted by a significant amount, but in that case the 10-30Hz would be boosted by at least 7 dB and the null at 37Hz would be up 19 dB instead of 11.
The Filters
The correction you see above is brought about thanks to judicious filter selection. Let's go over them one at a time
Peak Filters
Peak filters just means a filter with a center point. They are defined by their center Hz, the change in amplitude desired at that point and their width or Q. Higher Q = narrower effect. Lower Q = broad effect.
The biggest peak filter we used is to correct the null at 37 Hz.
Again, the only reason this works is because of the bass traps. No matter the money you spent on DSP you simply cannot EQ nulls without them.
The other peak filter is used at 25 Hz.
It is actually a combination of two filters I was using before, but after reviewing the results of them I realized they were fighting each other and that I could combine them, and this should be another lesson:
Make sure to review the individual filters to make sure they are doing something useful. If not, bypass or remove them.
The last peak filter we used was a subtle 1.7 dB suppression at 70 Hz.
Shelving Filters
Shelving filters reach a plateau above or below the frequency that they are set to. They behave like you'd expect a traditional bass or treble control to work, only we can control the frequency and slope precisely.
As we can see I'm using this filter to create the downward slope I was discussing at the very beginning.
Summary
- A single subwoofer with bass traps and appropriate EQ can sound phenomenal.
- Bass traps which are effective in the range of 40-80 Hz make EQ possible. GIK Acoustics Soffit Traps are excellent choices for this.
- Don't EQ for flat, use a descending line of 1.5 to 2.5 dB/octave as your target
- Be subtle and gentle in your filter choices.
- Avoid boosting the bass when you can cut the treble.
- Always review your filters individually, make sure they matter.
- EQ the individual components before you integrate them. That is, establish a good subwoofer curve first, then worry about mating it with your main speakers.
Not writing much about diy speakers these days.
ReplyDelete