We have written before about the benefits of using Digital Signal Processing (DSP) as part of your total room correction solution. In this article we look at a specific application and the benefits that DSP provided.
DSP and digital room correction add valuable features and adjustments to a system and can make dramatic changes that are impossible to correct any other way; but what’s important to note is DSP is only the last step in a 3 step process to achieve an optimal listening experience: Placement, Room Treatments, and finally DSP (or room correction).
For this project we were integrating a subwoofer into an existing 2 channel stereo system; a system we discussed briefly in Modernizing a classic Marantz 4300 Receiver where we added a MiniDSP 2×4 signal processor. This processor has 2 inputs and 4 outputs and was inserted between the pre-amp and amp stages of the receiver. 2 of those outputs drive the main speakers, and the other 2 outputs were used to drive the subwoofer. A standard twisted pair RCA cable was used to temporarily make the connection and after verifying there was no noise induced in the system, it was decided to leave them permanently connected.
The subwoofer we were integrating was an older Focal Chorus SW700V; an 11″ ported sub with a 300 watt amplifier. As mentioned previously, placement is step 1 in the multi-step process to achieve optimal results, but optimal placement is sometimes not an option. For this multipurpose space, cosmetics and space were of primary concern so it was decided to place the subwoofer out of the way in the corner. Consulting the product manual, it turned out that Focal recommended corner loading this particular sub.
To verify the subwoofer was working correctly, we quickly set up a crossover between the mains and sub at 45 Hz using MiniDSP, and then by playing music, slowly brought the subwoofer level up until it blended well with the towers using the knob on the back of the subwoofer. Finally, by selecting back and forth on the 0 to 180 degree phase switch on the sub, we did a couple of listening tests and decided to leave it set to 180 degrees. All of this was done by ear and achieved a fairly pleasing result. It was left for 48 hours and through that time it was noticed that some notes came in stronger than others. Clearly there were peaks in the subwoofer response that needed investigation.
Once the house was clear of any excess human noise, we brought in our acoustic measurement system and did some baseline acoustic readings. Below are three graphs that show the baseline readings. The green graph is the L&R speakers combined, with the subwoofer off. The Cyan graph is of the subwoofer ran full range, without a crossover. Finally the red is the system response with the subwoofer set by ear. Clearly the strong peak around 45 Hz is an issue and the transition from sub to main speaker was ragged.
Acoustic treatments could have been considered to help tame the strong 45 Hz peak, but bass traps tend to stop working below 80 Hz as well they tend to be very large for low frequency use. Also, as this is a multi-use space, the addition of acoustic treatments was decided to be minimized. Without DSP the only option left available would have been to turn the subwoofer level control down, which would have linearly affected the low frequency extension of the system. By using DSP parametric Equalization to knock down the peak we would be able to maintain the low frequency extension.
Turning again to MiniDSP added multiple filters to the subwoofer to smooth out the response across the response band we wanted. We also increased the crossover to 75 Hz at a shallower slope then before in order to observe the results. The below graphs illustrate the effect that parametric EQ had on the subwoofer. We applied 6 different filters to the subwoofer to achieve these results. The Cyan graph shows the subwoofer before and the magenta is after equalization.
Finally bringing the speakers back in and running a full system measurement, we achieved the following result. You can clearly see a large suck out at about 130 Hz. This is a room mode at the listening position and cannot be fixed with DSP. It is a much narrower suck out then shown with 1/6th octave smoothing and isn’t noticeable in practical listening. The large increase in sub-bass from 200 Hz to 40 Hz is purposeful to emphasize the low end and tracks to our House Curve very well.
Comparing the baseline to final results we can see the dramatic difference that we were able to achieve with use of parametric EQ.
After being pleased with the equalisation results we turned to fine tuning the time alignment between the subwoofer and towers. This is a feature that nearly no subs on the market have controls for and usually can only be brought in with DSP. By looking at advanced graphs available through our acoustic system we were able to determine that the arrival time of the subwoofer signal was 22ms delayed from the towers. This might be surprising because the subwoofer was physically right beside one of the main speakers but this is common and is due to the natural delay caused by the subwoofers active circuitry and amplifier. We dialed in the required time delay into the DSP (Specifically we delayed the towers by 22ms, to align to the subwoofer) and re-measured the system to verify, and all was in alignment.
Playback on the system was noticeable improved with no more ‘one note bass’ as was observed before. Also, because of the time alignment the subwoofer truly disappeared acoustically and we preserved a strong center image between the tower speakers.
As a closing note, while playing some music loudly with the system, we noticed that the cone excursion on the subwoofer was not at maximum; we could be getting ‘more’ from the woofer. We turned the subwoofer level up further and then backed it off a bit to a confident safe level. We did another measurement and achieved the following; and linearly 10db increase in bass from 50 Hz down. While this is too much bass for our needs from 40 to 100 Hz, we could add significantly more parametric EQ to smooth out the response, but still maintain the 10db increase from 40 to 15 Hz. This increase would extend the systems -6db point from about 33 Hz to 28 Hz. The red mark up below shows what we could expect to get with more equalization of the sub woofer.
**WARNING TO ABOVE** Do not take this on lightly. What we are effectively doing is causing the amp to push much more power to the low end of the sub-woofer while backing off on the higher end. This is going to cause distortion to increase in the boosted area in comparison to the rest of the range. Only do this if you’re confident in your abilities to detect distortion, non-linearity and clipping in the amplifier.
Closing Thoughts: If we did not use DSP in this situation we would have had two strong negatives with this project. 1) The strong resonant peak at 45 Hz would have overpowered the rest of the bass. More then likely, over time, the customer would have turned down the subwoofer level to control the peak thus nearly eliminating all sub-bass in the system. With parametric EQ we were able to remove the peak and leave the surrounding level unaffected. 2) The speakers and subwoofer would have been out of time alignment causing muddy bass and the subwoofer’s location to be pin pointed. This also would greatly hurt stereo imaging of any bass notes in the recording.
Focal Chorus SW700V
Focal Chorus 826v
Room EQ Wizard