Last week I was busy grading a batch of reports from my second year acoustics module where they were tasked with redesigning a run-down music venue in terms of internal acoustics and noise pollution. In addition, they had to specify a suitable sound system for the space. There was a common approach used for the subwoofer system in many of the reports; one which, if implemented, would be completely counterproductive to what the students were trying to achieve.
Many students came up with the idea that to save space, they’ll place their subwoofers underneath the stage. To ensure musicians on stage have a reasonable working environment, they specified that the subwoofers will have a cardioid polar response (something which features in a number of current units on the market).
At first glance, this seems absolutely fine and logical. There’s a big problem, however…
When you place a cardioid subwoofer underneath a stage, it will lose the majority of its directivity, more or less becoming an omnidirectional source. This is due to the interaction between the source’s output and reflections from the stage (which will wreak havoc on the intended interaction between the front and rear drivers).
In addition to this, a stage could resonate due to the close proximity of the subwoofers, causing further issues (in my experience, temporary staging is very good at resonating around 120 – 160 Hz, likely a higher harmonic of the driving fundamental frequency originating from the subwoofers).
Interestingly, this is the same misunderstanding I encounter when talking to many practitioners. It is blindly (or deafly?) assumed that cardioid subwoofers will retain their polar response no matter what.
Thinking back to the beginning of my time working with Gand Concert Sound, I remember a couple Nexo engineers visiting us. During our conversations they very clearly explained that CD18s need roughly 6′ (approx 2 m) of open air around them to maintain their intended polar response. There were many sketches on the back of napkins passed around and it made perfect sense. I suspect this lesson hasn’t made it to many in the industry.
Out of interest, a couple years ago I had one of my final year undergraduate students focus his dissertation research on this in order to carry out experiments to better illustrate this issue. We started by measuring a d&b Y Subwoofer in a large room (it was a roughly 10 m x 10 m) – don’t judge. It’s the north of England… we usually only get a couple weeks of good weather and that’s in the summer!
The measurements weren’t time-windowed since this would have caused a loss of detailed information in the low-frequency band. As the subwoofer would remain in exactly the same position throughout the experiment, we didn’t consider this an issue.
d&b Y Subwoofer measured in-room polar response (left) and magnitude response (right)
In our test space, the measured front-to-back rejection of the Y Subwoofer at 40 Hz, 80 Hz and 120 Hz was 16 dB, 14 dB and 16 dB, respectively. From documentation provided by d&b, the predicted front-to-back rejection is 17 dB.
So far so good!
Next, we placed a single piece of stage deck (8′ x 4′) over the Y Subwoofer and re-took the measurements.
d&b Y Subwoofer measured in-room polar response (left) and magnitude response (right) with a single piece of 8′ x 4′ stage deck placed over the subwoofer
In these measurements, the front-to-back rejection of the Y Subwoofer has significantly reduced. The most significant issue is seen at 55 Hz, where rejection is approximately 15 dB worse than before the deck was put in place. As a side note, the on-axis (in front of the stage) response is roughly the same as without the stage, so the audience experience isn’t changed by stage effects, it’s just the stage personnel that suffer.
To follow manufacturer recommendations, we then moved the stage so that the subwoofer was slightly in front of the stage (no part of it was underneath) and repeated the measurements.
d&b Y Subwoofer measured in-room polar response (left) and magnitude response (right) with a single piece of 8′ x 4′ stage deck placed just behind the subwoofer
Now we were back to where we started. The Y Sub is behaving as expected. And there was much rejoicing.
We carried on with the experiments to include a larger stage (a 3 x 3 grid of 8′ x 4′ pieces of deck) as well as testing for subwoofer placement on the front edge of the stage. To summarize what we found, here’s a table indicating front-to-back rejection for each test configuration with reference to the rejection achieved with no stage present. A value of 0 dB indicates that the rejection perfectly matches that without a stage present. Negative values indicate a reduction in rejection (* indicates simulated results using my software, LowFAT, which was required due to experimental limitations).
The reason behind testing two frequency bands was to inspect the subwoofer range exclusively (39 – 110 Hz for the Y Sub) and then to look at a bigger picture (20 – 300 Hz), which picks up stage resonances and other negative effects.
For a little while, this seemed to be the only focused experiments published in this area. Luckily, though, my colleague discovered some work by an Italian firm which replicated our experiments in late-2017. Their report can be found here. In this case, the tests were conducted outdoors with a simplified stage (plywood was used). Results seem to indicate that the degradation of the cardioid response is even more pronounced than our experiments had shown!
This isn’t quite the end of the story, though. There’s something else to keep in mind. Most large festival stages will have stage skirts to improve the “look” of the event. It’s important to check whether these are acoustically transparent or not. If they aren’t, you’ll find that your subwoofer output will decrease drastically. This is due to the reflection coming off the stage skirt. The reflection, when it reaches the front of the subwoofer, will be nearly the opposite of the direct sound. This will cause partial cancellation of the forward-moving wave, thus reducing efficiency.
The same goes if you’re indoors and place a cardioid subwoofer back up against a wall. I ran into this situation when on a gig with Adam Rosenthal a number of years ago at Chicago’s Auditorium Theater. The only place we could place our CD18s was on the offstage wings in front of the proscenium.
When we fired up the system the low-frequency seemed to be extremely low-level, even though we were driving the system fairly hard. After thinking about it for a few minutes, we realized that the reflection off the proscenium was cancelling out our direct sound. The solution? Turn off the rear drive unit and make the CD18s omnidirectional. After that, the low-frequency content came back to the expected levels and we had a good show.
So… what are the takeaway messages here?
- Don’t place directional subwoofers under a stage.
- Don’t place directional subwoofers in front of a wall.
- Beware of non-acoustically transparent stage skirts.
- Do place directional subwoofers slightly in front of a stage (1 m clearance is recommended).
As a final note, in case you think all of this stuff is new and I’m brilliant for highlighting these issues, I should note that Harry Olson and Richard Waterhouse figured out just about all of this before most of us (and even some of our parents) were born. Read their papers and books! There’s nothing new here…
Academic papers this post was based on:
- Hill, A.J.; J. Paul. The effect of performance stages on subwoofer polar and frequency responses. Proc. Institute of Acoustics – Conference on Reproduced Sound, Southampton, UK. November, 2016.
- Hill, A.J.; M.O.J. Hawksford, A.P. Rosenthal; G. Gand. Subwoofer positioning, orientation and calibration for large-scale sound reinforcement. 128th AES Convention, London, UK. May, 2010.
Selected “historical” reading:
- Olson, H.F. Elements of Acoustical Engineering, 2nd Edition. D. Van Nostrand Company, New York. 1957.
- Waterhouse, R.V. Output of a sound source in a reverberation chamber and other reflecting environments. Journal of the Acoustical Society of America, vol. 30, no. 1, pp. 4-13. January, 1958.