Ruede Hagelstein (Watergate Records, Upon.You Records) recently got in touch with me because he was unhappy with the acoustics of his Berlin based studio.
Especially the bass was causing him frustration, he told me when we met, as it varied so much through out the room and was inconsistent at his sweet spot. It made the process of getting the bass right in his tracks difficult, having to reference on a lot of different sound systems, and simply taking to many steps of refinement to get it to translate properly.
A first visit to the studio immediately uncovered the first major problem: His sweet spot was setup so that he was sitting right in the centre of the room, generally the place where all the problems in the bass region are at their worst. On top of that, the speakers were positioned asymmetrically in the room, and way too far apart form one another, which wouldn’t allow the stereo image between the speakers to properly develop.
A quick listening test confirmed my suspicions. Although there was plenty of low end energy in the room, it wasn’t present at the sweet spot. Meanwhile my reference track to check the stereo image and phantom center (4Hero’s “Give In”) showed it to be very inconsistent and stretched.
Measuring at the sweet spot supported these finding with solid data. The response varied drastically between the left and right speakers. Room modes had a strong effect on the low end of the response, while their ringing would make masking a real problem. The reverb time in the high frequencies was already quite low, but that was to be expected in a room of this size (roughly 22m²).
However, an initial calculation of the room’s modes based on its dimensions showed an acceptably even distribution with increasing frequency, so no dramatic problems were to be expected that would be difficult to solve.
As with all small rooms, the biggest difference in the frequency response can actually be gained by finding the best spot for the speakers and the location of the listening position.
No amount of acoustic treatment can remedy a bad choice of speaker and listener placement in a small room.
It quickly became clear that the best solution in this case would be to position the whole setup as close as possible to a wall. This is often the case for rooms under about 25m². The reason being that the effects of room modes tend to not all pile up. Cancellations due to speaker boundary interference are also higher up in the spectrum, and can easily be treated with absorption. The setup does lead to more energy in the low end due to half space loading, but this is easily remedied using the available EQ on the speakers, if necessary.
Ruede Hagelstein had the good idea of putting down laminate floor instead of the carpet, which I gratefully accepted. This would bring back a bit of reverberation in the high frequencies, and so would give me more flexibility to work with absorption.
The choice was quickly made to setup the speakers and listening position against the longest open wall. It allowed for the most ergonomic use of the available space, with the setup not getting in the way of the door, or being right in front of the window. Windows tend to be curious “walls” from an acoustic perspective, as they usually reflect back most the incident sound energy except for the lowest frequencies, which they absorb to some extent. At first glance this could be seen as an advantage, but in my experience putting a setup too close to a window always reeked havoc on the low end.
To decouple the speakers from their stands, I recommended Ruede invest in a pair of IsoAcoustics decouplers. Apart from stopping any vibration entering the stand and thus the floor, these decouplers also optimize the movement of the woofer which really does wonderful things to the time response, improving the definition of the bass and of the mid range.
With the initial placement roughly in the right spot, I started a series of measurements to confirm the frequencies of the room modes below 100Hz. These would determine which frequencies exactly the bass traps would be tuned to. The decision finally fell on 38Hz, 52Hz and 68Hz, as addressing these would cover the entire low end response below 100Hz. This is possible through the use of limp membrane traps which, by nature, have a fairly low Q and thus a reasonably wide bandwidth.
The positions for the bass traps were then found by identifying the areas of highest pressure along the walls at the corresponding frequencies.
As it turned out, the front left and back right corner were ideally suited.
Based on the position of the speakers and listener, it was an easy task finding the positions for the broad band panels to treat the first reflection points above, and to the sides of the setup. A row of three horizontal panels behind the speakers made sure any problems due to speaker-boundary interaction were suppressed.
Two broad band panels respectively were positioned in the wall-ceiling corner above the speakers and the mirror position on the rear wall to counteract the higher room modes around 100Hz, and take care of any corner reflections.
All broad band panels were built sufficiently deep to make sure they absorbed as low as possible. With a material thickness of 16cm and an extra air gap of 4cm, they are effective down to around 100Hz. Together with the bass traps this allowed me to evenly reduce the reverb time across the spectrum, and not leave the room sounding “dead” or “muddy”, even when the overall reverb time ended up being very short.
“Deadness” or “muddiness” tends to only become a problem when the reverb time in the high frequencies is severely reduced, but everything below is left uncontrolled.
A great side effect of proper broad band absorption across the spectrum is that the room feels very quite, an important attribute to hear the details of space, and really focus on the dynamics in the music.
Finally, a set of poly diffusors was placed on the rear wall at ear height to treat the mirror point reflections from behind the sweet spot. Diffusion needs some space to fully develop, which makes it difficult to be used properly in small rooms. But in this case enough distance from the listening position was available, and so choosing diffusion rather than absorption allowed the reflection to be treated without reducing the reverb time even further.
Once everything was installed I calibrated the monitoring volume using Bob Katz’s “K” system, where an average level of -20dBFS being played back results in 83dB SPL emanating from each speaker.
Calibrating a monitoring system this way has a range of advantages. First and foremost, working at a standardized playback volume allows our hearing to always have the same frequency curve. With some experience, this gives the engineer a very detailed feel on where to place elements in a mix, and what a good balance sounds like on that particular playback system. 83dB SPL in the meantime, has proven to be a sweet spot in playback volume, being loud, but still allowing many hours of listening without seriously tiring the ears. Working at an average level of -20dBFS in the digital domain on the other hand gives plenty of headroom such that clipping never becomes an issue, as well as being ideally suited to drive most plugins with an appropriate gain without causing unwanted distortion.
To verify the results I started a series of measurements to check whether all the goals were met both in the time domain and the frequency domain. And finally, a listening test with well known material confirmed that the measured data translated into real world results. Only an actual listening test can reveal if a system behaves as intended. While collecting data can point out problems, and makes finding a solution easier, certifying that a room actually sounds good can only be achieved by using your ears.