Studio SOS: Narrow Room Solution

As you can see, the room was quite narrow — which is far from ideal from an acoustics point of view!

The SOS team travel to the big smoke to deal with some tricky acoustics in a media composer’s writing studio.

Trained as a musician and composer, and able to play keyboard, guitar and violin, Adrian Leung came to London from his home country of Australia to make a career out of writing music for the media. He’s already produced several high-profile jingles, and countless tracks for libraries, but the small studio space he has available has proved challenging when it comes to accurate monitoring. When we arrived at the North London house he shares with his girlfriend, the entire place had been gutted and workmen were everywhere — but the studio room was still intact. So, ignoring the sound of power drills and circular saws, and being careful to avoid the wet paint everywhere, Hugh and I got to work.

Dimensions

The room shape and size was a real challenge as it was only around two metres wide, three metres long and around 2.5 metres high, with bare painted walls on all sides and a window to the left of the mixing position. Adrian had his Adam A7X monitors set up on basic Auralex MoPads atop a Zaor Miza studio desk, which also accommodated his iMac running Logic X, a Universal Audio Apollo Duo interface, a Golden Age Pre-73 mic preamp, an Arturia MicroBrute synth and a Novation Impulse 61 master keyboard.

We knew that sorting out the overly lively character of the room could easily be accomplished using standard acoustic foam, but in such a small room the modal response inevitably plays havoc with the lower frequencies. There was just no space for any significant bass trapping in such a small room, so our plan was to find a position for the monitors that gave the least uneven response and then use some monitor EQ to tame the worst of the peaks.

As the Apple iMac has no CD drive we couldn’t use Hugh’s carefully compiled test CD, so we constructed a semitone-increment sequence of sine-wave tones using Logic’s EXS24 sampler, making sure that the MIDI velocity was set to ‘fixed’ so that all the notes would be played at the same level. (As the lack of CD drives is becoming more common, Hugh says he will put the test CD tracks onto a USB thumb drive for future Studio SOS sessions!)

Even with the MoPads in place, we could still feel some vibration from the speakers reaching the desk. The most likely reason for the apparent lack of damping in the MoPADs was that the Adam speakers were insufficiently heavy to compress the foam and get it working properly as a spring. A simple solution for this common problem is to add mass to the speaker in the form of heavy ceramic tiles or steel plates between the speaker and foam — enough to compress the foam by 15 to 20 percent of its unloaded height. However, in this situation it was easier for us to swap out the MoPads for a set of IsoAcoustic ISOL8R 155 stands, kindly donated by SCV London. These needed only the short posts fitted with the big spacers added to the rear posts. The latter tilted the monitors downwards slightly so that the tweeters were aimed towards the monitoring position. The reduction in vibration reaching the desk was dramatic.

Running the bass test signal again (while using the spectrum analyser in Logic’s EQ to identify the specific problematic frequencies) revealed significant peaks at around 50Hz, 100Hz, 150Hz and 200Hz. Although moving the speakers forward on the desk and moving one of them closer to the computer monitor to disturb the room symmetry did reduce the severity of the peaks slightly, they were still clearly going to be a problem.

Foam Home

As we knew in advance that there was no room for effective bass trapping, we brought along an IK Multimedia ARC2 room-correction system as part of our bag of tricks. However, there was little point in trying to do anything more about the low end until we had tackled the mid and high reflections that made the room sound overly bright. This was dealt with in the usual way by installing acoustic foam panels, kindly supplied by Universal Acoustics. In this case we had eight Saturn Pyramix 600 tiles, featuring a pyramid surface pattern in 600mm squares, with a 50mm overall depth. We glued the tiles directly to the walls, using three along the front wall behind the monitors, two to the right of the monitoring position around the mirror point, and two on the rear wall. All these were placed to be level with the ears of a seated listener. That left one spare panel that could be placed in front of the window when mixing. We used the spirit level in my phone to ensure they were all level.

As we’ve come to expect, the application of a few foam panels really dried up the acoustic of the room dramatically, and also improved the accuracy and stability of the stereo imaging. It was also obvious that relatively thin foam was going to have no significant effect below around 250Hz, and that’s where the worst modal room resonances were occurring.

Up goes the foam...Both Hugh and I view active room-correction systems as a last resort and something that should only be considered after other avenues have been exhausted. In this case, though, there was little we could do in terms of mechanical room acoustic treatment other than trying to find the monitor position that gave the ‘least bad’ results. Adrian already had the speakers firing down the long axis of the room, and his seat was forwards of the problematic dead-centre position that afflicts small rectangular and square rooms, so there was little more to be done other than use some form of corrective monitor EQ.

Corrective EQ

The IK Multimedia ARC2 system takes a minimum of seven room response measurements around the listening position (each derived from the average of 10 ‘chirps’ emitted by each speaker in turn), and then works out a correction curve to give the flattest in-room response. We ran the analysis several times, but although we’ve had good results from the ARC2 system on previous occasions, it didn’t seem to be able to adequately tame the excessive peaks and troughs in the response of this little room. It computed a correction curve, but when we replayed our staircase test some peaks we’re still way too prominent.

As the resonances seemed to be mainly at 50Hz and its harmonics, we decided to experiment using notches set up in one of Logic’s parametric equalisers in combination with adjustments to the low shelving controls on the back of the monitors. After some trial and error using Logic’s standard Channel EQ, we had some success in evening out the response but we all noticed that the stereo imaging seemed to suffer. I then tried the same thing using Logic’s Linear Phase EQ, which sounded much more transparent in this application, allowing us to bring down the worst of the peaks, though the overall level of bass then sounded a little light on recordings that we knew well (transferred into Logic from iTunes). Hugh suggested using some gentle low shelving boost overlaid on the notches to restore the overall bass level and that actually seemed to work very well. We also put in a little 6dB/octave high cut above 15kHz to tame an excess of highs shown up by the ARC scan, and that seemed to make the high end sound a lot more musical.

IK Multimedia’s ARC2 room-correction system identified some strong resonances at 50Hz and its harmonics.By jumping back and forth between test songs and our sine-wave staircase we made further small adjustments to the notch widths and depths until we felt the bass was being properly represented and that the bass response was as even as we could get it. After an hour or two of experimentation we ended up with the monitors set flat and the EQ as described, which we saved as a preset in Logic. We explained to Adrian the importance of remembering to bypass the monitor correction EQ plug-in in Logic (whether our own linear-phase EQ preset, or the ARC2 plug-in) when bouncing the final mix. This is because Logic doesn’t have a dedicated control-room monitor bus, so the corrective EQ has to reside in the main stereo mix output channel. Consequently, if it wasn’t bypassed when bouncing out a final mix, the room correction EQ would be imprinted on the mix... which wouldn’t be a clever thing to do! Some other DAWs have a dedicated monitoring bus and inserting a corrective EQ plug-in there obviously avoids any risk of ‘contaminating’ the actual mix. I also suggested to Adrian that the easiest way to use the corrective EQ was to add it to the master output of the default song template in Logic, so that it is always in place at the start of a new project.

Coffee Break

At this point Adrian treated us to lunch at a local café, which gave us chance to chat and to get away from the aroma of spray adhesive and the noise of the building renovations. Over lunch we explained the limitations of what can be achieved in very small rooms such as this one, and that using EQ in this way is always something of a compromise, so it is always worth double-checking mixes on decent headphones and, where possible, on other systems in different rooms. We also discussed the pitfalls of using plug-in presets when mixing, especially for dynamics processing like compression and gating where it is vitally important to adjust the threshold control to set up the required amount of gain reduction. A dynamics plug-in preset is unlikely to be exactly right in respect to the threshold setting as the preset designer can’t know at what level you recorded your track.

After: With the foam in place, the corrective EQ applied, the foam monitor pads swapped out for some IsoAcoustics stands, and the monitor positions adjusted, the monitoring response in the room was significantly improved.Back in the studio, Adrian mentioned that he recorded vocals as well as acoustic guitars and violins (he is an accomplished violinist). His main mics for these purposes are a Rode K2 and a Neumann TLM102, but he also had an AKG C3000. We explained that it is vitally important to appreciate that while the monitoring may sound more even with the corrective EQ plug-in in place, the room itself still exhibits the same strong modal resonances, and these will cause problems when recording voice or acoustic instruments in the room.

Monitoring on good-quality headphones is essential in such situations to remove the room completely from the monitoring environment, and thus provide an accurate appreciation of the actual recorded sound and acoustic space. In this case we suggested to Adrian that he leave the door open when recording (to minimise the room resonances) — when the builders had left, of course! — but also that he might try hanging up some thick duvets, ideally folded double, behind the vocalist/instrumentalist. This can help to tame the lower frequency room modes considerably, especially if the duvets are spaced well away from the wall behind allowing them to act as make-shift bass trapping. This arrangement usually cleans up the lower end of the human vocal range and of acoustic guitars, bowed strings and suchlike to a useful degree. It won’t do much for really deep bass, but then bass instruments are usually DI’d in studios such as this.

By the time we came to leave, the stereo imaging accuracy was improved, the liveness taken off the room, and the overall monitoring response usefully improved, although we were all aware that the corrective EQ approach we had been forced to adopt was a compromise. Nevertheless, Adrian could now get on with composing and mixing his work in the knowledge that it would better translate to other playback systems elsewhere. When doing any serious projects in such a small room, it would also be worthwhile having the mastering done elsewhere, but then that is true of a great number of home studios.