Studio SOS: Reducing Spill

Although the track limitations of the band's Tascam 788 multitracker restricted their recording options to a certain extent, they also weren't using its facilities in the most effective way.

The intrepid SOS team travel to Nigel Helm-Nurney's London studio to help his band Pombokiwi do battle with unwanted spill, uninspiring guitar sounds, and masses of egg boxes!

This month's column saw us back in north London after Nigel Helm-Nurney called us to help out with recording his band Pombokiwi, which comprises himself, Etienne Baird, Simon Davies and Howie Hughes — although Howie was unable to join us for the session. Nigel plays acoustic guitar, Simon electric guitar, Etienne takes on vocals, and Howie plays the same model of Roland V-Drum kit as I have in my own studio. Pombokiwi's studio is a rather well-converted garage at the rear of singer Etienne's house, where studding walls, plasterboard, acres of dense Rockwool, double-glazed windows and a double door have created a fairly soundproof working environment. The project didn't take long, as all the band members are involved in the building trade in various capacities!

Inside The Studio

The band had made good use of Rockwool between the outer wall of their garage and the inner plasterboard wall, and this had soundproofed the room quite well. However, when it came to acoustic treatment in the room, they had fallen for the egg-box myth...The inside of the studio is festooned with fabric, which helps tame any ringing, but prior to contacting us the band had bought a huge stack of egg boxes to use as acoustic treatment. They'd made a start sticking these to the walls, but called a halt when we suggested they'd make very little difference, especially with all that fabric hanging over them. For the record, egg boxes do produce a little useful scattering at higher frequencies, but they have negligible soundproofing benefits and far better acoustic improvements can be achieved using drapes and/or foam panels. Where did that egg-box myth come from? Nevertheless, the room did sound very usable, with reasonably well-controlled reverberation across the spectrum and no obvious bass problems other than the choice and location of monitors.

The recording system comprises a Tascam 788 eight-track digital recording workstation augmented by a Mackie CFX16 mixer and a handful of microphones, including an Audio-Technica AT3035 large-diaphragm capacitor model and the ubiquitous Shure Beta 58. For monitoring, the band has a small pair of Sony hi-fi speakers (part of a compact Sony music system) and a pair of moulded PA speakers, one of which was hidden behind the drapes, plus there was a powered wedge monitor for foldback duties. They had been using these PA speakers as their main monitors, but as playing back some music soon proved, these produced a hugely coloured, bass-heavy sound, in part due to them resting on the floor close to the room's corners. Even correctly placed, they were not going to be anywhere near accurate enough to use as studio monitors.

The little Sony speakers fared somewhat better, but were not really up to the job of serious monitoring, as they had no real bass extension and were voiced to flatter a budget music centre rather than to be accurate, so Hugh and I suggested that upgrading the monitoring system should be made a priority. The speaker position also needed changing, as the monitors were set up much too far apart. As the Tascam recorder and Mackie mixer were set up side by side on a narrow bench at the centre of the long side of the room, placing the monitors either side of this control centre would be optimum, though raising them above the bench rather than resting them directly on it would give the best results, as it would help avoid reflections from the bench surface.

Here you can see the way Paul & Hugh set up the musicians in the studio for recording. The electric guitarist recorded through a modelling preamp while monitoring at low level through the floor wedge. The acoustic guitarist muted the speaker of his Carlsbro amp while feeding a DI signal to the recorder. In the corner of the room the singer was set up with his back to the walls, singing towards the centre of the room — this allowed his mic to reject spill from the acoustic guitar and the wedge monitor. Although the drummer was not present for the session, his playing would not produce much extra noise in the room, and he could monitor himself through the Roland TD8's internal headphone amp. Diagram: Tom Flint.Photo: Diagrams Tom Flint

Because of the style of music, the band like to record everything at once, including vocals, but this was causing problems with spill into the vocal mic, a nice Audio-Technica AE5400 hand-held vocal condenser mic. Nigel was using a Carlsbro acoustic guitar amp, which he was then DI'ing from the effects send socket, while Simon was using a small Marshall combo, also DI'd via it's rear-panel speaker-emulated DI jack. The drum kit, being electronic, was plugged directly into the recorder (as a mono signal) via the Mackie mixer. The Mackie insert jack trick was used to provide direct outs from the guitar amps, drum kit and mic to feed the recorder. By plugging a mono jack plug only halfway into the insert socket, a direct output can be taken without disrupting the signal path through the mixer.

The PA speaker was used to provide foldback to the drummer, while the floor wedge provided foldback for the vocalist. However, not only did this setup cause spill problems, there was also a nasty ground-loop hum which we traced to Nigel's DI feed from his acoustic combo. This could have been cured by putting a DI box in series, but as neither Nigel nor Simon were completely happy with their guitar sounds, we decided to try to find an alternative recording method. But first we took a look at the spill problem.

Reducing Spill & Improving The Vocal Sound

Because everything was being DI'd, both Hugh and I concluded that a multi-output headphone amplifier fed from the recorder's single headphone outlet would provide the most practical solution, enabling each of the band's members to wear headphones for monitoring. This solution would mean that they would all hear the same balance, but would at least be able to adjust their individual monitoring levels. The inaccurate PA speaker and the wedge monitor could then be switched off, thus preventing spill from those sources, and the headphone mix would be much closer to the actual sound being recorded. Numerous low-cost headphone distribution systems are available from companies such as Samson and Behringer, so this is not an expensive solution. A further possible refinement would be to make use of the V-Drum kit's external audio input, so if the drummer required a different balance, this input could be fed from one of the mixer's pre-fade sends to provide a custom balance of instruments and drums via the drum 'brain' module's own headphone amp.

By setting up the singer in the corner of the room, the drapes helped to reduce the levels of spill from the acoustic guitar. The hand-held condenser mic that had been used for the vocals was also changed for the band's large-diaphragm condenser mic.

Although there would still be acoustic guitar playing and some audible tapping on the drum kit, judicious mic placement would be able to reduce their spill on to the vocal mic to an acceptable level, something we proved by setting up the vocal mic at one end of the room facing a corner with Etienne singing towards the rest of the band. However, we felt that, for the best results, he should invest in a thick king-size duvet (whatever would we do without them!) and hang it from the ceiling about half way down the room to form a semicircular enclosure. He could then stand with his back to the duvet when recording, keeping him far enough from the walls to avoid coloration and damping any reflections in his immediate vicinity, while still allowing him to see the other musicians.

To improve the vocal recordings still further, Paul recommended adding a pop shield to the vocal setup, for a start. In addition, he suggested hanging a duvet from a rail to enclose the back and sides of the singing position, helping to reduce spill while keeping sight lines open.Photo: Diagrams Tom FlintEtienne had been using his live, hand-held Audio-Technica AE5400 capacitor mic for recording until now, and although this is a good mic we persuaded him to use the Audio-Technica AT3035 large-diaphragm condenser mic on a stand instead, although this would need a pop shield, which Etienne agreed to buy as soon as possible. He was used to holding the AE5400 mic and using it very close up, but we explained that this caused handling-noise problems, as well as being more prone to wind blasting, while variations in distance could compromise the tone due to the proximity bass boost. It may be a good technique live, but not in the studio! Maintaining a distance of four to six inches from the AT3035 on a stand, with the pop shield between the mic and singer, would produce better and more consistent results, and testing this (without the pop shield, as we didn't have one) confirmed that we could get a good basic vocal sound with minimal spill.

Acoustic & Electric Guitar Tactics

Tackling Nigel's acoustic guitar problem revealed that he didn't necessarily want a natural miked sound, as his rhythm playing also provided the bass end of the mix — the band has no bass player. He'd been using quite a lot of compression to get a warm, sustained low end, but this ended up sounding a little boxy and boomy. As luck would have it, I'd brought along my Line 6 Pod XT, so I suggested we try using that on the Piezo Acoustic setting using its own compressor.

Straightaway this produced a less woolly sound that still had the strength and sustain Nigel wanted, but rather than suggesting he buy a Pod XT (which is really designed for electric guitars), I recommended he try some dedicated acoustic preamps such as those made by Yamaha and Boss, as I felt these would give him more flexibility. Most such devices also include a sweepable notch filter to help counter feedback, which could be useful in live performance. I also mentioned to Nigel that I'd seen some acoustic guitar feedback stoppers that were effectively large rubber bungs that covered the sound hole. While I haven't tried these personally, they seemed worth checking out for live use, as they don't require any modification to the instrument and they are relatively cheap.

Another strategy for reducing spill in the room while recording was to use a modelling guitar preamp for recording the electric guitar, allowing a floor wedge to be used at low volumes for monitoring purposes. Paul's own Line 6 Pod XT (left) was able to create a sound the guitarist was happy to record with, and because the band's own Korg AX1000G (right) was unable to do so, despite concerted programming efforts, they decided to budget for a new modelling preamp.

One possible method of using a DI processor live is to feed one of the stereo outs to the combo for on-stage monitoring and the other to the PA for the main front-of-house sound. I often do a similar thing with my Pod XT for electric guitar, and at smaller venues this works extremely well. Where the processor has balanced outputs and is being connected to a device with balanced inputs (such as a mixer), balanced leads should be used to cut down on interference and hum.

Simon plays a really nice Japanese Fender Squier Stratocaster, and while his Marshall sounded fine as a live amp, the DI'd sound was pretty lacklustre, with no bite or energy. Again we tried my Pod XT as an alternative, this time using a mildly overdriven patch augmented by a little reverb and stereo echo that I'd set up for my own use. Simon liked this much better, but before putting a Pod XT on his shopping list, he produced a Korg AX1000G guitar processor from under the bench, so we set about re-programming it to see if we could get a similar sound. The closest we got was using the Modern High Gain amp setting, but when switching back to the Pod XT, we always preferred its sound, which was less fizzy and more focussed, as well as having a better sense of existing in a real space. So I think he might be going shopping in the very near future...

Simon was also experiencing the inevitable single-coil pickup hum problems that plague Strat players. There was one fluorescent light in the studio that could be turned off, but the proximity of so many power amplifiers and other devices with transformers and wall warts, combined with the relatively small size of the studio, meant he couldn't really move far enough away from all the potential sources of interference to avoid the hum altogether. Being a Strat player myself, I suggested he try Kinman replacement pickups, as I've found these to be excellent both for recording and live performance. These are humbuckers designed to look and sound like the single-coil pickups they are designed to replace, and they are very effective in reducing hum without compromising tone. Simon, being from New Zealand, was wary of anything Australian, but seemed prepared to check them out anyway!

Tweaking The V-Drums

The Roland V-Drums were the least troublesome of the sound sources, though I was told the snare drum was re-triggering sometimes when adjacent drums were hit. On checking through the menus, I found that no crosstalk cancellation had been dialled in for the snare drum, so I adjusted this and the problem went away.

The band also felt the kick drum wasn't coming over strongly enough in their mixes, but this was a simple matter of choosing a kick with more attack to it, and the TD8 'brain' module has lots of kicks to choose from. A little rebalancing of the drums sorted out a usable, basic sound, after which I set up a shorter, bright reverb using the 'Bathroom' setting with 'Glass' surfaces.

Nigel had been adding compression to the drum kit via the Tascam's effects section, and the fast attack time he'd been using had robbed the drum sound of some of its impact. My feeling was that, as the V-Drum sound was fully produced right out of the box, it shouldn't need any compression at all and only minimal EQ — possibly a bit of top lift to liven up the cymbals and snare. So, we bypassed the compressor in the Tascam 788 and immediately improved the drum sound.

Nigel also asked about adding overall reverb to the drum sound while mixing, but we advised against this on the grounds that kick drums should be left relatively free of reverb and the TD8 can provide all the reverb needed on the other individual drums and cymbals. Although track limitations in the Tascam 788 might mean recording the drums in mono sometimes, this shouldn't be too detrimental to the overall sound.

The sound that the band were getting from their Roland V-Drum kit wasn't working in their mixes, so Paul set about choosing different sounds and balancing them more suitably. He also adjusted the effects settings to avoid having to tie up the multitracker's effects for the task.

While on the subject of compressors, we also suggested not compressing the DI outputs from guitar processor boxes such as the Pod XT, as these devices include all the processing needed. Where such devices have to be recorded in mono to save tracks, a little stereo reverb may be added while mixing to restore a sense of space, but that's pretty much all the processing that should be needed. We set up a nice generic reverb on the Tascam 788 that was useful over the 1.8-2.2s range, and that worked nicely on vocals and other sounds that needed it. I then fine-tuned the vocal compressor setting using a 4:1 ratio, a fast attack and a release time of around 200-250ms.

Sadly, the Tascam compressor doesn't have a gain reduction meter, which is a serious omission on a device designed to be used by musicians who may not have a lot of signal processing experience, so the threshold level had to be set entirely by ear while listening out for the artefacts of over-compression. It isn't possible to create a compression preset where the threshold doesn't need adjustment, because the correct threshold setting depends on the recording level of the track and also on the dynamics of the recorded signal. Our advice was to lower the threshold until the vocal level sat evenly in the track, and not to take it any lower than necessary.

As I'd only taken one Pod XT with me, and because we didn't have a multiple-output headphone amp (or multiple sets of headphones come to that), we set up a test recording using minimal speaker monitoring levels, where Simon used my Pod XT for the electric guitar and Nigel DI'd from his Carlsbro combo as usual, but with a jack inserted into its phones socket to kill the speaker output. Etienne needed a little vocal monitoring from the floor wedge, but we were still able to make a recording with very little spill. We were confident that with a suitable headphone system in place, the end result would have been even better. The acoustic guitar didn't sound too good, but using a suitable acoustic processor instead of the amp would fix this. Simon was happy with his guitar sound using the Pod XT, and the vocals came over clear and natural sounding, needing only a little compression and reverb to make them sit nicely in the mix.

Replacing The Bass

During the session, the band were also discussing what to do about their lack of a bass instrument, other than recruiting a bass guitarist. Suggestions ranged from using synth bass pedals to overdubbing a bass line using a guitar via an octave-divider pedal, but a further opportunity is available via the V-Drum kit. This kit includes a General MIDI sound module as well as other instrument samples, and it also has a simple in-built sequencer that could be used to play pre-programmed bass parts, though this would mean the drummer following a click using headphones.

Own up — who's scoffed all the biccies?A further option is to use the pad sequencing capabilities of the TD8 brain where, for example, you could set up a simple octave, root and fifth bass-drone sequence to be triggered one note at a time every time the kick drum is hit. If a bassy enough sound was used, this might underpin the rest of the mix without its musical simplicity becoming too evident. The band decided to explore these options when they next met up with the drummer.

Another solution occurred to me after we had left, again from the Roland stable. Three or four years ago, Roland/Boss introduced a line of GK effects boxes designed to be used with their GK split pickup system, allowing independent processing for each string. One of these processors was an octave device that could be switched to operate on all the guitar strings or just on the lower strings. By fitting a GK pickup to the acoustic guitar, it would be possible using one of these processors to generate bass drones from, say, the bottom two strings of the guitar only. The same thing could be done in a more sophisticated way by using something like a Roland GR33 guitar synth fed from the pickup, again generating sounds based on the output from the bottom couple of strings only. In this case, the GR33's Hold pedal could be used to sustain bass drones indefinitely.

The band's main problems were their monitoring speaker system and their lack of a suitable multi-headphone foldback monitoring system, but we also demonstrated that dedicated DI processors could produce a better-sounding result than using DI feeds from their existing stage amplifiers. Even with such a simple recording setup, they could record together as a band and produce recordings of adequate sound quality to be used for making independent CD releases. The main obstacle, unwanted spill into the vocal mic, was largely overcome by the DI-plus-headphone monitoring approach, though we also stressed the importance of getting the right sounds at source so as to minimise the extent of further processing required.