The SOS crew head to the 14th floor of a London tower block to help a pair of readers improve their drum sound.
The Loose Cannons comprise Kaiser Saucy and Lord Fader, and the latter invited us to his lair (we thought it was going to be a hollowed out volcano with integral mono-rail and sharks with lasers!) to help sort out a few recording problems they'd been having. The pair have been DJs, promoters, remixers, writers, and producers for approaching four years now, but thanks to a brief and volatile marriage to a major label (recently divorced due to irreconcilable differences), the boys have at least been able to up their profile considerably in the last 18 months. Their new album 'Perbe is due out in late spring 2005, the first fruits of which, including the single 'Raw', are doing the rounds now.
This particular situation was a bit of a first, as Lord Fader (his stage name, you understand, and just Fader to his friends!) has his home studio in a 14th-floor flat. Fortunately Hugh and I were able to take the elevator so as to avoid doing battle with the Death Stair... After acclimatising ourselves to the rarefied air of the South London stratosphere, we were rewarded by the appearance of some excellent Sainsbury's triple-chocolate biscuits backed up by the now mandatory Hobnobs and a nice cup of hot tea. This was going to be a good day!
Fader's studio is based around a Mac G4 running Emagic Logic v6 using a MOTU 828 as an audio interface and a Korg Z1 as the main keyboard — all very familiar. He also had a Mackie 1604VLZ Pro mixer hooked up, but he was monitoring from the main outputs and re-patching just about everything for each new job, so we thought it might be best to rationalise that aspect of his system first.
His monitoring was via a pair of Alesis Monitor 1s, which were positioned rather asymmetrically in the room, so Hugh sensibly suggested that he move everything about a foot to the left to at least get the right-hand monitor away from the corner. DIY acoustic treatment and curtains gave the room a reasonably neutral sound, and the rest of the setup was pretty much OK.
While Kaiser assembled the drum kit, Fader and I set about re-knitting the mixer wiring so that the monitor amp was fed from the mixer's dedicated control-room outputs, and so that the main stereo out of the MOTU 828 fed into the two-track inputs on the Mackie. This configuration has the advantage that the control room section is isolated from the rest of the mixer if the two-track input is chosen as the only monitoring source, leaving the rest of the mixer (in this case two main outputs and four subgroup outs) free to feed into the MOTU's six line-input jacks. As this particular MOTU interface has two mic/line inputs and six line-only jack inputs, wiring the Mackie this way meant that all eight audio inputs could be fed with mic signals simultaneously if necessary (two direct and six via the Mackie).
Fader had made a note to ask us about hum and ground loops, as he was experiencing some problems with his setup, although this actually improved a little after the rewiring. However, all his jacks were unbalanced, and as both the Mackie mixer and the MOTU 828 have balanced line connections, using balanced jack cables will improve the situation; so balanced cables went on Fader's shopping list.
Once the drum kit was set up facing the recording setup to give us good access to the mics, we decided to remove the front kick-drum head, as the small hole that had been cut into it wasn't large enough to allow us much flexibility in mic positioning. Also, the damping (folded blankets) needed modifying to get a decent tone out of the drum. This done, Kaiser went on to apply a little damping to the upper and lower heads of the two toms using pieces of black gaffer tape. (Why is gaffer tape like The Force? Because it has a light side and a dark side and it holds the universe together!)
Knowing that we were going to be recording a drum kit, we took along a set of CAD drum mics that we'd had in for review — this seemed like an ideal opportunity to give them a practical workout. It turned out that the included drum shell clamps wouldn't fit the snare drum in this kit, but everything else went up smoothly and we used a standard boom stand for the snare mic. The clamping hardware worked perfectly on the toms, supporting the close mics a couple of inches above the edges of the drums facing inwards. While I fitted the tom mics, Hugh rigged the two electret overhead mics from the CAD kit as high as practical in the form of a spaced pair using two more standard boom stands, and finally rigged up the snare mic.
Initially, we tried resting the kick-drum mic on the pillow inside the drum, but the sound we obtained was very dull, so we used a second mic stand to support the kick mic about halfway inside the shell and slightly to one side of where the beater hit the head. The overhead mics were fed to the two mic inputs of the MOTU (and fed phantom power) while the snare, kick, and tom mics were patched into the Mackie mixer and then routed to four of the MOTU line inputs via the mixer's group buss outputs.
After setting the levels, we made a test recording directly to four mono tracks and one stereo track in Logic, with Kaiser playing the drums. We had no chance to monitor the sound prior to playback, as the drum kit was right next to us, but with no EQ or other processing, we achieved what was actually a very decent drum sound, with quite adequate separation. However, the separation on the kick-drum mic wasn't as good, and it sounded just a bit too leaden and soggy as well. Messing around with the mic positions helped a bit, but we eventually decided to switch to a Sennheiser MD421 instead. This produced a better balance of weight and click. Fader said that when he'd tried recording the kit, he'd had trouble getting any separation, and that all the mics had picked up more or less the same thing, so my guess is that he didn't have his close mics close enough.
After checking all was well with the CAD mics, and having proven that it was certainly possible to get a good drum sound in the room, we re-rigged the kit with the mics that the guys had available in the studio, to try to get similarly good results. We ended up using an SM57 on the snare, the MD421 on the floor tom and something that looked like an old AKG D12 on the kick. Very much a case of The Usual Suspects! The overheads were initially an old AKG C451/CK1 and a newer AKG Blueline SE300 with CK91 cardioid capsule. However, the C451 seemed to have stopped working, so we used the SE300/CK91 together with an Audio-Technica AT4033 large-diaphragm mic instead.
These are obviously quite different microphones, but once the levels were matched, the subjective end result was actually rather nice. Capacitor mics are always preferred in this role, as their extended high-frequency response captures the transients and cymbals more effectively than dynamic mics. The shockmount for the AT4033 had seen better days — the upper elastic support band had broken, and the cradle elastics had stretched. Fader had bodged together a replacement system using several large elastic bands which seemed to work well enough, even if it wasn't too pretty. He had also attached a homemade pop shield that seemed to work well for vocals too.
If anything, the hi-hat element of the drum sound could have been a little louder, but they didn't have a spare mic to record this separately, so it was picked up mainly in the overheads. Repositioning the snare mic may have helped, along with some judicious equalisation, to emphasise the hi-hat more, but for real control a separate hi-hat mic would be the best solution. Any small capacitor mic should be OK for hi-hat recording provided that it is positioned just above or just below the point where the two cymbals meet, otherwise the rush of air as the hi-hats close could upset the mic's performance and compromise the sound.
The drum sound we achieved using their assortment of mics was a little tonally different to that achieved with the CAD mic kit, especially in the overheads, but still pretty good. However, we felt the kick needed more definition to fit in with their urban style of music, so after checking the mic position again and deciding we were getting about the best sound we were going to get, I tried a few Logic plug-ins to see what could be achieved.
I have to admit that I was feeling quite nervous at this point, as we'd been playing the drum kit on and off for over an hour and I half expected to see a crowd of locals with torches and pitchforks coming to get us. After all, this was a flat with neighbours above, below, and to either side! Fortunately, Fader had already warned his neighbours of our visit and it seemed they were either incredibly tolerant or out at work during our recording session.
The starting point for processing was to use the channel EQ, where around 6dB of boost at 76Hz (the kick drum's fundamental) combined with a 5.1kHz boost of 8dB (to bring out the click) started to work quite nicely. A little mid-range cut at 280Hz to tame any boxiness also improved things, but it is important when doing this to listen to the kick sound both soloed and in context with the rest of the kit, as the overheads make a very significant contribution to the sound. The other trick I tried was to use Logic's Exciter plug-in to bring out the kick mic click, as this can synthesise high-end harmonics based on what's going on lower down the spectrum. Again this turned out to be quite successful. The same plug-in has also proven useful in the past for rescuing dull snare tracks.
As a final kick-drum experiment, I opened Logic's SubBass plug-in, which creates frequencies an octave below those selected for analysis. You can think of this as being like the Exciter plug-in in reverse — instead of synthesising harmonics above those that are actually there, it synthesises sub-harmonics below those present in the program material. There are two sets of controls, so that you can generate two lots of sub-bass from two parts of the original audio spectrum, and by turning down the dry sound you can adjust the parameters until you get the kind of depth you're looking for. Once I added the processing back into the original kick-drum sound, this gave a welcome impression of extra depth without making the kick drum sound unnatural. I could see this was one plug-in the guys would be playing with some more in the near future, as they were quite enthusiastic about the results achieved.
Fader had bought a TLA5060 compressor to use on his final stereo mixes, but to do that he'd been recording the MOTU output, via the compressor, to a stand-alone CD recorder. I explained that, as he had a couple of spare inputs on his MOTU 828, he could send his mix through the compressor, then back into the 828 and route it to a new stereo track in Logic. All you have to do is set the record levels, then record the mix directly into the Logic Song — remembering not to route this track to Logic 's main output while recording, though, otherwise the audio will feed back into itself and cause a howl. Then again, these guys would probably even find a way to use that!
Rather than record entire drum tracks, the approach taken by the guys was to record sections (often against a click) to be used to create loops, so toms were rarely used. On a normal session they probably wouldn't have set up the toms at all, and if you don't need them, leaving them out reduces unwanted ringing. The snare sound was deliberately big and ringy, so we went on to try overall compression using Logic 's own compressor just to show how different settings can help create different feels.
Conventional settings gave added weight and evened out the sound in a fairly predictable way, but what really seemed to hit the spot was using fast attack, fast release, and a ratio of around 4:1 (hard knee, peak sensing) to make the kit sound pump in a magnificently trashy way. At gain reduction readings of over 10dB, the sound got quite animated in exactly the right way to create exciting loops.
Fader had tried adding reverb to his own recordings, but didn't like the washy sound he was getting. Just to demonstrate the principle, I used Logic 's Platinumverb set to almost its shortest possible decay time, and with the Balance ER/Reverb setting at around 65 percent early reflections, 35 percent reverb tail. With full brightness, and the bass controls turned down, this gave us a short, bright sound that worked well to add weight and presence but without adding any noticeably cloudy reverb at all. I would have preferred to try some of the Space Designer reverbs, but these weren't available in that particular version of Logic. Nevertheless, the exercise demonstrated the production value of using a short, bright reverb to liven up a drum sound without diluting the impact.
Where possible, you should record and mix at 24-bit resolution, reducing to 16-bit at the last stage in a CD-burning program such as Roxio's Jam. However, Fader works at 16-bit throughout so as to maximise track count, and, given the deliberately grungy, urban nature of his mixes, this is perfectly fine. However, the issue of CPU and disk resources had still been worrying him. His songs are all mixed as audio tracks, and the one he opened by way of illustration turned out to have 41 audio tracks running together. This was enough to occasionally knock the disk's data transfer capacity over the edge and cause glitches, as confirmed by Logic 's performance meters. (He was using an external 7200rpm Firewire drive for audio, though tests with the internal drive had produced similar results.) Sensibly, he'd used the track freeze function to conserve CPU power on those tracks that included a lot of processing, but that doesn't help with disk access problems.
However, I noticed that several of his unfrozen tracks comprised very short repeating audio loops and segments, which make the hard drive work much harder than pieces of continuous audio. Freezing the busiest of these (which renders them into a single temporary audio file), got the disk meter reading down to around 70 percent, which I felt rather happier with. The other simple trick is to zoom the screen to show the entire song, so there are no processor-intensive graphic re-draws as the cursor reaches the end of the screen — every little helps.
Both Fader and Kaiser reported that they'd experienced timing problems when composing in Logic, and Kaiser, who works mainly on his G4 Powerbook, said that with his Metric Halo 2882 Interface he was hearing too much latency to allow him to record anything unless he monitored only the source while recording. Looking at Fader's Logic settings, it turned out that he hadn't switched on the plug-in delay compensation in Logic 's Audio preferences menu, and when I checked the same thing in Kaiser's Powerbook, it turned out his was switched off too.
I still don't understand why this is switched off by default when Logic is installed, but it is, so you have to turn it on manually. Once engaged, plug-ins loaded into track insert slots don't throw out the mix timing due to the processing delay within the plug-ins, as the timing of the other tracks is delayed to match. However, be aware that this doesn't apply if plug-ins are inserted into either the buss, master, or aux channels.
In addition, Kaiser's ASIO buffer size was set to 1024 samples, and a setting this large results in audibly high latency. I managed to reduce it to 128 without detriment to his system performance. Logic had to be restarted for these settings to take effect, but a quick test with his guitar proved that the latency was now low enough to allow software monitoring.
Fader was using a buffer size of 512, which is good for stability but slightly too high for real-time monitoring or playing of virtual instruments. My usual strategy is to use a buffer size of 128 or 256 when recording, as this gives an adequately low latency, but then as the mix builds up and the CPU or disk access starts to struggle, I set it to a higher value, which gives good stability when mixing and adding effects. Usually you can leave most of the CPU-hungry effects off until you come to mix.
We finished the day with a couple more practical questions, both of which were related. Normally Fader records vocals in the same room as the computer, but we felt we could improve both the sound and spill by having the singer stand in the curtained corner furthest from the computer, facing the computer. This would place the cardioid vocal mic with its least sensitive rear axis to the computer, and the curtain would also mop up some reflected sound giving a drier vocal.
Fader also records a classical acoustic guitar in this room, and the same strategy would apply there — set up with the back of the mic directed towards the computer and as far away as possible. He was also unsure as to the best mic position, and though the textbooks, including some of mine, suggest aiming the mic at the point where the neck meets the body, there's no substitute for monitoring through headphones while you move the guitar relative to the mic until you find that magic sweet spot — every guitar is different! One of the small-diaphragm capacitor mics would be best for this application.
The other related issue is that of computer noise, and it is possible to quieten G4s to some extent by putting a folded towel or rug over them, leaving the front and rear open but covering the top and sides right down to the floor. This doesn't obstruct the airflow and so shouldn't cause any heating problems. Also, hanging a folded towel in front of the computer, but again leaving room for the air to circulate, reduces the noise level noticeably, especially the high-frequency 'whine' components. Finally, as the back of the shelf holding the computer was open, with a hard plaster wall behind, we suggested hanging foam or heavy fabric on the wall itself to cut down on reflected sound from the rear of the machine.
That about wrapped it up for our visit too, and as the biscuits had all but gone we decided it was time to head for home!
The team also encouraged us to try out recording more stuff (percussion, bongos, claps, ambient vocal 'vibe' tracks) all while someone plays the drums. This added loads to the recordings, and was much better than trying to add it all later. I think worrying about getting miking 'right' puts off too many people, but it's simple really, and once you start, you end up miking up everything to see how it sounds. You just need enough people to make the right noises...
Thanks again for all the tips!"