Studio SOS: John Farrell

As is the case for many students, Jon's studio is set up in a bedroom. Though the bed dominates the room and will inevitably compromise the acoustics, it is still possible to turn it into a useable mixing environment.

The team take a trip up north to a multi-student household in West Yorkshire.

Jon Farrell first contacted us during our seminars at the 2006 Music Live show and asked us to help him squeeze a bit more performance out of his bedroom studio — which is quite literally in his bedroom. Jon currently lives in an eight-bedroom student house house in Headingley, Leeds, where he and five other students are in their second year of Music Production at Leeds College of Music. All are regular Sound On Sound readers and they seem to have a better understanding of their recording systems than most, but in common with students everywhere, they are rather less familiar with the workings and application of the common vacuum cleaner!

After an introductory cuppa (the name of Jon's Myspace page — see 'Reader's Reaction' box — suggests that this is an essential part of his day!), we moved on to discussing the apparent demise of the plain chocolate Hob Nob and then sampled a few of his special edition chocolate orange Hob Nobs to help damp our disappointment. While we chatted, Jon explained his predicament: "My problem is that my room is relatively small, and the only place I can set myself up is in a corner, against the window. The sink occupies the middle of the left-hand side wall and my bed on the right takes up over half the room, which is around 11 feet square. The ceilings are pretty high as well, so the dimensions are close to the dreaded cube, which I know is not a good thing. I have the monitors set up either side of the window, which means that I spend hours looking into the flat of the girl opposite. I hope she doesn't get the wrong idea!

"My setup comprises Logic Pro running on a 12-inch G4 iBook, going through a M-Audio Delta 410 and feeding a pair of Alesis Active 520 monitors. A Yamaha MG10/2 mixer acts mainly as a monitor level controller. I also have a master keyboard and a couple of hardware synths, specifically a Korg MS2000B and a Studio Logic SL990 Pro weighted keyboard. The other lads are running pretty much the same type of system, albeit with an M Box here and there and sometimes using different software, so they may be able to benefit from anything I learn today. I need to improve the accuracy of the monitoring as much as possible, and I could use some tips on maximising my limited CPU power. A few mixing tips or tricks wouldn't hurt either."

Making The Best Of It

Auralex Mo Pads were used to improve the isolation of the speakers from Jon's desk, while more acoustic foam was placed to the side and rear of the speakers (see photo on previous page) to absorb unwanted reflections and tighten the stereo imaging.Jon wasn't wrong about the room, and though the ideal studio setup is symmetrical, this clearly wouldn't be possible here unless we threw out the bed and made him sleep on the floor. The table really was in the only possible place, so we just had to make the best of it. Jon had his Alesis monitors set up on cubes of polystyrene packing to get them to around the right height, which, while less than ideal, actually didn't sound too bad, though room reflections were diluting the stereo image. We'd brought along some Auralex Mo Pads, so we put these on top of his blocks, which added stability and helped decouple them from the table.

To kill the reflections we used four-inch thick Auralex foam, but we didn't want to stick it directly to the walls, as this was rented accommodation. In the end we used self-adhesive Velcro strip but, as it comes, this isn't really sticky enough to hold securely on either the foam or the walls. We used our familiar trick of spraying a little Auralex adhesive onto the wall and onto the tile where the Velcro was to go. After this had dried for a few minutes and gone tacky, we pushed the sticky backing of the Velcro onto it, where it stuck fast. Using this technique we fixed a vertical 2 x 4 foot panel on each of the side walls and put a horizontal panel above the bed. A further vertical panel was fixed to the wall behind Jon's mixing position, and this could also be useful when recording vocals, as it would reduce reflections coming back from that wall. Recording with the singer's back to the absorbent foam is generally the most effective option. We had to cut a couple of slots in the left hand piece of foam to clear the ends of some small wall shelves and for this we used Jon's bread knife.

The outcome was noticeably better stereo imaging and a generally drier sound in the room so, although we hadn't been able to create an ideal studio setup, we had made a significant improvement. However, we warned Jon not to mix while sitting close to the centre of the room, as in square rooms there's usually a big bass dip in the centre. Having made short work of such acoustic treatment as was practical in the room, we listened to a few commercial tracks played from iTunes, which confirmed my earlier impression that extra bass trapping would be unnecessary because of the plasterboard/studding walls, and because of the absorption properties of a double mattress. The speakers were also fairly small, but such bass as there was came over as tight and even. Jon was quite impressed by the difference that hanging up a few panels of foam made.

Computer Improvements

Our next challenge was to help Jon get the best performance from the limited CPU power of his Apple G4 iBook. He was already using Logic 's track freeze facility to reduce the load imposed by plug-ins inserted into the channels, but a further thing you can do when using Space Designer as your main bus reverb is to use the half-bandwidth mode, as this also halves the amount of CPU power taken and generally makes no subjective difference to the reverb sound — reverb tends to have little going on as high as 10kHz anyway. If you select this option and also click the Preserve Length button, the reverb duration will remain the same. It's also a characteristic of convolution-based reverbs that the longer the IR (impulse response) you load in, the more CPU power is required, so it's best not to pick a longer reverb time than you really need.

Though a large audio buffer size alone won't guarantee infinite performance, it is true that the larger the buffer size, the more stable the computer is likely to be when working close to its limit. When recording, you need the smallest possible buffer size to reduce latency (ideally 128 samples or less, but certainly no higher than 256 samples at 44.1kHz), but when you come to mix, you can afford to set it to maximum. Yet another horsepower saver for Logic users is to open the Preferences inside their EXS24 sampler's Edit window and select the option to load samples in 32-bit float format. This takes up more RAM but lowers the CPU overhead, so you can achieve greater polyphony.

For rented accommodation, velcro can be used to attach acoustic foam to the wall. The velcro wasn't strong enough, so we applied a little of Auralex's foam adhesive to the back. This did the job without copious amounts of adhesive on the wall.

To make working on a 12-inch screen easier, Jon was already using Logic screensets but he hadn't come across the Show Channel Strip Only option in the Keyboard Commands section. This affects the left of the screen and closes all the various parameter boxes so that you get a full view of the channel strip, and it is well worth assigning this to a key command. In conjunction with the Hide Parameters key command, this enables you to use the whole screen to view arrangements when necessary, yet you're only two keystrokes away from bringing up everything else you might need to see in that window.

Though largely fluent with Logic, Jon hadn't come across the Audio Preferences option to have plug-in delay compensation on only the tracks or on 'All' (Tracks, Buses and Masters). By default this comes set to Tracks, which is the best option for recording as it reduces latency and also avoids situations where newly recorded audio plays back out of time with what you've previously recorded (which it can do if you have bus plug-ins that involve a lot of delay). However, if you want to use the buses during mixing for submixing groups of tracks, such as drums or backing vocals, then it pays to switch to 'All' mode, otherwise any plug-ins you insert in a bus will delay the audio through that bus by the delay time inherent in the plug-in. It doesn't matter so much if you only use the buses as effects returns, as a few more milliseconds of delay on a reverb or echo effect rarely makes any difference.

An editing feature Jon hadn't come across was Cut/Insert Time, which is available in the Region menu of the Arrange page of Logic Pro, but sadly not Logic Express. Jon already knew how to set up a skip section in the ruler bar, to temporarily jump over a few bars to see how the song might sound without them, but he hadn't twigged that the Cut/Insert Time/Snip feature would make this permanent, including dealing with any chopped loops to assure the correct outcome. You need to select all the audio regions first, unless you only want to edit specific tracks of the song, and if you use Insert Time instead of Snip, whatever segment you've marked in the ruler at the top of the Arrange screen will become a gap and all the audio following your edit will be moved up correspondingly. You can also chop a section out of one part of the song and insert it elsewhere in the song using the same tools.

Tips & Tricks

Jon then played us a dance remix he was working on, and it sounded very accomplished. However, he was keen to see what alternative production tricks could be useful within this genre so I set up a few of my favourite plug-in tricks for him to try out. One that always works well, and one that I've described before, relies on using the Tremolo plug-in, which by default works as a stereo tremolo that pans the signal from one side to the other at a speed set by the rate knob. However, you can adjust the Stereo Phase knob to get the upper and lower modulation waveforms in line, so that both channels pulsate together more like a conventional tremolo. Adjusting the modulation wave shape to a hard square wave produces a nice rhythmic chopping effect, and if you turn the rate control anti-clockwise you eventually get to a list of tempo-sync'd settings. Eighth-notes and 16th-notes work particularly well. Now the chopping syncs to the song tempo, enabling you to turn pads or backing vocals into rhythm tracks. If you add generous reverb to a backing vocal before chopping it, the resulting texture really adds something special, yet the chopping action prevents the reverb getting in the way. Another variation is to feed a 16th-note chopper into a quarter-note chopper, so that the resulting chopping follows a rhythmic pattern with breaks, rather than being constant.

Our next move was to show Jon how to fake a double-tracked vocal, by copying the vocal to two different tracks and applying Pitch Correction to only one of them. B ecause even a well sung vocal includes some pitch slurring or drifting, combining the treated and untreated vocals creates the impression of two people singing together, with close but not quite identical pitching. You can also use the Detune slider in the Pitch Correction plug-in to move the pitches of the two tracks further apart if you feel the song needs it. You don't need to be a Logic user to try this — you only need a pitch correction plug-in of some kind.

To emulate automatic double tracking (or ADT), as favoured by artists from Elvis Presley to John Lennon, there's a neat trick you can do using Logic 's less processor-intensive Platinumverb (again, many other reverbs can be used to do the same thing, providing they are sufficiently adjustable). Platinumverb has separate pre-delay times for the early reflections and the later part of the decay tail, and it also has a slider that adjusts the mix of the early reflections and the decay tail. By setting the early reflections delay to around 100ms, and then moving the mix slider so that what you hear is almost entirely early reflections, you start to hear a nice slap-back echo, but one that is rather less clinical sounding than if you did the same thing with simple delay. If you can reduce the density and diffusion settings, the early reflections will come over sounding more coarse and gritty, which usually helps the effect. If you can't mix out the contribution of the reverb tail, set the decay time as short as you can, and if your specific reverb won't give you a 100ms pre-delay, simply insert a delay plug-in before the reverb, with the mix control set to 50:50. Jon liked this effect, and one of his fellow students also tried it on a rock mix he was working on and agreed that it created the right effect.

The only possible position for Jon's setup was in the corner, by a window, which is rather less than ideal! This shot shows the foam panels that the team mounted on the walls to the left and right, to help improve the room acoustics.

Weird & Wonderful

Just to sneak in something weird, I went through the steps needed to create a reverse reverb effect similar to that which is achieved when you turn over an analogue tape, record the reverb to another track, then play the tape the right way round again. The result is a 'Buffy'-style supernatural reverb that starts to build up before the word or sound that triggered it. To do this with plug-ins is pretty simple, providing you have a reverse reverb you can use. We used Space Designer and reversed the impulse response for maximum authenticity, then applied this to a copy of the dry vocal part on a new track. By setting the reverb mix to 100 percent wet, you get all the dry vocal on one track and all the reverse reverb on the other. All you need to do then is slide the effected track two or three seconds before the dry track, then fine tune the offset so that the reverse build-up stops exactly when the dry sound starts. When you get this right, the effect is exactly like the old reversed tape method.

Back in safer territory, we tried applying a slow rotary speaker plug-in to the backing vocals, and also using flanging on a backing vocal track before adding reverb. Either of these can be surprisingly subtle, with the movement added by the plug-in being just enough to stop your ears getting used to the sound and tuning it out. Jon had also been experimenting with filtered vocals to produce a radio or telephone sound and, of course, you can do this by band-limiting the track using steep high- and low-cut filters set either side of the band you wish to keep. However, there are often Impulse Responses available that have been taken from real radios or other devices, for use in film post production, and, used in a suitable IR reverb plug-in set to 100 percent wet, these often produce a more plausible result because they include some of the finer artifacts of the original that simple filtering fails to capture. Anyone who uses Altiverb will find some splendid examples on the Audio Ease website.

Another way to achieve a similar effect is to use a fingerprint equaliser such as Logic 's Match EQ or TC Electronic's Assimilator, which allow you to take a spectrum reading from a reference sound, then apply this to one of your own tracks. They do this by looking at the audio spectrum of both the reference and the 'track' sounds, then computing a 'difference' EQ curve to give the 'track' sound the same spectrum as the reference. You can have lots of fun with this by simply recording a radio speech program from a radio or a recording of your own voice from a portable cassette machine using a microphone, then applying its spectrum to your own track. You can also build your own speaker simulators by miking a guitar amplifier, then applying the spectrum to a DI'd guitar sound. This way you can DI guitar via a distortion pedal, which usually sounds dreadful on its own, then use the fingerprint EQ to add the effect of a real speaker cabinet.

Jon didn't seem to use that many software instruments but we did show him how to add a bit of angst to a simple TB-style square wave using Logic 's Phase Distortion plug-in. This uses a short delay and then modulates the delay time using a filtered version of the input signal. The result is reminiscent of FM synthesis or Casio's Phase Distortion synthesis. Where you need a bass synth to be a bit more prominent, or if it needs dirtying up a bit, this is a good plug-in to try.

Compression Counts

Though Jon didn't do any serious mastering on his system, we showed him how to use Logic 's Multipressor as a three-band compressor, using low ratios of 1.1:1 or 1.2:1 combined with a very low threshold (-30 to -40dB), so as to gently squash the whole mix rather than to heavily squash only the peaks, which is how compressors are generally used when mixing. I usually set the crossover frequencies to around 200Hz and 5kHz, so the important mid-range isn't split up. This means that if you feel the sound needs a bit of mid-range scoop, you can always drop the level of the mid band by a dB or two. Following this up with Logic 's Adaptive Limiter helps achieve maximum level without the risk of clipping, but the amount of limiting needs to be modest if you're after results that still sound natural.

A Little Help For His Friends

Jon made notes on all we'd tried, and we also saved the settings in a Logic song so that he could go back and play with them more after we'd gone. After a brief session on vocal EQ (at all times chanting the mantra "broad boosts, narrow cuts"), we looked in on his friends, who were experiencing an odd problem. Whenever they used Jon's SE2200 mic (which worked fine on Jon's system) they got a very asymmetrical recorded waveform via their Mk1 Digidesign M Box, but only in input one. Input two worked fine and other mics also worked fine in input one. They said they'd tried a second M Box and had the same result. They repeated the experiment while Hugh and I watched and, sure enough, that's exactly what happened. Hugh measured the phantom power voltage, both with and without the mic connected, and it was well within normal limits. We could find no reason for this, but if anyone else has had similar experiences with the same combination of mic and interface, please let us know and we'll make further enquiries. As it is, they now know to use the mic in input two only, to avoid the problem.

Having done what we could, and virtually exhausted the supply of chocolate Hob Nobs, we set a course for home, in the vain hope of getting out of Leeds before the rush hour!