Studio SOS: Michael Tefula

Although it's amusing, the pop-shield improvised from a pair of socks by reader Michael won't work in quite the same way that a nylon stocking would, as Paul points out...

This month we look at monitor placement, tighten stereo imaging and tackle what was probably the world's smallest vocal booth!

Michael Tefula and his friend Arthur specialise in remixing and hip hop, using Michael's small studio setup in a spare room in his brother's house. The studio is based around an Apple iBook running Logic Express. Michael uses a small Soundcraft mixer and a pair of KRK Rokit 6s for monitoring, and his main vocal mic is an SE 2200A. He'd been getting good results from this simple setup, though he'd run into some computer overload problems running Logic Express with plug-ins. He asked if he should invest in a new audio interface, but it turned out that his M-Audio Firewire Solo was working fine and, as far as we could see, was not to blame.

The studio clearly needed some basic acoustic treatment (though it sounded better than it had any right to with bare plaster walls!) and his closet vocal booth really wasn't viable, as it was too small to acoustically treat in any effective way.

Room Response

The speakers were widely spaced and the highly reflective walls of the studio were untreated. Both these issues meant that there was plenty of scope to improve stereo imaging.Michael's KRK Rokit 6 monitors were set up on Quicklok speaker stands, with the tweeters around six inches higher than his head height. when seated. The speakers were also very widely spaced and closer to the corners of the room than I would have liked. Hugh and I played some test material from Hugh's ancient BBC test CD and discovered the bass end to be surprisingly even (given the total lack of acoustic treatment), though the stereo imaging was suffering due to side wall and ceiling reflections.

We both felt that moving the speakers closer together would improve the imaging and avoid any low-end problems resulting from the speakers being close to corners. We also dropped the speaker height so that the tweeters were at ear level, and ended up with the stands much further forwards and inwards, so that they were fairly close behind the computer desk. To test the outcome, we set up a series of sequenced bass notes using Logic's EXS24 sampler default sine-wave (semitone steps over three bass octaves), set all velocities to maximum and then listened for any obviously loud or quiet notes. There was a little variation, but the result was still surprisingly good for an untreated room.

To further tighten up the imaging, we used some two-inch Auralex foam panels at the 'mirror points' on the side walls. As this was Michael's brother's house, we didn't want to glue it directly to the walls, so we decided on strips of double-sided carpet tape. Using this was very straightforward, other than peeling off the backing — Hugh seemed to be the only one able to master the technique! Using this tape, we fixed one panel either side of the monitoring position at head height, to cut down high-frequency side-wall reflections. We then positioned one on the back wall and another on the ceiling above the mixing chair. The tape seemed to hold the foam, once we'd pressed it down firmly, though if the ceiling panel starts to give problems a few map pins should help keep it in place. After fixing the foam, we played back some more test material. As expected, the stereo imaging was sharper and the room sounded a little drier than before.

A Better Booth

If you're renting, you can avoid upsetting your landlord by sticking adhesive strips to the back of acoustic foam panels and using these to mount them, so that there will be no messy glue marks on the walls.Michael had been using a small closet at the back of the room as a vocal booth, with a mattress propped against one wall to help control reflections. The results were, predictably, very boxy, and, as it was a tiny space, there was really no room to put in adequate acoustic treatment. A better solution was to run a mic cable to the adjacent (unused) bedroom, where a duvet or blanket hung behind the singer and an SE Reflexion filter behind the mic would do away with any need to acoustically treat the whole space. Fortunately, SE had given us some Reflexion filters to use on Studio SOS projects and we had the last one in our car. As usual, we had to use the mounting hardware in a slightly unorthodox way, in order to prevent the weight of the Reflexion filter toppling the rather lightweight mic stand being used, but once it was set up we got a great vocal sound that was both clear and dry. We also tried using the mic preamp in the M-Audio interface and found that it produced a perfectly clean sound, which means that there's no need for Michael to record via the mixer.

I noticed that there was no pop shield on the mic stand, but Michael produced a home-made one. At first sight this looked like the usual pair of tights stretched over a coat hanger, but closer inspection revealed that it was made from a pair of mens' socks — which would cause a significant loss of top end. Ordinary nylon stocking or tights material is best for this purpose, though you can also use a metal chip-pan splash guard, or even a fine plastic kitchen sieve at a pinch (or, of course, a pop-shield!).

To test the new room we had to stretch the mic cable to its limits, and Michael asked where he might find a longer one. Of course, the simple solution is to simply hook up two standard mic cables to provide double the length. Cable length isn't really an issue with low-impedance mics until you start exceeding several hundred metres, so running two five-metre cables in series won't be a problem.

Computer Concerns

Michael was concerned that his supposedly fast Mac laptop seemed rather sluggish. Logic took a long while to load if the song included a lot of plug-ins but, more seriously, the disk access seemed unable to keep up if a lot of audio tracks were in use. Michael tends to use a lot of CPU-intensive plug-ins, so I suggested freezing some of the instrument tracks. While this saves on CPU loading, it also places a heavier burden on the hard drive, and it didn't seem to be much help, but I soon discovered the problem: Michael was recording to his internal drive, which was almost full. We moved all of his EXS24 sample library to his external USB2 hard drive, which gave us around 12GB of free space and significantly improved the situation.

Michael had tried to construct a vocal booth inside a small cupboard, with a mattress to dampen the reflections. It didn't work well, as the space was simply too small. We persuaded him to use the vacant room next door instead (as pictured at the beginning of this article), and this brought about a big improvement in the vocal sound.Another issue that was causing heavier than necessary CPU loading was that Michael was using multiple software instruments for separate drum parts that he could then process individually, the main instruments being Stylus RMS and NI's Kontakt 2. With Stylus RMX, once you insert the multi-output version of the plug-in you can use the routing windows seen at the left of each part in the Stylus RMX window to route the parts to different destinations. These parts are designated A, B, C and so, on so you can set up as many as you need, but in Logic you have to have an Aux channel set up in the Environment window to act as a destination for each part, otherwise everything will still emerge in stereo. Once you've created enough stereo Aux tracks (up to eight, in this case) to handle the different parts, you use the input selection button of the Aux track to select the required RMX part as the input source. These now all show up as valid sources in the drop-down menu that appears when you select and input on the Aux channel. Other multi-output instruments operate in a similar way. You can then automate the levels of the Aux tracks, insert plug-ins, and so on, to process the different outputs.

Using a single instrument with multiple outputs can save significant CPU overhead compared with running half a dozen incarnations of the same instrument, each just handling one type of drum hit or loop. Another CPU-saving strategy is to set the EXS24 to load its samples in 32-bit float format, which takes up more RAM but significantly reduces CPU loading. However, in Michael's case there wasn't a lot of spare RAM either. He uses a lot of samples, and when you run out of RAM these are streamed from the hard drive, placing a further burden on the drive's read/write head, so we suggested that he consider upgrading the computer's RAM from 1GB to the maximum 2GB.

In the meantime, to help get Logic to load a little faster, we went to the AU manager and unticked all the plug-ins that Michael never uses. This helped, but the additional RAM will be needed to really speed things up. Because of the need to maintain free space on the 80GB internal drive, the ideal way of working is to keep only current projects on the internal drive and to archive finished or shelved projects onto an external drive. Of course, digital data is never secure unless backed up, so a second (and even a third) external drive — and backing up onto DVD-R — wouldn't go amiss (for more suggestions on backing up your data, see the 'No More Data Disasters' feature in SOS October 2007).

In this shot, you can see the studio after it has been treated, with Auralex foam at the mirror points on the walls and ceiling, and the speakers brought much closer together.The other essential tip we gave Michael to help him avoid computer bottlenecks was to set the buffer size to 256 when recording (to achieve an adequately low latency for soft synths), but then to change it to its maximum of 1024 for mixing: this reduces the risk of the CPU not being able to keep up with spikes in demand.

As Michael was fairly new to Logic, we went through a few basics, including setting up a default template with the required number of audio and instrument tracks, as well as inserting the two software instruments that he usually uses as starting points for a song. We added a few useful screen-sets to the Template and, as the laptop screen is pretty small, I also set up a key command that would 'show channel strip only' in the Arrange window.

Michael was unaware that in Logic's Audio Environment you can select multiple tracks (something you can now do in most sequencers) and then configure their bus sends or their fader levels all at once. Providing you haven't already started adding automation data, this can be very useful, as you can drop all the fader settings by a few dBs if you're starting to run out of headroom. I also showed him how to set up Environment Layers, to simplify navigation. For example, rather than having miles of audio channels in the Audio layer, you can split it into a virtual instruments layer, a bus and output layer, an audio tracks layer, and so on.

Departure

We felt we'd done most of what we could do on the day, and Michael was aware that he would need to get more RAM and a proper pop shield. We also recommended that he invest in a headphone monitoring system with a talkback facility, which he will need to get the best from the new 'vocal booth'. Given his budget, we thought the Samson C Control should be a good option, as this would also allow him to hook up as many as three different sets of monitors in the control room for comparison purposes.

With this advice dispensed, it was time for us to head off — and we left happy that the improvements we had made to the studio acoustics, together with the computer tips and tricks we'd passed on, left Michael much better placed to get on with the business of making music.