You can achieve a lot with just a computer and a soundcard, but when you want to accommodate external sound sources or record with high-quality microphones then you really need to add a hardware mixer to your setup. Here's how to hook it all together for maximum flexibility.
Although it is possible to compose and mix music entirely on a computer, most practical music recording systems include some external hardware elements, which is where life can get confusing. For example, you may wish to use a high-quality signal from a capacitor microphone rather than relying on your soundcard's mic/line inputs and bundled plastic microphone, which means either using a separate recording channel or mixer. This important subject area was covered in detail in SOS April 2002, although the use of a mixer in this capacity is also covered briefly here. What may be less easy to sort out is the best way to integrate external MIDI instruments and effects boxes into the system.
For users with a simple stereo-in, stereo-out audio interface, the solution shown in Figure 1 is both simple and affordable. Small desktop mixers are no longer high-cost items, and even a modest mixer provides a convenient means to combine the outputs from your soundcard and the outputs from your external MIDI sound sources. The number of inputs required depends on how much hardware you want to plug into them, but you don't need mic inputs on every channel. Because systems have a habit of growing, try to get a mixer with a few more channels than you currently need.
The mixer output carries a stereo analogue mix of your music that can be recorded to any mastering machine of your choice or even back to a new audio file via the stereo inputs of your audio interface/soundcard. The system shown here uses a conventional hi-fi amp and speakers as a monitoring system, and this also has the advantage that the signal routing to and from the stereo recorder is handled by the amplifier. If you're new to recording, you can simply plug your mixer outputs into any spare Aux, CD or Tuner inputs on the back of your hi-fi amp.
A conventional hardware effects box can be used via the post-fade aux sends and aux returns of the mixer and this allows you to add different amounts of effect to each sound being mixed. However, what you have to bear in mind when using a soundcard with stereo outs is that all the computer audio tracks (and maybe some MIDI parts too if you're using your soundcard's synth chip) emerge ready mixed, so there's no way to add hardware effects to some tracks of your computer audio mix and not to others. Because of this limitation, your computer audio tracks need to be treated using plug-in effects and processors before they are sent out to the mixer. The MIDI tracks may or may not have effects available to them depending on the design and type of soundcard you're using.
Provided that you have one spare mixer channel with a mic input, you can use it to feed a signal into the computer for recording, thereby saving the expense of a dedicated voice channel or stand-alone mic preamp. In many home studios, vocal and instrumental parts are overdubbed one at a time, so a single input is often sufficient. On a standard mixer with no multiple busses or other fancy routing options, the easiest way to use a mic channel for feeding a soundcard is to turn its fader fully down and then use the pre-fade send control to send the mic signal to the mixer's pre-fade output jack. This is normally used to set up monitor mixes, but in the smaller studio it can be fed directly into the soundcard input as a means of routing the mic signal separately. Essentially, the mic signal goes through the mixer channel, via the pre-fade send and out of the pre-fade send jack without interacting with anything else the mixer may be doing, almost as though it were going through a separate piece of hardware. All you have to do is keep its channel fader down when mixing and ensure that the pre-fade aux send is turned down on all other channels when recording.
A further tip here is that you can turn up the mic channel fader when recording to hear the mic signal in the stereo mix (which you'll need to monitor via headphones while overdubbing). By monitoring the mixer channel in this way and by switching off 'through monitoring' in the computer, you'll avoid the distracting effects of any system latency, though you'll also lose the ability to monitor the effects of any software plug-ins being applied to the input signal.
Figure 2 shows a slightly more flexible solution for users with multi-output audio interfaces. If you have eight physical analogue outs, these can feed eight inputs of your analogue mixer, and because these eight signals are independent of each other, different amounts of effects can be applied to them. Furthermore, the mixer insert points may be used to place signal processing devices, such as compressors or equalisers, into individual signal paths.
If you have more than eight tracks of audio/MIDI on the computer, then clearly some mixing must still take place within the computer to reduce the number of audio streams to eight. However, mixing sounds into logical groups still leaves you plenty of flexibility for adding effects. For example, you might use one output for your lead vocals, a pair of outputs for a stereo mix of your backing vocals, one for bass parts, one for guitar or lead instruments, one for rhythm parts (or subsections thereof) and so on. If you take a little time to think about what effects you are likely to add when mixing, it soon becomes obvious which sounds can be grouped and which need to be kept separate.
One important point to note is that when you're using an external mixer with a multi-output audio interface, the panning assignments on the computer's internal mixer are more likely to be used for routing than for actual panning. In the example where the lead vocal is on the first output, this would be achieved by assigning the signal to interface outputs one and two, then panning the vocal channel hard left in the computer's mixer page so that all the signal is routed via the odd-numbered soundcard output (ie. number one). That's because the routing in most software mixers follows the hardware mixer convention of using the pan control to adjust the balance between odd/even-numbered output pairs. When the vocal signal reaches your hardware mixer, it is simply a mono source that can be panned anywhere you like in the mix using the mixer's pan knob. Where stereo mixes have been set up in the computer using an odd/even output pair, the hardware mixer channels into which these are routed need to be panned hard left and right to maintain the original stereo perspective.
The pre-fade send trick used to get a mic signal into a computer soundcard works exactly the same as described earlier, but if you want to record more mics onto separate tracks at the same time, you'll either need a mixer with more pre-fade sends or you'll want to use channel direct outputs, where fitted, to send the channel signal directly to the soundcard input. For any larger-scale projects, a recording mixer with multiple output busses is recommended.
Because there are so many good software plug-ins available today, using a two-in, two-out soundcard or interface doesn't impose too many restrictions on how you can mix your tracks. Even so, it does preclude you from using hardware effects and processors to treat individual parts of your mix, other than elements coming from external hardware such as MIDI modules.
An audio interface with eight outputs offers a good compromise between flexibility and cost, as it allows you to separate important sounds for processing with external effects and processors. In each of these cases, although it need not be large or expensive, an external mixer is still needed to combine the sound outputs of your computer with the outputs of external MIDI hardware instruments.
Using a high-quality voice channel with a digital output is a good bet if you record into your computer one part at a time. However, once you have a voice channel, you may only really need to use your mixer for monitoring the outputs of your soundcard and hardware MIDI sound modules. In which case, it could be more cost effective to go for a mixer without mic preamps — in other words, a line mixer.
Using a line mixer, you can leave your voice channel connected to your soundcard's digital inputs, and use the line outputs of the voice channel to feed into the line mixer for direct monitoring purposes.
If you're one of those people who only run their MIDI modules live while composing tracks, recording their audio to disk for mixing purposes, then you may find that you don't really need EQ or aux sends on your line mixer either — after all, lots of sound modules now have EQ and effects built in, as do most sequencing packages. In that case, you can probably get by with only pan and level controls on each channel.
There are some seriously affordable units available which provide just these facilities, from manufacturers such as MAM, Midiman, Behringer, and many others. Because such mixers only need a small number of controls, you can get a large number of channels into a small physical area, which is great if you're short on space — for example, MAM offer a eight-channel stereo line mixer in a 1U rackmount box!
The only fly in the ointment here is if you want to hear reverb on your monitoring signal while recording through your voice channel. If you need this, then you'll want to make sure you get a line mixer with at least one effects send. Mike Senior