READY FOR INSERTION
Insert effects got their name from the “insert” jacks located in the channels of hardware mixers. By plugging into the insert jacks, you’re inserting (duh!) the effect into the channel’s signal path.
In hardware mixers, you’ll find channel inserts after the input preamp and before the fader/panpot circuitry. This allows finding the sweet spot for gain-changing, as effects generally want a diet of high-level, not mic-level, signals; the preamp can bring up the incoming signal to a consistent level for feeding the effect. Also, if the effect generates any noise, pulling down the fader reduces both the signal and the noise. With software hosts, the hardware audio interface has already set the level going into any input channel, so gain-staging isn’t much of an issue.
Insert effects affect only the channel into which they are inserted. Typical insert effects include dynamics control, distortion, EQ (because of EQ’s importance, it’s often included as a permanent insert effect), flanging, and other effects that apply to a specific sound in a specific channel.
Insert effects are patched in series, where the output of one feeds the input of another. One exception is Tracktion 2, where you can insert a “rack” of effects connected in series and/or parallel (see Power App Alley, July 2004 issue). A program may have a maximum number of inserts, such as the eight found in Cubase SX. Other programs allow for an unlimited number, although “unlimited” translates to “how many effects your computer can run at one time without screaming pathetically for mercy.”
In a hardware mixer, insert effects are mostly mono. But in a virtual mixer you usually want to apply stereo insert effects to stereo channels. While it’s often possible to insert stereo effects into mono tracks, there may be unintended results. For example, if you use a mono effect with a stereo signal, perhaps only the left side of the stereo signal on that channel will be processed. On the other hand, some programs won’t let you use a mono plug-in with a stereo channel, and will return an error message if you try. Mono/stereo protocols vary from program to program; check your manual for details.
Master effects are patched in series like insert effects, but they insert in the mixer’s master output section, after the separate channels have been mixed to a stereo or surround signal. As a result, master effects process the entire mix.
These effects usually patch in prior to the master faders, so turning down the faders also reduces any noise contributed by the effect. However, with virtual mixers, there may be post-fader effects slots as well. This is important in case you want something like a reverb or delay “tail” to continue even after you’ve pulled the main fader down. A post-fader slot is also where you would apply dithering (at the very end of the chain). If you applied it before the master fader, moving the fader would re-create a high-resolution audio signal, and defeat the point of dithering.
As master effects process the entire mix, these effects are generally mastering-oriented processes like limiting, multiband compression, noise reduction, EQ, loudness maximization, etc. If you want to add more than one master effect, the options are the same as for insert effects. ‰
For mastering effects, you want really clean-sounding programs . . . which unfortunately, have a reputation for being greedy when chowing down at the CPU Cycles Café. If you throw a lot of quality master effects on top of insert effects, your computer might go on strike (and even file an abuse complaint). One possible solution is the bus (send) effect, which uses one effect to do the work of many.
These effects receive their signal from an aux (send) bus, and therefore process only signals that get on the bus. The signals that show up on the aux bus come from the send controls on individual channels.
For example, suppose you want to add reverb (probably the most common send effect) to the instruments on channels 1, 5, 7, and 10. Each of these channels will have one or more send controls, each of which sends some of the channel signal (the amount depends on the send control knob setting) to a dedicated aux bus.
If a channel has multiple send controls, each can feed its own bus. So you would turn up like-numbered send controls for channels 1, 5, 7, and 10 that correspond to the bus feeding the effect. The aux bus out then goes to the effect input, and the effect out comes back into the mixer so you can hear the processed sound. The processed sound’s level depends on either the mixer’s return level control and/or the effect’s output control.
Cubase SX3 has an interesting take on send effects, as it makes it easy to treat external hardware processors like a plug-in. The only catch is that you need an audio interface with multiple ins and outs to send signals to the processor, then pull them back into the computer again.
Note that because send effects produce sounds that are mixed in with the main channel faders, the faders take care of sending the dry signal component to the main output, so the send effects generally produce only the processed (wet) sound. Blending the unprocessed signals with the processed sound returning from the effect is what gives a good-sounding balance.
Send controls often have a pre/post fader switch. This chooses whether the signal going to the aux bus comes from before the channel fader (in which case the amount of signal feeding the aux bus is constant, regardless of the channel fader level), or after the fader (pulling down the fader also reduces the amount going to the effect). In other mixers, some sends may be permanently set as pre-fader while others are post.
If you want to add more than one send effect, you can usually patch several effects in series: Feed the input to the first effect from the aux bus, and feed the last effect output back into the mixer. However, if you want to have a number of separate send effects, you need an equivalent number of buses. For example, suppose you want to add reverb to some channels and delay to some other channels, some of which have reverb and some of which do not. You would then need two aux buses: one to feed the reverb, the other the delay.
How does this save CPU power? Easy. When using plug-ins, every effect you add requires more computing power. If you want to process several channels with the same effect, it makes more sense to stick a single plug-in into a send bus and use the channels’ send controls, instead of inserting, for example, a reverb into each channel that needs reverb.
INTO THE FUTURE . . .
Most recording software follows the hardware mixer paradigm fairly closely. But as people become more familiar with operating in a virtual world, this is starting to change. For example, Samplitude can treat effects as something you attach to a single audio object; if there are multiple objects on a single track, each can have its own processing.
Another trend is the inclusion of onboard effects in virtual instruments, making additional processing unneeded. However, in some cases you still might prefer the sound of separate plug-ins compared to what’s included in the instrument.
Finally, although you can’t do parallel effects inserts with most hosts, remember that there’s a simple workaround: “Clone” the track, and put one effect in one track, the parallel effect in the other track, then mix both tracks together.
And now you know where to stick those effects. Happy processing!