But why does tape make drums sound so good? The answer is tape compression, which occurs when the recording levels are driven hard and the hot signals hit the tape. Saturating tape causes transient signals to get chopped off, leaving a fuller sound because more of the body of the source or room tone is squeezed into the mix. In effect, air is sucked into the sound, the “smack” is chopped off, and the body is turned up. This is part of the secret to John Bonham (Led Zeppelin) and Roger Taylor’s (Queen) thunderous tom sounds.
No Tape? No Problem!
So you don’t have an old Studer or Ampex tape machine. Does that mean you are S.O.L? Not necessarily. While the best way to get the sound of tape is to record with tape, applying bus compression to your drums can produce a similar sound. The best part? All you need is a hardware or software stereo compressor.
What is bus compression? In the case of drums, this technique involves busing (or summing) your drum tracks together by routing some or all of the individual tracks to your stereo bus—creating what is known as a submix—and then applying compression to the submix.
Many people apply a small amount of compression at the stereo bus (also called the left/right bus or the mix bus) to glue the overall mix together before mastering, and that’s cool. However, applying processing at the mix bus affects all of the tracks, and that’s not what we are trying to do here. We want to compress the drums independently, so we need to make a submix using a separate stereo bus.
Serial Versus Parallel
Before you start compressing the submix, you need to decide whether to use serial or parallel processing. Serial processing, in this context, is when you apply compression to the drum submix, and the compressed submix will comprise the song’s total drum sound. In other words, the original drum tracks will not be sent to the mix bus. Conversely, parallel processing means that you keep your original drum tracks in the mix bus, and blend in the compressed drum submix to add some punch and vibe. If you want to keep your drums sounding fairly natural, you might want to take the parallel processing route. However, if the compressed submix is exactly the drum sound you want for your mix, just say “damn the torpedoes,” and use serial processing.
I prefer serial processing, as it’s a much simpler technique. On the other hand, parallel processing provides more options when it comes time to mix, such as allowing you to make your compression a little more transparent. Parallel processing also allows you to leave certain elements of the drum kit out of the submix—such as the overheads—that you may not want to hit with the bus compressor. There is no right or wrong here. Trust your ears, and use whichever technique meshes best with your workflow.
Note: With parallel processing, it is important to make sure you have virtually no latency on the processed submix, as latency causes phasing problems (or even actual delay) between the submix and the original drum tracks. This is rarely a problem when working with hardware mixers and compressors. Plug-ins, however, inherently have some latency, so make sure you use delay compensation in your host program.
Setting Up and Dialing In
With the drums subgrouped to a stereo bus, you can now apply compression to the submix. You have a few options in terms of physically implementing the compressor. If you’re working completely in the box, you can simply add a plug-in compressor to the subgroup track. If you’re using a standalone, stereo hardware compressor, you will need to either insert it into the master buses, or send the master buses’ outputs to the unit and return the processed signals to two channels.
Whether you’re using software or hardware, you’ll want to use the compressor in stereo mode (see Figure 1) to avoid any undesirable ping-pong effects within the stereo field. Stereo mode tells the compressor to treat both sides the same when the threshold is broken on either channel. This way, louder sounds panned to either side won’t get compressed while the other side is untouched. Again, while there is no right or wrong in applying compression, you should understand your compressor—and how it will affect your drum sound—before you start twisting knobs and pushing buttons.
I like using a fast attack, as this allows the compressor to clamp down on the transients of the drum strikes. Compressing the smack of the drum fattens the tone the way tape compression does, because you are pushing down the loudest part (the transient), and getting it closer in volume to the room tone. This creates more headroom (because the loudest part has been compressed), which therefore allows you to turn the signal up with the makeup gain, emphasizing more of the body of the drum and the tone of the tracking room.
If you’re using a fast attack, I also recommend using a medium to fast release. Once you’ve manhandled the transients, you don’t want the compressor to hang out too long and squash the body, as well. In fact, you can actually thin the sound out if you compress too much of the signal. Slower releases can be more useful when parallel processing, because the “over-processed” sound is really more of an effect that you’re blending with the original tracks, rather than a final drum sound.
Lower compression ratios and higher thresholds can help you create your desired effect without making things sound too processed. Try somewhere between a ratio of 2:1 and 4:1, and a gain reduction of 2dB to 4dB. That being said, certain mastering limiters (with inherently higher ratios) could be effective when used sparingly. For example, the Waves L3 (see Figure 2) can help create a very thick drum tone if you don’t overdo it. For a more exaggerated, over-compressed sound, higher ratios and lower thresholds with more gain reduction are the way to go, as this heavy-handed approach can be very effective when parallel processing.
Keep in mind that a compressor’s parameters greatly influence one another. For instance, a fast attack allows for makeup gain, while a slower attack might not leave enough headroom. Also, an effect somewhere in the middle can be reached by using higher ratios and higher thresholds (like the L3 example), or lower ratios and lower thresholds.
You must also be aware of all the equipment in the signal chain. Every piece of gear—be it hardware or software—imparts its own sonic characteristic to the tracks. Using the exact same settings on different equipment will yield different results. In the end, just turn the knobs until it sounds good—with “good” meaning “appropriate for the song” or “pleasing to your ears.” You don’t have to clone what your favorite producer deems as a good drum sound, unless that’s your idea of bliss, as well.
Words to the Wise
It’s common to apply compression to individual drum tracks in addition to the submix. One approach is to get the drums sounding the best that you can before you create the submix, allowing the bus compression to be the icing on the proverbial cake. However, you may find that you don’t need as much processing on the individual channels once you get the submix in place. For this reason, you might want to try dialing in the individual tracks with the bus compressor active from the beginning.
Some people are eager to compress and re-compress every track on a recording. I highly recommend you not be one of them. While applying bus compression to drum submixes can really add to your sound, this technique isn’t appropriate for every mix. A jazz drum recording, for example, probably wouldn’t benefit from a squashed sounding “rock” drums on a laid-back jazz drum track where the player is using brushes and playing very dynamically. Mixing approaches are all relative, and though you don’t always need to apply processing sparingly, you do need to apply it appropriately.