Mixing the Drums

On this song I was presented with some unusual and creative drum tracks to mix. As Rob explained, the drum kit was made to sound very trashy. No mics were used on toms, as they were not played. Instead we had mics on the kick, snare, hats, and overheads. We also had a very crappy sounding mic under the snare, and another ugly sounding mic in the room.

Presented with these choices, I decided to take a reasonably straightforward mix on the kit using the 5 main mics on the kit, and leave the ugly snare and rooms mics out of the drum submix. That way, I could use those colorful mics later in the mix.

The drummer had used an undamped kick drum with quarters taped to the head and beater. The mic was placed very close to the head, giving the drum a strange, empty, clicky sound. I rather liked it as it was - raw, boomy, round, and thumpy - but it had too much "roar". That roar can step all over the mix - and with 48 tracks, I was going to have to get everything to sit nicely in its place. So I did not EQ it, but instead used a Waves C1 gate to clamp down on the roar, and an Ultrafunk compressor with a fast attack and high ratio (8:1) to really make the attack pop out. Listen to the unprocessed and processed sounds and you can hear the difference. I think because the mic was so close to the drum head, it picked up a lot more snare than I would have liked, and the snare shows through in the processed kick track. But, remember, the main point was to gate out the roar.

Then there was the snare. The top mic had a very conventional snare sound, a little too much ring for my tastes, but I never would have guessed that there were 1500 BBs in it! I use a Waves Q10 to EQ the ring out:

As you can see I performed notch cuts at 120 Hz, 482 Hz, and 595 Hz. To find the drum ring, listen to the drum, and hear its primary ring tone. Get that tone in your head as you take a band of EQ, set the Q as high as it will go (100) and boost all the way up. Now sweep the frequency up and down until it matches the tone in your head. When you find it, it will ring REALLY LOUD! Now cut it, all the way if you have to. Once this frequency is removed, you can hear the next overtone. Repeat the process until the nastiest overtones are gone. The less EQ you can get away with, the better.

As you can hear from the dry snare track, the snare was kind of thick, so I did a severe bass cut as well as a high shelf to shift its tone up and bring out the "crack". I also used a C1 gate to damp the ring and roar and a Renaissance Compressor to accentuate the attack. (For more information on using a compressor to bring out an instrument's attack, read the "Squish This!" series of articles here on ProRec.com). After the EQ, gating, and compression, the processed snare track was bright and had lots of crisp attack.

For hi-hats, the drummer had used old metal turntable platters. The dry sound was brash and mechanical, and there was plenty of it already in the overhead mics. So I used the hi-hat mic as a "hi-hat treble boost control" by adding a radical EQ to it. I used a Waves Q4 EQ with a low cut set at 1500 Hz and a high shelf boost of +6 @ 6000 Hz to essentially remove all midrange and hype the bright treble. I would use this processed hi-hat track sparingly, just to add brilliance to the hats without making the rest of the kit too bright.

The dry overheads really got a lot of the tom roar. I felt the overheads must have been too close to the kit to have this much tom ring in them. Whatever the cause, the overheads just had this thick midrange roar that needed to be minimized. I used a radical EQ on them to lighten them up - a low shelf set to -8 dB @ 650 Hz combined with a +4 high shelf @ 8000 Hz gave the kit a severe treble tilt, and a deep dip @ 3000 Hz helped minimize the brashness of the "oven rack" crashes and "turntable platter" hi-hats. The processed track was a lot emptier sounding. And it would need to be: if the drums had a wall-to-wall sound, then there wouldn't be any room for the other 43 tracks! Also, I was planning on using the "ugly drum room mic" heavily, and it was very thick and midrangey.

Once I had the drum sound tuned in the way I wanted it, I bounced the drums over to two tracks and muted the individual tracks. I knew that I was going to be using a lot of plugins on this mix and wanted to keep the CPU as free as possible to try new stuff. You can hear the completed two-track drum submix here (without the "ugly snare" and "ugly room" mics).

Coming into the final chorus you will notice a reverse drum loop. That loop provides a somewhat psychedelic feel to the end of the song, as well as adding another texture. The only effect used on the reverse drum loop was a MondoMod used to spread the otherwise mono image out into the speakers. The drummer also cut a track with pots and pans placed on his toms and snare. This track was left unprocessed and placed back into the mix. It makes an appearance during the final chorus.

Filling out the Bass

Presented with this rude, clanging drum sound (which I loved), I was underwhelmed by the dry, clean bass track that was provided in the mix. On the verses, the bass really carries the rhythm, providing the phrasing that is the verse's hook. I wanted something mean to be happening there. And, on the double-choruses, there is a killer bass figure that really rocks. But with the round, plain sound that was recorded, that part was just lost. I wanted growl. Lots of it.

To achieve this, I used a process similar to the approach used to record Chris Squire's famous bass lead on "Roundabout." The idea is to split the bass signal into two paths. One path is essentially full-range, direct, and a little compressed. This provides the bottom end to the signal. The other path is filtered to remove all of the bass content and run through a loud, distorted amp. When the two sounds are mixed together, it almost sounds as though the bass is being doubled on a distorted baritone guitar.

To do this in Vegas, I used a Cakewalk Amp Sim inserted into an effects bus. I also inserted a Waves Q4 into that bus before the Amp Sim. The Q4 was set up with a bass cut @ 70 Hz and a low shelf set at -15 dB @ 250 Hz. The low shelf de-emphasized the midbass while the low cut removed the low bass. It is essential to remove the bass content before hitting the amp, or else the amp will just fart out sloppy bass notes. By providing the cuts, the amp "sees" a lot more midrange, and it is this signal, not the bass, which generates the distortion.

On the bass track, I used a Waves Q4 to cut lows @ 30 Hz to remove some speaker-flapping stuff that was in the track. I used a Waves Renaissance compressor set to 4:1 ratio, fast attack and release and about 6 dB gain reduction to "ride" the signal, evening out the bass notes a little bit. I then followed the compressor with a plugin that I rarely ever use - a Hyperprism Bass Maximizer. This is a sub-bass generator that adds in low harmonics (as opposed to the Waves MaxxBass which works the other way, adding in high harmonics). I wanted this effect because I wanted the bass to retain its beefiness even when the highest notes were hit. Otherwise, the bass content seemed to vanish.

Now I had the bass sound that I wanted, but there was a problem. There was a slight amount of hum in the track, and the Amp Sim made it very obvious. And this wasn't just a 60 Hz hum - it was a square wave, with even-order harmonics at 120 Hz, 240 Hz, 480 Hz, all the way up to the highest treble frequencies. You can't EQ this noise out. So I took the track into Cool Edit 2000 and used CE's noise reduction to eliminate the hum. CE uses a sampling noise reduction that removes only the specific frequencies sampled. Sonic Foundry also provides a similar noise reduction, however, I am always able to get better noise reduction from the Cool Edit system. Don't know why.After the track was cleaned up, it sounded great (or as Rob described it, "really pissed off.").

Article Continues >>>