Flirt With Danger - Production techniques

This is all based around using Logic Pro X but can be easily adapted for any other DAW or even using analogue outboard gear.


On any tracks of real instruments or vocals, pull up the editor and put it into file mode. Then use search peak to find any bits where it distorts digitally (which causes a very unpleasant sound) and use Change Gain to knock it down. If it's a little in the danger area, use, say, -5dB. If it's way out of hand, use, say, -20dB.

On each track, add a Gain and Level Meter after all other effects, then solo the content and adjust the gain until it reads no higher than -10dB at any point. Others reckon -12dB but I found the result too quiet and needing a lot of boosting at final mix.


Remove anything under 40Hz

Anything down this low is barely audible, is usually just noise, and wastes energy by making speakers work hard. Use a high-pass EQ to severely cut it. Set to about 35Hz for the kick, 50Hz for anything else.


Initially, the recorded material will probably sound a mess as there'll be several tracks competing for each area of bandwidth, particularly, kick drum and bass. By assigning separate areas and cutting one track while boosting the other on each, separation can be achieved.

On the kick, boost in the range 65-80Hz, cut 100-180Hz. Vice versa for the bass.


Frequencies around 3KHz can cause harshness, try a small cut here.

A shelving filter about 10KHz gives air. Useful on vocals and overall mix.


Make a copy of the vocal track and loop around a particularly bad area, 2 bars max. Add an EQ around 6KHz, boosted around 20Db, low Q. Play with the frequency until the offending area's found, then sandwich with high pass and low-pass EQs, so this track will have the S's insanely boosted but little else audible.

Mute this track then on the original vocal track, add a compressor and side-chain the copy. Play around with threshhold and ratio until the S's are reined in. Don't overdo it unless you want a lisp on the vocals.

Sub mixes

Use aux channels for this rather than groups, which are a pain if you want to vary individual channels. Usually, 10 = drums, 11 = bass, 12 = guitars, 13 = keyboards and 14 = vox.


Generally try to make this 'light touch'. Usually used on bass (which is DI'ed), guitars and vocals. The opto compressor sounds great on bass, usually use Studio VCA (simulates Focusrite Red?) on most other things.

Additionally use the Classic VCA across the drums submix just to beef them up a little.


Only use Tape Delay to give that 'vintage' sound. On most places used, set the Saturation to max.


As there limited apparent depths to what the brain can process, use just two reverb models on aux channels. 'Bigger bathroom' for vocals, 'Ambience' for anything else that needs it. Avoid on kick and bass, which adds mud to the mix, however on a low tom it works fine.

Additionally as my studio amp, a Rift 5C1, has no reverb, use a short spring reverb for the guitar channels.


Don't pan anything until after separation with EQ. In fact, leave it until one of the last tasks.

Output channel chain

First a linear phase EQ for final touches to the whole mix. Generally, a harsh cut of anything below 40Hz and a little 'air' above 10KHz.

Second, a Classic VCA compressor, very light touch, for 'mix glue'.

Third, tape emulator, on a fairly light setting.

Next, stereo spread, mid and high frequencies only, in order to keep the bass and kick centred.

Finally, adaptive limiter. Output level set to -0.1Db, other controls set to cause just a small amount of hitting the max, while still retting some movement in the levels.