The audio files available on this page accompany the Mix Rescue column for SOS October 2012 (/sos/oct12/articles/mix-rescue-1012.htm), featuring the song 'On The Line' by SOS reader James May. The filenames should be fairly self-explanatory, but the descriptions below should help you understand a little more about what you're hearing. In addition to these demonstrations, you can also download both the raw multitrack files and my full Cockos Reaper remix project from the 'Mixing Secrets' Free Multitrack Download Library at www.cambridge-mt.com/ms-mtk.htm#JamesMay.
Here's a section of the raw acoustic-guitar recording that James originally submitted as part of his multitracks for Mix Rescue. He was already determined to retrack it, because he wasn't happy with the part he'd played, but I was also keen for him to do this to get a more satisfying raw sound, preferably with some inherent stereo width — after all, this part was meant to be providing the harmonic bedrock of the whole song.
The tracks that James recorded to replace what you heard in the AcGtr01_Original file comprised a straight DI track (which I didn't use), a 'pseudo-miked' stereo track he'd generated from the DI signal using convolution technology, and a crossed coincident stereo microphone array. This file shows what my mix of these sounded like early on in the balancing process — the only processing at this point is 4dB of Sides attentuation of the convolved track and a couple of decibels of 100Hz shelving boost from Sonimus SonEQ. The two signals were blended pretty equally, with perhaps 1-2dB more of the combined signal coming from the mics than from the convolution signal.
If you listen carefully to the AcGtr02_RetrackBalanceEQ file, you can hear how the low-frequency thudding of James's hand against the guitar as he damps the strings is rather overwhelming, and the pick noise is also in general a little edgy. In this audio example you can hear the results of my processing solutions to these problems: -7dB of Attack gain from SPL's Transient Designer plug-in on both the miked and convolved-signal channels, followed by fast compression of the frequency extremes from Universal Audio's Precision Multiband dynamics plug-in. The latter compressed the regions under 300Hz and over 2kHz, both with a 3.5:1 ratio and fast time constants.
Parallel compression from Stillwell Audio's The Rocket compressor proved to be one of the secrets magic ingredients in the mix, and here's what the guitar sounded like after this was mixed in. Compare this with the AcGtr03_PickThudReduction file to hear the difference most clearly.
My only concern with the parallel compression effect you can hear in the AcGtr04_ParaComp audio file is that it unduly emphasises the guitar tone's low-end sustain, whereas it tends to be the upper strings that usually need more help in this regard. Therefore, for the final remix sound I applied a 3dB EQ cut at 54Hz in the parallel channel, pre-compression, to push the sustain more into the higher-frequency ranges. Note that I employed DDMF's LP10 linear-phase algorithm for this purpose, because I wanted to avoid destructive phase-cancellation of the overall guitar timbre.
The drums in this mix were emminently mixable, not least because of the lovely capture, as illustrated in this audio example, which features the unprocessed overheads recording. Notice how the snare in particular sounds great, and the whole kit (with the inevitably exception of the kick, is naturally very well well balanced too. As a result, I didn't have to work hard to rebalance the kit with the close mics, as I often seem to have to do with Mix Rescue submissions.
Here's the processed kick-drum close-mic from the final remix. In order to create slightly better attack definition, I applied an instance of Christian Budde's Lightweight limiter. In order to let the drum's attack phase pass through the limiting stage untouched, I set a 17ms attack time. A release time of 100ms allowed the gain reduction to mostly reset between hits, and also drew up a little extra of the drum's 'rumble' into the bargain.
So that you can hear the effects of the limiting applied for the KickLimiterIn file, I've bypassed it in this audio example, roughly matching the apparent loudness so that you can appreciated specifically the change in the attack sound.
Here's the snare close mic as processed for the final remix. I EQ'd it with Cockos Reaper's built-in ReaEQ plug-in to add low midrange weight (a 5dB peak at 570Hz with a 2.6-octave bandwidth, contained at the low end with a 90Hz high-pass filter), reduce vocal masking (a 3dB shelving cut at 3.5kHz), and tame a discordant pitched harmonic (a 7dB 1/10th-octave notch at 905Hz). Other than that the only processing was another instance of Christian Budde's Lightweight Limiter.
The Lightweight Limiter processing on the snare track involved a fast 1.5ms attack, deliberately to reduce the cut-through of the drum's transients, and for this example I've bypassed this plug-in to demonstrate what it was doing — compare this with the SnareLimiterIn audio file.
Here's the bass part from James's original tracking sessions for the song. Notice how the tone and low-end weight vary a great deal from note to note. Notice, for example, how the first and third notes here are much less bassy than the second and fourth, and the fourth note also suddenly has a big peak at around 150Hz. The note at 0:11 is another good example, featuring a momentary boost in energy at 800Hz and concomitant tonal nasility. It was for this reason that I decided that the bass would benefit from retracking.
Here's the raw retracked bass DI part delivered by Studio Pros to replace the original part you heard in the Bass01_Original audio file. This is altogether more even tonally, albiet with a more muted timbre overall.
This audio file shows what the processed bass sounded like until the latter stages of the remix. EQ was courtesy of Cockos Reaper's ReaEQ (40Hz high-pass filter to remove rumble as; -2.5dB at 70Hz and 150Hz to remove a some excess warmth; and 2dB boost at 1.5kHz for a bit more presence) and Stillwell Audio's Vibe EQ (another 3.5dB at 800Hz, again for increased melodic audibility). A parallel channel of Mokafix's NoAmp distortion gave a stiffer dose of tonal brightening, achieving a wirier sound closer to the original recording, and then two instances of Universal Audio's LA2A plug-ins helped control the dynamics a little, both on the DI itself and on the mixed DI+distortion channel.
Towards the end of the mixing process, I decided that the bass sound needed more low-end reach, so I applied an additional layer of processing from Cockos Reaper's bundled ReaXcomp multi-band compressor, as you can hear in this audio example. This operated below 120Hz at a ratio of 4:1, reducing the gain by 4-10dB on most notes (although this was then made up using 7dB of make-up gain in that band) with an attack time of 15ms and a release time of 116ms. Compare this with Bass03_EQDistComp to hear the effect this processing had.
Here's a section of the song's opening which most clearly showcases the effects of the automated EQ band I inserted on the acoustic guitar track late in the mixing process. The EQ in question was a 2.4-octave wide peaking cut centred at around 3.3kHz, which cut 2.5dB from the guitar's midrange whenever the vocal signal was present, thereby clarifying the vocal sound a small but significant degree. I used Cockos Reaper's built-in ReaEQ plug-in for the purpose.
For this example I've bypassed the automated EQ band on the acoustic guitar parts so that you can hear how much difference it made. Compared with the GtrAutuEQIn file in your own MIDI + Audio sequencer to hear the subtle differences most clearly.
In the main article, I discuss the subliminal effects of sudden drop-outs in the 'air' between vocal phrases, and why I tend to try to bridge them in some way, perhaps with copied background noise from another section of the recording. As you can imagine, this is a very subtle effect, and engineers disagree as to whether it's important or not, so I've created a couple of audio files to give you the opportunity to decide for yourself. This example features a gap in the vocal 'air' and a missing breath before the third phrase — the drop-out occurs at around 0:12-0:14. Now compare this with the VoxAirGapIn file to see if it makes an appreciable difference for you.
For comparison with the VoxAirGapOut audio file. This example has had the 'air' between the second and third phrases (around 0:12-0:14) reconstituted using a section of another take. For me this makes the verse as a whole flow a fraction better musically, and it feels more as if the whole verse hangs together. However, the effect is extremely subtle, so you may decide that it's all snake oil!
The original mix of 'On The Line' sent in by SOS reader James May to me for Mix Rescue.
Here's the first draft remix I sent to James, based on his original multitracks but with retracked vocal, acoustic-guitar, and bass parts. Although the arrangement and rhythmic feel are both already quite different from the original mix, James felt that it could still have more contrasting arrangement dynamics, so in the end I made some more creative changes for the final remix version.
This is the final remix version of James May's 'On The Line', complete with the additional rearrangements I carried out in response to James's first-draft feedback.