We aim for more punch, depth and clarity in a quirky urban track.
The mix files for the song 'Daydreamin' came to me as a massive project, with no fewer than 39 separate vocal parts (some with both 'A' and 'B' versions), all of which were to be used in my final mix. Along with the instruments and layered percussion, the original project came to around 90 tracks, but I managed to whittle that down to a more manageable 45 or so before I started work on the mix in earnest — and even then the Logic mixer took up two whole 20-inch monitor screen-widths!
I like this song — it's slightly quirky, it feels contemporary, and it has a very well‑conceived structure. Placing it in a single category is quite difficult, though, as it features urban-style vocals underpinned with ad-libs and multi-layered harmonies. All this sits over a strong backbeat, with the most prominent instrument being a ukelele! There's also an acoustic guitar rhythm part, an acoustic guitar lead part, organ and piano segments, a string-like synth pad and some nice synth arpeggios. Add to that a multi-layered percussion part, a section of steel drums, a kick drum and a fat but somewhat 'lacking in top' bass part, and you start to get the picture.
With the exception of some slight tuning issues on the acoustic rhythm‑guitar part, it was all well played, well recorded and well sung. Equally importantly, the arrangement worked pretty much as it was.
I didn't want to try to mix all 90-something tracks in one go: 'multing' is all very well where it's required to do a job, but mixing is not about keeping all options permanently open. It never hurts to commit to some decisions, so after determining that the vocal pitching was perfectly adequate without resorting to artificial pitch correction (or perhaps it had already been subtly tweaked?), I bounced down all the double-tracked and triple-tracked parts to fresh stereo tracks after panning the individual lines to give a bit of stereo spread. There were also vocal segments recorded separately, but that could be moved onto the same track, which also helped to simplify the mix.
When it came to instruments, the ukelele was split into verse tracks and chorus tracks, and I simply combined these. All the audio tracks placed the segments at their correct time positions, other than the ukelele chorus part, which I had to copy and move so that it could be used in both choruses.
To aid navigation, I coloured the various parts according to whether they were vocals, rhythm section or whatever and routed the various vocal parts to their own stereo subgroup. A second subgroup was set up for the main drum and percussion parts. I'd also been supplied with dry and processed versions of the instrument tracks, so the processed versions were all muted and then hidden to free up more arrange-page real estate.
To my ears, the original mix had the vocals pushed a little too far back, while the somewhat boomy kick-drum tended to dominate proceedings. There was no real definition to the bass part, and I felt that some of the effects were clouding the mix and robbing it of clarity, but even so, the overall mix wasn't a bad start.
However, before I started to rebuild the mix, a little housekeeping was in order. Having bounced the vocal tracks that I could, I set about dividing the vocals into separate phrases, so that any spaces between phrases could be silenced. There were quite a few lip smacks and anticipatory breathing noises between phrases, so these were all removed in the process. While you can use a 'Strip Silence' function to do this (most DAWs now include one), I feel much more in control of proceedings when doing this sort of thing manually.
In my experience, vocal tracks often include a fair bit of low end, due to breath getting through the pop filter, so although there may be no audible popping, putting a spectrum analyser on the track will often show up a significant amount of 20-40 Hz frequencies flapping about at the low end. To combat this, I routinely put a 12dB/octave, 80Hz filter on each track and then fine-tune the cutoff frequency by ear, sometimes taking it up to as high as 200Hz if it doesn't affect the vocal tone. This goes before any compression, as the unwanted low end can otherwise trigger unnecessary gain reduction. I used Logic's own EQ for this purpose, and also took the opportunity to add a little shelving air EQ above 8kHz to open out the vocals slightly.
In addition to low-cut EQ and 'air' lift, compressors were put on all the vocal tracks to even things up a touch. Logic's own compressors were used for most of the supporting tracks, and a UAD 1176 was used in Nuke mode (all buttons in) to firm up the urban vocal part. Most of the other compressors were set to work quite gently, levelling the peaks by no more than 4dB or so. The UAD 1176 was preceded by a Waves Vocal Rider to help keep the level of the urban vocal part even and followed by a limiter just to catch any errant peaks, though with the final mix settings the limiter proved unnecessary. Similarly, a safety limiter placed over the drum subgroup turned out not to be needed, as the peaks never triggered it.
At this early stage in building the mix, I set up an initial vocal reverb using the Universal Audio EMT 250 plug-in (for their UAD DSP platform) to create a short, fairly bright reverb with 60ms of pre-delay, adding a sense of spatial presence to the vocals without actually making them sound processed. (The same principles would apply if using another reverb plug-in, but I'm particularly fond of the sound of this one.) More processing and tweaking would come later, but this gave me a good starting point. In fact, the only vocal part that was given more blatant processing turned out to be the one that carried the 'Dreaming' line on the choruses, which was treated to a generous helping of 625ms delay with 30 percent feedback, via a Logic Tape Delay plug-in that I'd set up on aux send 2.
After the vocals, the percussion and ukulele formed what I felt to be one of the most important parts of the track. The ukulele was brightened up slightly using the Noveltech Character plug-in, running on my TC PowerCore system, and again I rolled out any unnecessary low end, this time below 270Hz, but otherwise left it dry so that it would cut through cleanly. (There are plenty of enhancer plug-ins that run natively and can add some brightness, but none that seem to do quite the same thing as Character. The nearest equivalents are probably Flux's Bittersweet and the BBE Sonic Maximizer in the Sonic Sweet bundle.)
Most of the rhythm-section parts worked well with little or no added treatment, although the kick needed a bit of work to bring it into focus without losing weight. This involved using the SPL Transient designer to shorten its release time, after applying EQ to tame the sub-50Hz frequencies and to plump up the 76Hz region. A little 200Hz cut prevented the bass boost from making the mid-range flabby, while a touch of boost at 3.5kHz added some clarity. I also placed a limiter at the end of the chain, just to catch the occasional high peaks and maintain some headroom. Noveltech's Character was again put to work, this time on a multi-layered drum fill close to the end of the track, to give it a little more definition.
The snare was livened up with a coarse early-reflections treatment from Logic's Platinumverb with 16ms of pre-delay, preceded by Logic's Overdrive plug-in set to full brightness, to introduce some snap. A mild dose of distortion was also added to the handclap sound, and this was followed by a couple of short delay repeats from Logic's basic stereo delay plug-in at 70ms and 90ms, with very little feedback. By now, the rhythm section was starting to come together, but I still craved more punch, so I set up a parallel compressor fed from aux 3 and inserted the UAD plug-in version of the Empirical Labs EL7 Fatso Jnr, set to compress really hard using its 'Spank' setting, and driving it to achieve up to 25dB of gain reduction. When mixed back in with the main drum mix, this made the kick sound noticeably more solid and punchy.
I was a little worried that the bass sound was going to get lost, as it had very little definition, so I added a modest amount of overdrive (this was getting to be a habit!) to stimulate a little harmonic growth, and then used EQ to try to push the sound a little in the 170Hz region while trimming away excessive low end using a 24dB/octave low-cut filter set just below 80Hz. A 12dB/octave high cut applied above 5kHz took care of any HF mess. This improved the sound a little, and now it actually sat quite well in the rhythm section without muddying the kick sound. Feeding some of the bass part into the parallel compressor along with the drums also added some welcome weight.
Playing the vocals, the ukelele and the rhythm section showed that the track was starting to take shape, although much balancing and some automation would ultimately be necessary to keep the vocal balance correct. The band had very strong ideas as to how the vocal balance should sound, so after improving the clarity of the vocals and bringing them up in level, I sent the band several mixes so that we could fine-tune the vocal balance to their liking.
Now it was time to bring in the other instruments, which I treated far less than on their original mix. The piano was given just the barest hint of reverb and some presence boost, while the UAD's Roland Dimension D plug-in lent it just a hint of movement. The Dimension D provides a strange and rather unique chorus effect, that's much more subtle-sounding than most. (If you don't have a UAD card, you could also try equivalent plug-ins by ERS and WOK.) The string-like pad was left completely unprocessed, while only a tempo-locked 312ms tape delay was added to the arpeggio synth. A further chopped synth part was left as it was, other than rolling off the lows, while the organ was treated to EQ to add 3.5kHz boost and to roll out those unnecessary lows.
As the acoustic rhythm guitar was not entirely in tune, and because it largely underpinned what the ukelele was already playing, I trimmed off some low end and then used it quite low in the mix, where it added a bit of body to the sound but wasn't loud enough to make the tuning problem obvious. In a perfect world, I'd have asked for this to be played again, but in this instance it wasn't a major part of the arrangement, so it could be mixed quite far back. The acoustic guitar solo, which is in tune, plays towards the end of the song, accompanied by more vocals, so its level was automated to keep it audible while avoiding conflict with the vocal parts.
Some 'broad strokes' vocal level automation was required to maintain the balance between the various sections, but nothing really forensic was needed. The balace was fine-tuned after I had sent rough mixes to the band to check, which is just as well, as they managed to identify a couple of short vocal snippets they wanted to retain that I'd faded down in my initial mixes.
Panning was used to spread out the vocal layers and, to a lesser extent, the instruments, trying to maintain a nominally even left/right balance, as always, but as many of the multi-tracked parts had already been submixed in stereo, (and many of the effects were already stereo) it wasn't necessary to use extensive panning to paint a wide picture.
Overall mix treatment was provided by an SSL bus compressor emulation, and a very mild application of passive EQ from PSP's NobleQ, just to lift the 80Hz region a hint and the highs by a couple of decibels. Then I put the mix through the a tape emulator (UAD Ampex), prior to limiting, as this seemed to knit the sound together nicely, and it also made the highs sound less aggressive without detracting from the overall clarity. Limiting was restricted to around 3dB of gain reduction on the loudest peaks, while the bus compressor added no more than a couple of decibels or so of overall 'squash', so the final mix processing was actually quite gentle.
Comparing my version of the mix with the band's original rough mix, the vocals were standing out more clearly and the bass end sounded tighter, but I'd still managed to convey plenty of punch and depth. The separation between the various instruments in the mix had also been retained, which was as much down to the good musical arrangement as it was to my mixing. I was very pleased with the ukelele sound and, unlikely as it may seem, I thought it was exactly the right choice of instrument for this track!
What can you learn from this mix? Well, if I had to summarise the main elements of my approach to this type of mix, it would be to first simplify the track layout as much as possible and to use colours to identify the various types of track, just to aid navigation. It makes it so much quicker and easier to get on with the mix. For similar reasons, groups were used, where possible, to provide overall control over the main parts of the mix, such as the vocals and the rhythm section. Then I'd go on to trim away unwanted lows from any tracks that weren't a legitimate part of the rhythm section's bass end, to edit out any sections that might include low-level noise, such as breathing or ambient background sounds, and to use any effects sparingly, especially reverb.
Parallel compression is a valuable tool for adding weight without smearing the transient detail of a mix and is invaluable in urban, dance and rock production. It certainly suited this mix. Reverbs often work best when understated — I used just enough here to stop the performances feeling unnaturally dry: a fairly long pre-delay helped to create a confident vocal sound without smearing it all over the place. Where something more noticeable is required, as with some of the chorus vocals in this track, a basic repeating delay often does the trick.
It's also interesting to note that overdrive distortion isn't just for 'dirtying' things: it can occasionally be useful to make things actually sound cleaner, which is the opposite of its usual effect on electric guitar. Sources like snare drums and handclaps can often be sharpened up nicely using a simple overdrive plug-in and, as in this case, a harmonically-challenged bass can be given a bit more mid-range grunt by adding a touch of dirt. Then there are the old fallbacks of using coarse, early‑reflection reverbs to liven up percussive sounds without filling up all the valuable spaces and of adding closely spaced, discrete repeats to claps to make them sound bigger.
Ultimately, then, there were no unusual techniques applied to this mix. It was really just a matter of deciding what belonged at the front, what should stay at the back and how the lows should sound. As I said at the outset, having a well-thought-out musical arrangement really does help make life easier for the guy doing the mix!
This article won't make much sense if you don't listen to the original and remixed versions of the track! We've placed both of those, along with some other before-and-after examples, on the SOS web site.
Ivy League Sound are a production team based in Pittsburgh and comprising Mike Burger and Joseph Faulk. Mike studied piano performance at the Eastman School Of Music in New York. Joseph began playing guitar at a young age and was part of various bands writing and recording music throughout high school and college. The two began writing for artists in the Pittsburgh area in 2010, which led them to the artist Nesia Beatz. At that time, Nesia had been successfully producing songs in Atlanta and Pittsburgh for artists such as Wiz Khalifa, Yung Joc, Bonecrusher and Waka Flocka, but was interested in extending his talents as a vocal artist. Nesia came to Ivy League Sound in search of a more organic song with live instrumentation and an island/hip-hop vibe. Mike and Joseph created the entire instrumental track for 'Daydreamin' and handed the song over to Nesia to write the lyrics and melody. They recorded the vocals, ukulele and acoustic guitar using a Peluso 2247 SE mic, Vintech x73i preamp and EQ, Empirical Labs Distressor dynamics processor, and Lavry AD10 A-D converter. The rest of the song was constructed using a combination of Logic's soft synths and samplers, Lennar Digital Sylenth and Spectrasonics Trillian. They also used Yamaha HS80M studio monitors.
Joseph Faulk: "Mike and I have been reading SOS ever since we decided to start producing and recording artists, and it has been extremely helpful for us in so many ways. We are always devouring the Inside Track, gear review and Mix Rescue articles. After hearing a lot of the mix rescue 'before and after' MP3s, we were confident that our song would be in great hands. We had a few mix engineers eager to get their hands on this song, but with all that Sound On Sound has done for us, it seemed fitting to send it there.”
"Paul's mix added a lot of clarity and presence to the numerous elements of this song, helping it realise its true potential. He was extremely easy to work with and really understood our vision for this song.”
Mike Burger: "We were most concerned with the sound of the drums and the ukulele, since those elements formed the backbone of the track. Paul's adjustments, especially to the kick drum, took those elements above and beyond. We also had a lot of ideas both instrumentally and vocally as to how we were going to make each section of the song new and interesting. While good from an arrangement point of view, all of the ideas created a bit of a cloudy rough mix. Paul was able to clean up the tracks in such a way that all the parts can be heard clearly in their own space. Paul's mix is everything we hoped it would be!”
Nesia Beatz: "The mix is very organic and rich. There was so much to play around with but Paul put everything in a certain place. Sounds like candy: I love it!”
Ivy League Sound: