I have long advocated that the best, nay, only way to become a better music mixer is to practice. I have also privately lamented that I don’t myself take the time to practice more – see, I’m always too busy making my own music. A worthy endeavor, indeed, but I work so slowly that there’s just not that much music to mix. Enter HOFA.
HOFA is a German company that does many things musical, including making high-quality plug-ins. Somewhere along the line I got on their mailing list, and in early December they sent out an invitation to join a mix contest. The rules were simple – download their tracks and make the best mix you can. There were two versions – one a simple mix contest where adding instruments or rearranging the song were verboten (see what I did there?) and a re-mix contest where anything goes. I listened to their version of the mix and decided that I liked the song enough and heard enough in their mix that I would do differently that I downloaded the 48 (48!) tracks and loaded them into a new Sonar project to give it a go.
The CEO of HOFA is a chap named Jochen Sachse and he has a musical project called INI. They have a number of videos on the YouTube. This song, called Warten, or Waiting, caught my ear partially because of the instrumentation – drums, bass, acoustic and electric guitar, and also piano, Hammond organ, mandolin, violin and harmonium (kind of like an accordion that sits on your lap), but mostly because I really like the song. It’s hard to place in a genre, but if I had to I would call it folk rock. Sachse’s sample mix leans more rock than folk, but my mix leans folk.
I spend a lot of time reading about mixing, listening to podcasts about mixing, and even watching videos about mixing – it was time to put some of that knowledge into use actually mixing. Based on what I had learned I had a good idea how I wanted to approach this. Step one involved bringing all the tracks into Sonar, seeing (hearing) what I had, arranging them on the screen to simplify my workflow, color coding them for convenience, and setting up the audio routing for the mix.
With all the tracks imported, I used used Sonar Remove Silence tool to break up the clips and I quickly learned that while there were five acoustic guitar tracks, none of them played at the same time. What Jochen did is break each part of the song into a separate track, anticipating that the mixer would want to apply different effects or panning to an instrument depending on the part of the song. To keep all this straight and avoid future confusion, the first thing I did was group tracks for the same instrument into a single track folder, and then routed all those tracks to a single buss track. This can all be seen in the composite screen shot near the bottom of this article.
Doing this allowed me to deal with only 7-10 channels when mixing rather than 48. I also planned to mostly use the same EQ and compressor settings on an instrument for the whole song. In the screenshot above you can see an example of having all the acoustic guitars routed to the buss “AC Guitars.” I did the same thing for the piano, electric guitar and organ. Finally, the drum and vocal tracks were all sent to their respective busses.
In addition to all the buss routing, I set up my effects sends – in this case a single ambiance reverb in the form of ValhallaDPS’s ValhallaRoom. It does the best ambiance reverb I’ve ever heard. The other initial set-up things I did were to apply the Console and Tape emulator to the Master Buss and Console emulators to the rest of the channels. Finally, I added the Density mkIII compressor, set to mid-side, and Cakewalk’s Boost11 limiter to the end of the chain. I’m starting to run into the limitations of Boost11, so I think the next time it’s on sale I will be picking up Sonar’s Concrete Limiter for the ProChannel. Oh, I also set up an instance of Native Instrument’s Guitar Rig on the bass since it had been recorded direct. First I’ve used Guitar Rig – it’s pretty sweet.
The final step once those were all dialed in was to begin looping the song and setting basic levels and pans. The idea is to get the best possible mix before adding any more processing other than the basics I’ve already mentioned. This involved soloing both the drum and vocal buss to work more carefully on the balances within those sets of tracks. Once I had that in place, then next step was to set up volume automation for all of the tracks that are getting mixed individually – the busses and several individual instruments like bass, violin and harmonium. I listened through each section of the song and tweaked the volume of the individual elements for just that section (see volume envelope on the bass track, above).
Top Down Mixing
Top down, or backward mixing is a practice of mixing that I’ve heard a lot about lately, and I’ve wanted to give it a go myself. The idea is that you listen to the whole mix pretty much all the time and do the least possible to fix what you don’t like. Therefore, you start with the Master buss and work back from there to the Aux busses and finally to the individual tracks as needed.
In this case the whole mix felt dark and thick. The answer to this was to open the Master Buss EQ and dip out the low mids a bit around 260 Hz and apply a high shelf to brighten the whole mix a little. I couldn’t take much more that a few dB out of the whole mix however, before I started losing power on some of the instruments and vocals, so I turned to the busses. I once got a tip that a cut in the 500 hz band on the drum buss is like magic. That tip was right, so that’s what I did. I also cut a little of the low mids from the electric guitar since the part it is playing is very low-mid heavy.
Satisfied that the instruments were sitting better in the mix the next step was to apply some effects and processing. I started with the guitars, which can be seen to the left. The Plektron WTComp has been my go-to guitar compressor for quite a few years, now. It really seems to do something magical with plucked strings. Also added is the NastyDLAmkII by Variety of Sound. It’s one of the more useful and user friendly delays I’ve ever used. In this case I made the mistake of not setting the project tempo prior to doing all the work loading in clips and splitting them up, and changing the tempo would have shoved them all around. Instead, I just used Nick Fever‘s on-line delay calculator to find the right values for a 16th and 8th note delay – one in the left and one in the right.
Supplied drum tracks included kick front and back, snare top and bottom, two toms, hi hat, ride and four room stereo pairs – overhead, near, far, and something called “boom” that contained a lot of lower frequencies. I muted one of the of the room pairs and mixed the other three together, using the overheads as the primary signal.
The drum buss also has some specific processing applied. Beyond the compressor, I wanted to change the stereo field throughout the song – the verses are narrower than the chorus, which helps make the chorus feel bigger and more important. Most more instruments I could just change the panning from closer in to further out, but the drum overhead and room mics were both stereo pairs panned straight up the middle. Bundled with Sonar Platinum are a number of effects from Nomad Factory, including the Blue Tube Stereo Imager. I used this on the drum buss to narrow the stereo field during the verses.
There are two violin tracks. The image to the right shows one of them – the processing between the two is very similar except that the high shelf cut in the track shown is a high shelf boost in the other track. My only complaint with the entire recording is that the violin tracks are really roomy. Since I don’t have access to any of the advanced (read: expensive) reverb removal software, I inserted the TS-64 Transient Shaper and used it to back off some of the ambience. It’s not perfect, and you have to be careful not to overdo it and begin losing signal, but in the mix it works reasonably well. While a dryer track would be nice, my biggest problem with the violin track is how it is played. The lines have no attack – each note gently swells from nothing. As a result, the line has little to no definition. The problem is that the track either melts into the mix and disappears or screeches over the top. I think I found the best balance I could with what I had.
One more treatment that I wanted to mention specifically is the harmonium. While an accordion can be a very expressive instrument in the arms of a talented player, I noticed that this harmonium is much more static. I decided to add a chorus to give it some movement. While it’s somewhat subtle in the mix, I think it’s cool in the end.
There were seven vocal tracks supplied with the mix – lead and double, lead performing ad libs, and four harmonies: one male and three female. The arrangement of the entire song was very good – so good, in fact, that the only part of everything I downloaded that I didn’t use was a bit of the doubled lead vocals (greyed out in the screen shot below). The ad libs provided me with a bit of a conundrum, though. I originally had them mixed way back, with some additional plate reverb (picked up VahallaDSP’s brand new Valhalla Plate just for this, although I love it and can’t wait to use it some more!)
The thing is that in the final verse the ad libs sounded more like they should be a lead vocal doing an improv over the top of the rest of the mix. The problem was that when I mixed them loud enough it a) sounded unnatural and b) sounded too roomy. At first I thought the roominess was the same problem I was having with the violins and tried to fix it with a transient shaper. No matter how much I fiddled, though, I couldn’t find the balance between dry and gone. Silly me, it turned out to be the ambience reverb causing all the problems…
So I removed the ab lib track from the ambience send (by pulling them separate from the vocal buss). Viola, dry vocals! If you look closely at the screen shot above you can see what else I did at the last verse – took the lead and double out of the center of the mix and panned them hard left and right, and dropped the level of the lead to equal the double. I turned them into part of the choir allowing the ad libs to soar over the top in the center of the mix.
While we’re here, a couple other things to observe from the screen shot: 1. Volume automation to the lead vocal helping to improve intelligibility (I laugh I little when I listen, while I don’t speak a word of German I’m amazed by how precisely everything is enunciated – must be a cultural thing). 2. Volume and plate reverb automation on the ad libs track. There was one phrase right before where the ad libs take over that I spiked the reverb a bit. 3. The automation lanes on the vocal buss at the bottom – the ambience send slowly grows as the song goes on – increasing the send also serves to increase the overall volume of the vocals a little bit, too. The bottommost automation lane is the mix for the delay – a very similar delay setup to the guitars. Note that it, too, increases as the song goes on and the mix gets denser. You can never really hear the delay in the mix (although you’ll sure notice it if I turn it off) but every time I hit stop in the middle of playback the delay is really obvious.
Automation; automation everywhere. The screenshot, left, shows the automation for the remaining busses that I haven’t really covered so far. The two curves at the end of the song are the fade-outs on the organ and final piano chord.
Not shown are my treatment of the jingle bells, which got compression and a high shelf to enhance the good; and eq and compression on the piano get it to fit in the mix and sparkle a bit.
Finally, at the bottom of the screen shot is the automation on the master buss. The bottom is simply the limiter output at the end of the song, helping to smooth the fade-out. Just above that, however, is the limiter input. I adjusted it by a total of 2 dB over the course of the song – quieter in the verse and louder in the chorus. I ended up lowering the entire input level by about 2 dB when I finished because the mix was beginning to sound a bit squashed – I especially noticed in the cymbals. Lowering the input level just that much made a huge difference – the whole mix breathed better.
I leave you with a composite of the tracking screen showing all the tracks and automation lanes. Enjoy, and please feel free to leave comments below.