Case Studies in DIY Recording
By Jason Lustig
Case Study # 5
Well, this column was supposed to be monthly, but you know…life. Sorry about that. I had a fun project recently that presented one of the most fundamental challenges in mixing. It was a hip-hop project where the tracks I received to mix consisted of (for most songs) a mono drum track, a mono instrument track (made of one or more samples/loops), a lead vocal, and some backing vocals (doubles/ad-libs). So, to mix in mono or not? If simulating stereo from the mono tracks, do you worry about mono-compatibility?
As with all recording/mixing/mastering, please start by asking yourself the question: “What is the purpose of this project?” Put another way: “What is the artist trying to emotionally communicate?” Remember, all your gear are just tools, and you are an instrument of the emotional need of the music. Your job is to help the artist effect the listener as strongly as possible in the way that the artist intends. So if mixing in mono is the most effective way to communicate the intent to the listener, then don’t hesitate. However, with this project, I know that the artist was going to be played on Pandora and wanted to sonically compete with the bigger names. I felt that a mono mix would be a dead giveaway to their budget level, particularly with the prevelance of headphones/earbuds as the primary listening method these days. Headphones/earbuds will always make a stereo mix sound wider and a mono mix could feel that much smaller by comparison.
So, now that I decided I was going to mix the tracks to create some sort of stereo feel from the mono tracks, I had to decide how important to make their mono-compatibility. Given that these were going to be exposed to lossy compression coding (MP3, streaming formats, etc…) I felt that mono compatibility was going to be important. Also, and this comes from my more formal training, I just find that preserving mono compatibility whenever possible is good form. Additionally, this group would use the instrumental mixes for their live shows and my guess was that the majority of house PAs would be running in mono, probably summing the outputs of a stereo or iPod to mono, therefore mono compatibility would save me from having to do separate mixes for their live shows (definitely not in the budget).
Okay, fake stereo – check; mono-compatibility – check. On to the technique I used. This is definitely not the only way to accomplish the above two objectives, but it’s a quick, easy, and effective method that I’ve used to great success on several projects over the years.
Starting with the mono drum track. On most songs, this consisted of a kick, snare, hi-hat and usually not much else. Sometimes there were additional drums contained in the loop on the instrument track as well. So for the drum track, I created two identical copies to give myself three tracks of drums, one panned left, one right, and one center. If these were played back as is, they would just be the mono track but louder, so here’s how I made them stereo while still holding down the middle.
1) The center track was compressed leaving the transients largely in tact and with a full low end and meaty attack on the snare. This was done with some eq and some low end enhancement (I often use Waves RBass but a BBE or just eq will do). The compression was set up to let the initial transients through while bulking up the sound. I feel in this fake stereo technique it is important to keep your center channel drums with strong transients to root the listener’s ear to center.
2) The right and left tracks of the drums were compressed and eq’d differently from each other and differently from the center track. I used two different compressors on these with drastically different settings so that they would impart their own unique sonic characteristics on each side. Perhaps a SSL bus compressor on one and something like the default Nuendo compressor (like a DBX 166xl) on the other. Both of them were run aggressively to squash transients, but by choosing two very different compressors each side pumped and breathed at different rates and with different qualities of distortion. From there each side was EQ’d a little differently, perhaps the right side was a tad brighter to move the hi-hat to that side (audience perspective – sorry drummers). I also usually rolled off or dipped the low eq on the sides to leave the kick’s low end firmly centered in the mix.
So with that, we have a transient, full sounding center channel, a left channel which has been “destroyed” with compression on the left, and another similarly “destroyed” but with a different sounding compressor on the right. What started as three identical mono tracks is now much wider thanks to the differing compression artifacts and EQ. Play around with the ratio of the “side” channels to the center channel to suit your tastes (I’m often about 2:1 center to side). (Please read the note below on latency and delay compensation. Whenever doing this type of “parallel” compression it is a good idea to check your system and know if you need to do any workarounds to prevent comb filtering which can occur.)
Going further, you can set up two or more reverbs, each panned hard to one side or the other, and then feed the left drum channel to the left verb and the right channel to the right verb (with different sounding rooms naturally) to further create artificial space.
Next up is the mono instrument/sample/loop track. I was planning on keeping the lead vocals centered and with the drums having a strong center channel in their new form, I don’t want my music crowding that out. So for the music track, I create just a single copy. I pan one left and one right, no center version.
From here, I usually choose two different compressors, maybe from an LA2A, SSL bus comp, Fairchild 660, even an 1176, Waves RVox (best bang for the buck) or RComp just to name a few. Try and choose two that have sounds you like but are quite different. Play around with different combinations of them, one left and a different one on the right. Play around with driving them harder or softer. For the most part, with the types of tracks I got, I did a medium level of compression, nothing too drastic, but enough to bring out some details and even out some discrepancies. For these two, I would subtly eq each side differently and feed each side into a different reverb, each panned hard left or right. Very similar to the drum sound but no center channel and not pushing the compression as hard on the sides.
Bringing these up with the instruments and then using the mono button on the master section of my console I could flip back and forth between the fake stereo and the mono sum checking for two things: 1) mono-compatibility (are there weird artifacts or losses in frequency response – AKA comb filtering) and 2) Was all that work an improvement over the mono sum. As always, I do these comparisons with my eyes closed and I rapidly hit the button over and over until I can’t remember which I’m on, then I use my ears to pick which one I like best. Thankfully, I liked my fake stereo better, it create some noticeable depth and spread without sounding like I was just throwing things around (I guess I could have panned the drums to one side and the instruments to the other and called it a day – but this was meant to be modern Hip-Hop not a 3 position stereo recording like Brubeck’s “Take 5” or something).
I added in the vocals, did some panning with the adlibs and create a mix that wasn’t quite as complex as a mix where I had access to each track individually or stereo stems, but wasn’t bad considering all I had to start with was a mono drum track and a mono loop track.
A note on latency/delay compensation – The one thing I haven’t talked about is the implications in computer based (in the box) mixing when duplicating tracks but utilizing different plugins. In the analog world you can basically ignore the processing latency of the outboard gear you use, so this won’t apply if you use these techniques with analog gear, but in the computer, each plug-in you use takes a different amount of time to process and spit out the changed sound. All modern DAW software has plug-in automatic delay compensation which SHOULD account for the different plug-ins in use throughout the project and adjust the output timings to counter the varying latencies/delays created by the processing. I say SHOULD because in a multitude of tests I’ve done, I’ve found that this is almost never the case in actual practice.
Here’s how to test your system’s delay compensation. Pick a relatively full spectrum sound (distorted guitar is often a good one). Duplicate the track. Put two different plug-ins on the track, make certain they are ON but not actively changing the sound (for instance, compression ratios of 1:1 or EQ flat) then pan each track center and play back at the same time. If you reduce each tracks volume by 3db when you combine it, the combined sound will the same volume as the original track. When you combine them do you hear any changes to the frequency response compared to either track alone or compared to the original track with no plug-ins? If delay compensation is working then it shouldn’t make any difference. However, if each plug-in is spitting the sound out at a slightly different time and the DAW is not compensating, the waves will not line up exactly and the result will be comb filtering and usually a loss of low end with a very “Real Audio from 1998” watery like sound on the top end.
Think your computer doesn’t compensate. Here’s what you do and how to check that it is working. Let’s say you have a guitar and you want to process it with parallel compression (2 copies of the track with 2 different compressors but blended together). With delay compensation working this is no problem, but to do it without delay compensation, or just to be on the safe side, make your copy of the track, put your first plug in on track 1, go to track 2, load that same first plugin but make certain that it isn’t doing anything (again like a compression ratio of 1:1 or flat eq…), then load your active 2nd plug-in on track 2. Go back to track 1 and add that same second plug-in to track 1 making certain that it too is not doing anything. This way you have the same two plug-ins on both tracks so the output delay will be the same but only one plug-in is affecting the sound on each track. If you sum to mono or pan to the same side you shouldn’t hear any of the combfiltering. If you remove one of the dummy plug-ins you should hear the sound change – that’s when you know that your system doesn’t do delay compensation well and you need to use this work around.
So there you have it, faking a stereo mix from two premixed mono tracks with good mono-compatibility and adjusting for poor plug-in delay compensation (a must with this type of fake stereo). Thanks for reading.
Learn more at http://jasonlustigrecording.com
Copyright 2012 Jason Lustig