This post explains a technique I've been experimenting with for avoiding 'clutter' in my mixes.
About a month back there was a discussion thread concerning the tendency for mixes to sound 'cluttered' and the various ways of avoiding this situation. The solution proposed by Stephen Dunston and others involved ensuring that you don't have too many sounds with similar frequency ranges layered on top of each other. In cases where you do have such a situation the suggested solution was to use subtractive EQ - in other words to cut unnecessary frequencies rather than boosting the fundamentals.
If you play acoustic instruments or electric guitar you'll be able to find information on the fundamental frequency and ranges of your instruments and you'll be able to use that to help you set the EQ to cut the frequency range and increase the 'space' in your mixes while retaining the essential character of the instruments.
Unfortunately things aren't so easy if you play synthesizers - the characteristics of the sounds can vary wildly from patch to patch, and analogue synthesizers (OK, my JP-8000 isn't really analogue, but you'd never know to hear it) are infamous for creating clutter and poorly defined sounds when you start to pile on the tracks. The kind of nice fat sound you naturally aim for when programming new sounds can swamp a mix. However there's a well-known solution to this problem when you're using any synth that has a resonant filter.
You can use a synth's filter to determine the resonant frequency of a sound and then to restrict its range. The technique is to get rid of any filter modulation, set the resonance to a point just short of self-oscillation then sweep the cutoff to find the point where you get maximum output. This is the resonant frequency of the sound that determines it's basic character. You then use the filter to restrict the range of the sound by adjusting the resonance at this point until the sound retains enough of its character but is thinner than the original - in a similar manner to using subtractive EQ.
You can use the same technique to determine the resonant frequencies of your sounds using the VS (if you have OS v2), although I wouldn't suggest you try to solve problems in this way - use the EQ instead.
The technique is as follows:
Set up a track on the VS to take input from the instrument whose sound you want to 'analyse' and insert one of the effects processors into that track. Put the display into post-fade mode so you can see what the effect does to the sound. Set up the effect as follows:
Algorithm B92=Lo-Fi Proces LFP:LoFiPros = off RMF:RMF = on RMF:Type = BPF RMF:CutOff = 0 RMF:Reso = 98 RMF:Gain = 24dB NS:Thresh = 0 NS:Release = 10
Now play and sustain or repeat a single note in the range the sound will actually be playing in. I often use the synth's arpeggiator to keep it repeating the same note. If you're using an acoustic / electric instrument try recording a single note and using the scrub function to set up a loop that'll keep the sound playing. Go to the Cutoff parameter of the effect and sweep the frequency of the filter through its entire range - you'll hear peaks in the volume of the resulting sound, and see them on the display. The loudest peaks correspond to the resonant frequencies of the sound; the range over which lesser peaks are produced gives you the frequency range of the sound. Keep a note of where the peaks occur.
Repeat this process with the other sounds you want to use in the mix. If the results show your sounds all have their resonant frequencies at the same points then those sounds are going to lose definition when you layer them. If you find the sounds produce a wide range of resonant frequencies (they have a broad spread of many lower volume peaks) then you may need to EQ them (or edit the original sounds) to cut down their range to avoid clutter: this'll probably make the sounds thin when played solo, but they'll fit into the mix much better as a result.
The snag with all this is that the Lo-Fi Processor effect's filter cutoff is just a number in the range 0-100 so it's not much help when trying to work out how to set the EQ. I haven't tried to convert the numbers to frequencies yet, but this shouldn't be too hard since it's possible to make the filter resonate.
I've also tried this technique using the four band parametric EQ, but that wasn't so successful since you can't get a big enough boost to make the resonant effect clearly audible. Things are also made more difficult by having the four bands which overlap, and the fine resolution of the frequency control means you end up doing a lot of dial-twirling: mine's got a circular hole worn in it now....
My problems with cluttered mixes became severe when I attempted to produce some fairly complex pieces of music (typically 12 to 16 tracks) using only my JP-8000. My first attempts were very disappointing: strong bass sounds became muffled, rich pads lost their depth and lead parts tended to disappear into the mix. I couldn't understand this since the tracks taken singly showed the VS's sound quality to be impeccable. The posts to this list concerning how to avoid clutter in your mixes showed me what I was doing wrong, and the technique I've described here has helped me improve my final mixes tremendously.
If you come up with any variations or improvements on this I'd like to hear about them!
Alan Clifton
31 Jan 1998
1998/03