Modular Synthesis (Part 1)
Producing String Sounds
A new in depth look at building up complex patches
Over the last year or so in Advanced Music Synthesis, we have been looking at the various individual modules that can be found on a synthesiser and seeing the sort of things they are capable of and I've been giving you simple exercises to illustrate each modules facilities. Over the next few months we'll be putting all these modules together and seeing how they interact with each other and instead of trying out fairly nebulous ideas we'll be covering specific sounds such as orchestral, percussive and synthesizer sounds. But before all you Musicians Union members go wild about us showing how to imitate acoustic instruments let us explain our reasons for doing this.
Acoustic instruments are often quite complex sounds that have a good deal of harmonic and amplitude variation within the course of each note and across their musical range. Synthesisers, on the other hand, have a habit of being quite static and 'even' which is why a lot of people think that synthesisers are cold and clinical. It's not so much the fault of the instrument as of the players. With a bit of thought and application, however, it is almost possible to reproduce these 'acoustic' nuances on a synthesiser which will enable you to create more 'animated' sounds and hence dispel these unfounded criticisms of electronic music. By trying to reproduce as closely as possible an acoustic instrument it will give you an idea of how familiar sounds are made up so when you set up an electronic synthesiser sound, the principles you learnt when imitating an acoustic instrument can be applied to create more interesting synthesiser effects. This principle for 'learning through copying' is not new, however. Painters when embarking on their studies, are usually told to copy famous paintings by various artists so that the techniques the artist acquires when copying these paintings can be applied to his or her own original work. It is interesting to note that the best synthesiser players usually have their roots in classical music or instrumentation. Their awareness of orchestral techniques plays an important part in the creation of their own music which usually sounds more dynamic than many synthesisers whose music and sounds tend to drone on and on, sometimes for the entire side of an album, without variation.
Anyway, to business. This month we are going to look at 'Strings'. This is a good all purpose sound that can be used as a melody line or as chordal backing but with some thought it can be made far more interesting than the standard string-synth sound that has been flogged to death over the last couple of years. But before we see how to set up a string sound, let's take a quick look at how 'real' strings achieve their unique sound.
First of all, a stringed instrument such as the violin has a high harmonic content and because of the bowing action, the envelope has a slowish attack, usually a full sustain level (when played legato) and little or no release. A string section has a number of players, all of them scraping away at their instruments and because of human inaccuracy, none of these instruments will be exactly in tune with each other and also, each musician will have a slightly different style and vibrato technique so that the end result will be a rich, thick, ensemble sound. Synthesising, this, however, can be done in a variety of ways.
The ingredients you need for this 'recipe' are 1 voltage controlled oscillator (VCO) giving simultaneous sawtooth and pulse waveforms, 1 voltage controlled filter (VCF), 1 voltage controlled amplifier (VCA) and envelope generator (EG) combination, 1 low frequency oscillator (LFO) and a chorus unit. These will need to be patched up as in Fig. 1. The control settings aren't that critical: but the VCO should be set to give both the sawtooth and pulse waveforms simultaneously, the cut-off frequency and keyboard follow on the VCF should be set at about two-thirds or three-quarters, depending on the tone of strings you require. The LFO should be set to give a medium-slow output (around 5Hz) and patched to give vibrato and pulse-width-modulation (PWM).
I find it best to keep the vibrato on all the time and not delayed in any way, incidentally. The EG is connected only to the VCA and is set to give a slowish attack (the exact setting depends on the type of string sound you're after), full sustain and a fairly short release (long release times tend to transform your string sound into a 'theatre organ' and will blur melody lines). Messing around with ADSR controls will yield quite a large variety of string effects, however, and I advise you to experiment for the best effect - as long as the attack time is fairly slow you should get a passable string sound. Finally, all these are shoved through a chorus unit to further augment the ensemble effect provided by the vibrato and PWM.
This utilises the same sort of control settings as Method 1 except that we are using 2 VCOs to provide the ensemble effect. This method is preferable to the above as chorus units can sound somewhat electronic and can have an unnerving 'beat' to them which shows up on long sustained notes, while the effect of two detuned VCOs can sound more natural. Again we use the sawtooth and pulse waves simultaneously but if your particular synthesizer doesn't allow you to do this then select sawtooth on both VCOs or sawtooth on one and pulse on the other - two pulse waves will not sound so good as two sawtooth waves. The setting of detune depends on the thickness of the ensemble effect you require but I don't recommend tuning the VCOs an octave apart (or using a sub-octave) as this will sound more 'synthesiserish' and less authentic. To further augment the ensemble effect, you could use a chorus unit in this method as well but this can sometimes blur the sound if you are not careful.
This requires more hardware than either of the two methods outlined above. You will need a modular synthesiser that has 3 (or more) VCOs, 2 VCFs, 2 EG/VCAs, 2 (or more) LFOs and probably a small mixer of some form. Set all the controls as above but make small variations to each 'channel' so that the cutoff freq of one VCF is slightly different to the other and likewise with the attack, LFO rate, etc. This will allow you to create a more authentic string sound as it enables you to recreate the small differences that exist between two instruments and two players. For the more adventurous amongst you, if you have the hardware, you can also recreate the striking of the bow against the strings by feeding a sinewave into the control voltage inputs of one or both VCOs. The amount of modulation from this sinewave is fairly critical but you should be able to tell when the effect is right. The frequency of the sine wave should be around 80-100 Hz (experiment for the best effect) and it must be shaped via another EG/VCA giving a fairly fast attack decay with no sustain. This will add a 'crunchy' transient to the front end of string sound and is particularly effective on faster melodic lines. Routing the VCA via a foot-pedal to control the level of the sinewave will allow you to bring in the 'scrape' as and when you want it. It takes a bit of time to set up but the results can be very good indeed, especially if the synthesiser is then put under the control of a sequencer or Micro-Composer.
Those then are the three ways of obtaining good string sounds that will sound quite authentic. There are a few things one has to consider, however, in order to get a totally realistic string section effect. Firstly, violins, violas, 'cellos and contrabass are not simply the same sound spaced an octave or so apart. The heavier strings of the lower pitched instruments require more energy to get a sound. As a result, the attack and vibrato speed will be slower than on a higher pitched stringed instrument such as the violin. To more accurately reproduce this effect will necessitate you altering the respective controls on your synthesizer accordingly. On a polysynth, this can be difficult, especially if you intend performing the whole string part in one go unless you have a key-follow control on the EGs that allows you to vary the envelopes times over the range of the keyboard. Another point to bear in mind is that in a string section there are fewer 'cellos and contrabasses than there are violins and violas so the ensemble effect will need to be decreased in the lower registers by decreasing the amount of detune, chorus or PWM. To overcome these problems, I advise you to record each line separately - but more of that later.
One of the key points in obtaining a realistic string sound is not only the sound but the way that sound is played. In other words, don't expect to get authentic string sounds by playing large, keyboard type chords on the keyboard - you will have to play the parts very much as a string section would play them. Fig. 4. illustrates this. The block chording given in Fig. 4a will not sound as good as the arrangement in Fig. 4b. Block chords can sound quite good but for the most part I recommend you orchestrate your string parts as if it were a 'real' string section. Again, this will require you to record each part separately as the intervals can be quite large but the results can be well worth the effort, believe me.
Your playing technique can also play an important part in the final string sound as well. The slow attack on the EG enables you to obtain a touch-sensitive effect whereby playing in a more staccato manner will give - a soft, fairly quiet sound as the attack is not being allowed to reach to full sustain level. Keeping your fingers on the keyboard for just a fraction longer will allow more of the attack portion to be heard and the sound will be a bit louder so that varying the time you keep your fingers on the keyboard can give you notes of varying levels. To hear this demonstrated to good effect, I suggest you take a listen to 'Village Ghetto Land' by Stevie Wonder off the 'Songs in the Key of Life' album set. Although done on the huge Yamaha GX1, the same sort of effect is possible on even the smallest of synthesizers.
Finally, there are some things to bear in mind when recording string parts on a synthesizer. For a start, a 'real' string section does not play in mono, but is spread across the stereo image and so, once again, it is a good idea to record each string part individually. Not only will you be able to articulate each instrument separately and make allowances for the differences between the different pitch ranges, but you will be able to pan them to give a wider string sound. If tracks are limited on your tape machine then I suggest you record the violins and violas on one track and the 'cellos and contrabasses on another and during mixdown, send the violins/violas through a chorus unit (via the effects send on the mixer) and pan the straight and chorused signals left and right respectively leaving the cello/bass lines in the centre. This will help to spread the string sound quite considerably. If you have access to a harmoniser then that would be better than a chorus unit. If you have quite a few tracks to play with, try double or triple-tracking string parts in unison using detuned or varispeed effects.
Reverberation (NOT echo) will give you the correct acoustic environment for the strings to work in (Echo will give 'bouncing' repeats, especially on more staccato lines - this can be a good special effect, however). I also recommend that you tweak the EQ in the mid-range a bit in order to give that woody resonance of strings. A simple EQ section on a mixer will probably suffice but a graphic equaliser will give you more control. You can use one or more chorus units on your string sound and I find it best to route it via the effects send and bring the chorus back through another channel where it can be EQ'd and treated separately.
That just about concludes this months look at strings. As I said earlier, any of these techniques can be usefully employed for any sort of ensemble sound, be it and imitation of an acoustic instrument or a unique electronic sound. I have not covered pizzicato strings as this can be a bit involved and will require another 4 pages! I will, however, be looking at them later.
Next month, we'll have a look at brass sounds so until then, happy orchestrating.
Feature by Steve Howell
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!