Tips & Techniques:Strings: Reducing surface noise when recording live strings
Strings in the Studio – Reducing Surface Noise when Recording Real Strings…..
I am sometimes asked by clients on a lower budget (or with a smaller studio) whether it is possible to achieve a realistic string section sound by skillfully overdubbing a small group of players. After quite a bit of trial and error, I’ve concluded that the secret is to emulate the mic placements of a real string section in an orchestral recording. Whenever I have played in studio orchestras, a single stereo mic has been placed in front of each section - so one for the first violins on the left, another for the second violins left of centre, a third for the violas right of center, one for the cellos on the right and finally one for the double basses at the end. If the sections are particularly large then an additional second mic may be placed halfway down each section, with a couple of ambient mics suspended above the whole orchestra.
So, the sound is taken from an audience perspective and the complexity of the string sound is captured by recording from the various distances involved. This means that some players are naturally nearer to the mics than others.
A few colleagues and I experimented with overdubbing strings using a close mic placement on every take. The result was that the surface noise captured when the bow was drawn along the string (a natural sound on any bowed instrument) became multiplied on each take, as were sounds of the players breathing (because the violins and violas in particular are played right near a person’s face!). So, after say ten violin passes, we would end up with ten times the amount of surface noise, resulting in a ‘fizzy’ end sound. This ‘fizzy’, slightly phased effect seemed even more pronounced when recording cellos and double basses.
To overcome this, we tried using a single stereo mic (suspended from a distance of about four feet from the instrument) with an ambient mic at the back of the room. Let’s assume we were using five players all in all (one for each of the five sections in a string orchestra) and that each section was to be recorded separately (so only one player recording at a time). In the first two takes, the player emulated the front ‘desk’ of the section by assuming a position slightly to the right of the mic on the first layer and slightly to the left on the second (although it doesn’t matter what order you do this in). For layers numbers three and four, the player then sat around three feet further back from the same microphones, and again recorded a layer on the right then left. This now imitated the second ‘desk’ of violins, with a static mic placement. The process is repeated for how many ‘desks’ (or pairs) of violins you require. If you do this for 8 first violins and then the same for 7 second violins, 6 violas, 5 cellos and 2 double basses you will end up with a 28-strong string section, with a highly realistic end result. The engineer can then pan the tracks from left to right to accurately reproduce the seating arrangement of a real string section.
If more than one player is overlayering at one time, then separation can be used to give the engineer/producer greater control of the end result when it comes to post production such as panning or any other sound manipulation in the mixing stage. Alternatively, you could have all five players recording simultaneously, without separation, allowing the spill to occur freely between sections. This will result in a more natural end sound, as it emulates the spill that occurs in the recording of a ‘live’ string orchestra.
There is a note of caution here. This method can only work if you use exceptional players, capable of consistently tight rhythm and intonation over several identical overlayers – in other words, you don’t just need the best professional string players, but people who are conscientious about the quality of each take. Musicians ideally need to subtly vary their tone and vibrato from one take to another to mimic the different playing personalities that naturally occur within a string group. The phrasing (crescendos and diminuendos) also has to be consistent over all of the takes, so it is a painstaking process which demands great accuracy. Any tiny errors in tuning, timing or phrasing can become magnified as further passes are recorded, so the skill of the individual musicians and careful ear of the engineer are paramount.
Considering many rock, pop and particularly dance tracks will frequently require just upper strings (violins and violas), it is possible for this to be done using between one and three players. Very few tracks will require the double bass as this territory is often already covered by the bass guitar. When using this process with the right musicians, it will become apparent that the resulting sound can still sound more realistic than the current sampled string libraries, with the added advantage that scoring more creative articulations and prominent solos are then possible (which can sound rather unconvincing when sampled).
An engineer we once recorded with related that he played back one of our string section tracks in front of another client who was using his studio. This studio had a very small live room that could only really hold about 5 musicians. The client (who was a seasoned user of programmed strings) expressed surprise and asked which sampled orchestral program he had used as the strings sounded so real….the engineer replied….it’s because they are real.
Vaughan Jones, String Section
|Comments for: Strings: Reducing surface noise when recording live strings||Page Tools||Search this Page|
|Search this Page|