Ask the Doctors! Drs. David P. Berners and Jonathan S. Abel Answer Your Signal Processing Questions.
 |
|
& Jonathan S. Abel |
--Mike Barnes, UA Marketing Director via tap on the shoulder
A: The Fairchild was modeled using the same approach we took with our other vintage emulations: we used the schematic to construct a model of the unit’s operation, and then devised a set of tests to run on the hardware which would tell us how to choose parameter and component values, and also let us know the nature and locations of the nonlinearities within the unit.
The topology of the Fairchild is similar to the 1176, but it uses variable-mu tubes to control the gain rather than FET’s. The variable-mu tube is interesting in that the way the tube’s gain changes is very sensitive to its geometry, which in turn depends upon the details of its production. Without having production drawings for the tubes used in the Fairchild, we needed to find the gain characteristics experimentally. This made for an interesting problem, because the value of the particular Fairchild we used for our modeling prevented us from disassembling it, or pulling any of the tubes. Instead, we came up with a set of input/output tests we could use to characterize the tubes without taking the unit apart.
Because the variable-mu tubes are used as active gain elements rather than variable-impedance shunts, the Fairchild uses a transformer-coupled differencing scheme to reduce artifacts from the sidechain signal leaking into the audio path. This configuration can also lower the noise floor. However, mismatched tubes cause most Fairchilds to have a bit of “thumping” during attack, even when the bias currents are matched. Since this is a very common but unintended behavior, we decided to include a user control to select the amount of tube/bias current mismatch in our plugin.
While not on the front panel, the Fairchild has a ratio adjustment, which works in precisely the same way as it does on the 1176, by adjusting the bias point of the sidechain output. While the 1176 uses pushbuttons to select from a set of discrete resistors, the Fairchild uses a potentiometer to fix the bias point. This control also changes the width of the ‘soft knee’, with higher compression ratios having sharper knees.
The original Fairchild could be run either in dual mono mode or in ‘lateral-vertical’ mode, which was used for mastering phonograph recordings. In this mode the incoming signals are matrixed into L+R and L-R, and compressed separately before being teased apart into L and R at the output. Many Fairchilds have been modified by the addition of a linking switch, which ties the two sidechains together when enabled. This allows the compressor to be used as a true stereo compressor, with no L-R image-shifting. All of these modes have been incorporated into the plugin.
Another interesting feature of the Fairchild is the choice of attack/release settings. Some of the release settings on the Fairchild are program-dependent. Again, this dependence is achieved in precisely the same way as on the 1176. However, because of the different choices of time constants available, and because of the different gain characteristics of the compressor, the actual release behavior is very different from an 1176. Each channel of the Fairchild has six settings for attack/release, but when the stereo link switch is enabled, the controls interact, leading to about 20 unique selections for attack/release behavior.