UA WebZine "Ask the Doctors!" December 07 | The UAD Precision Buss Compressor

The VCA's (Voltage-Controlled Amplifier) gain is determined by the voltage supplied to a control input. Virtually every compressor technically uses a VCA, whether it is based on FETs, variable-mu tubes, or optical elements. However, in the parlance of audio, the term “VCA compressor” usually refers to a compressor using an integrated-circuit VCA for gain control. Integrated VCAs have the desirable properties of low distortion, high input impedance, low output impedance and high bandwidth. In addition, integrated VCAs provide precise relationships between control voltage and amplifier gain. Because of these features, integrated VCAs can be used to modularize the design of compressors or other gain control devices: High input impedance allows signal detection to be performed without worrying about loading the detector circuit. Low distortion and high bandwidth have obvious advantages. Most importantly, the tightly prescribed relationship between control voltage and amplifier gain allows for conscious design of compression curve shapes. By contrast, with most vintage discrete compressor designs, the form of the compression curves depends mostly on the vagaries of the devices employed to provide gain reduction. VCAs usually have an exponential dependence between control voltage and gain, so that the sensitivity of the amplifier with respect to the control signal is specified in dB/V. This type of VCA is said to have “logarithmic gain control.” The logarithmic-gain-control VCA is ideal for compression, because any constant compression ratio can be achieved by using the scaled, log-encoded signal level as a control signal for the VCA.

The tightly prescribed relationship between control voltage and amplifier gain allows for conscious design of compression curve shapes.

Part of the characteristic sound of a good VCA compressor is due to the low distortion exhibited by the VCA; most other gain-reduction technologies cause varying degrees of distortion, depending on the amount of compression. But many of the common properties of popular VCA compressors are due more to the time of their design than to the technology used. Compressor-design philosophies have changed over the decades according to applications, music styles and the tastes of designers. By the time integrated VCAs were available for practical use, compressors were being designed for different purposes than early “vintage” compressors. Attack times were being made with much bigger ranges on the “slow” end by companies like SSL, and auto-release functions were being designed somewhat differently than they had been previously. Compressors like the Neve 88R had separate detectors for dropouts or silence that would affect release behavior. The flexibility of the VCA would have made it possible to design VCA-based compressors that were somewhat similar to FET, variable-mu, or even optical compressors. But by the time VCA compressors became common, the philosophy of compression had gone in a different direction.

Hardware compressors are typically designed using feedback topologies, while expanders and gates more often use feed-forward control signals. Part of the reason for this is that with compression ratios greater than one, the dynamic range of the output signal is smaller than the dynamic range of the input signal. Design of the level detector is thus simplified by working with the reduced dynamic range of the output signal. (For compression ratios less than one, the opposite is true; the input signal has the smaller dynamic range.) Feedback systems pose a special problem for digital implementations. If a feedback system has a response time comparable to or shorter than one sampling period, special measures have to be taken to do an accurate implementation. For the case of the VCA compressor, typical attack times can approach one sampling period at a sampling rate of 44 kHz. Direct implementation of such a system as a digital algorithm would result in gross inaccuracy, or even instability. The most robust way to deal with the problem is to create an equivalent feed-forward system, if possible. For compressors with first-order signal detection (no program dependence) and constant ratio, such a feed-forward system is easily derived. But with program dependence, or a “soft” compression knee, a more complex transformation must be found if the system is to be implemented as feed-forward. This transformation is the key to the operation of the Precision Buss Compressor. Once a suitable model has been found, the details of different compression curves or program-dependent release functions can be fit to the model as ancillary parameters.

The Precision Buss Compressor is meant to typify the VCA buss compressor in terms of compression characteristics, attack, and release. As such, the behavior of this compressor is vastly different from previous UA software compression offerings, and the Precision Buss Compressor should be a nice complement to the existing line. A sidechain filter has been included to increase flexibility, as well as an auto-fade function. These two features are also typical of the VCA compressor era.