Funktion-one F81 - Appendix B - Gain structure; What is gain structure all about and why is it important?

To Next Page

To Previous Page

Appendix B – Gain structure

What is gain structure all about and why is it important?

When an audio signal passes through an analogue or digital audio system, it is important to maintain the

optimum signal operating level through the various sections of the signal path. Microphone levels must

be boosted to make them compatible with line level and analogue-to-digital converter stages and mix

stages must allow enough headroom for multi-channel summation.

The various gain settings throughout a piece of audio equipment are collectively known the system’s gain

structure.

Signal level too low

If the signal level is too low it will be wallowing around in the analogue noise floor losing resolution and

becoming fuzzy, distorted and noisy in the digital domain.

Avoiding digital garbage

If you have to go digital for your multi-channel live sound operations, use 24-bit systems - if they are

available - so that you can maintain enough headroom to mix without sinking into the digital mush. And,

of course, if you like to use lots of effects, try to use one, comprehensive, studio-quality effects system

that runs its internal processing at 32-bit floating point. This reduces the build-up of mush you get every

time your signal goes through a different processor.

Signal level too high

If your operating levels are too high, there won’t be enough headroom to allow for performance peaks

and multi-channel mixing. You may end up with distortion caused by peak clipping - or even long-term

clipping.

If you’re using all-analogue equipment, the odd peak clip may go unnoticed as long as it’s not sustained

enough to damage your HF drivers. This is because analogue clipping components are harmonically

related to the original signal.

Avoiding digital mush

Unfortunately, digital overload isn’t as simple. You get clipping when the digital stage runs out of digits –

usually referred to as 0dBFS – and the higher order harmonics generated can then alias with clock

frequencies creating all sorts of strange sum and difference frequencies. These aliasing products can be

interesting if you’re a robot but are generally heard by humanoids as nasty and unmusical. They can also

turn cymbal “tings” into “shhhhs” and make vocals excessively sibilant.

To make matters worse, any over-sampling or delay-based processes built into effects units, can stretch

these nasty artefacts in the time domain and make them far more audible than their relatively low levels

would suggest. Also remember that aliasing will be compounded as the overloaded signal passes through

successive processes.

Signal level just right

If your operating signal levels are just right, you won’t run out of headroom, clip or run out of digits –

even on musical peaks. And you’ll still run well above the analogue noise floor and stay clear of the digital

mush.