When we rail on the importance of “gain structure,” what we are really aiming for is less noise or “hiss” and a louder signal. In other words, we are trying to optimize signal-to-noise ratio that is typically expressed in decibels. Having discussed some basics on gain structure previously in this column, I want to recap a bit and move forward.
I believe we have nearly beaten to death the gain structure issue on the drive side of the mixing console a few years ago. But achieving good gain structure is mostly found from the input source through the mixing console. So for the rest of this column, let us focus on the diagram in Figure 1 and look at ways to preserve the signal to noise ratio.
From the Beginning
To start with our sources, whether a microphone or a direct box, these sources generally will have the best possible signal-to-noise ratio (SNR), and from here it is just a case of how many “dBs” of dynamic range and signal-to-noise ratio we choose to lose by raising the noise floor. If anything, it is a good object lesson to choose the best mics you can use in performance, being aware that some mics will never be rugged enough to leave the recording studio. Given good electrical signals from the chosen mic or DI box, reasonable cabling allows those really good signals to flow to the jacks of the mixing consoles (monitor and front of house).
If you have been digesting the mic preamp specifications of various consoles, you will stumble across a few useful parameters. The most important one is the input referred noise floor of the console. Typically, the best consoles advertise a spec close to a -130 dBu noise floor. This pretty much sets the initial dynamic range, from the -130 dBu input floor to the +22 dBu maximum unclipped signal that the power supplies within the console will allow you to use. So with that nearly 150 dB of theoretical SNR, the first thing most consoles will do is gain up the input signals to nominal levels, such as to 0 dBu average signal level.
Most dynamic microphones will deliver about a -50 dBu signal at moderate intensities. And condenser mics will offer a few dB hotter signals than dynamic mics. But when you crank up the gain on the signal, you also ratchet up the console noise floor dB for dB. Now suddenly that 150 dB dynamic range and 130 dB SNR drop to 100 dB and 80 dB, respectively, due to the increased gain. The good news is that the mic preamps are optimized for low-noise performance, and that means you will only lose a couple more decibels off the SNR due to passing through this gain stage.
A dB here, A dB There
This brings us to a basic rule of gain structure: Every gain stage taxes some decibels of the SNR and there is no way to undo this burden. A corollary to this is that you may wish to consider that the fewer gain stages passed through, the less additional noise added. Each amplifier/gain stage has a Noise Figure represented in decibels. As the signal passes through each stage, the SNR shrinks by the Noise Figure. Now there is nothing to cry about 80 to 100 dB of SNR, especially in comparison to the usual 60 dB SNR for vinyl record playback on audiophile grade stereos, 45 dB SNR for cassette playback and around 70 to 80 dB SNR for typical CD/MP3 playback. The challenge is to minimize losses in the rest of the console.
Attenuation is the biggest thief of gain structure, but judicious use of attenuation is what allows us to “mix” properly. The Noise Figure of an attenuating circuit (potentiometer, VCA, pan or fader) is the same as the number of dBs being reduced by the attenuator. In other words, the signal drops by a certain number of decibels, but the noise floor does not always drop and may remain about the same. So, every attempt to recover signal from prior attenuators not only amplifies a reduced SNR signal, but the amplifier’s Noise Figure is now reducing the signal a bit more.
Examples
An example of bad gain structure would be a console in which the operator really gained up the mic preamp to +10 dBu averages, then ran the smoking hot signal into an insert compressor in which the compressor’s input gain control had to attenuate the signal down. Then to get back up to the operator’s level of +10 dBu hotness, the insert compressor’s make-up gain is then cranked up and returned to the console. Even if the equalizer section were left flat or bypassed, there is a good chance that the signal source has to be attenuated back to sanity as the master bus summing amps may have been clipping if all the channels are hot and summed. And if the drive section needs a lot of signal for the speaker processor and amp settings, then the master fader gains are back up. All this up/down and more up/down, gain and attenuation is a recipe for a rising noise floor.
An example of a near perfect gain structure is the cool operator who dials up his/her mic preamp for 0 dBu average signal levels, cleanly passes the line levels to the insert signal processing without much level shifting, lightly fades down the appropriate signals for the mix, and runs the master faders as the ultimate “gas pedal” for venue loudness. These fader masters may be at 0 dB, but just as likely be at -5 dBu or -10 dBu if the venue requires less sound reinforcement. Setting the channel preamp gains around 0 dBu provides a lot of clean headroom upwards for accidental/intentional performance crescendos, while still allowing the noise floor to stay plenty below the audible range (unless you like sticking your ears on speaker grilles for every last dB of dynamic range).
Temptations
Now after saying all of this, you may be tempted to “mix” using the mic preamp gains, and leaving all the channel and master faders at 0 dB or unity gain. The only issue with that philosophy is that the consoles are not designed for that purpose and the auxiliary send levels would be all over the place and need re-adjustment each time you regain the channels. My answer is to not mix from the mic preamps, and use the faders for what they are designed for. Mixing from the mic preamps makes every aux send a “post” send. Yes, you could say that using the faders to attenuate the less prominent signals invites noise floor creep, but these signals can get away with less SNR since they are placed below the main sources in the mix.
E-mail Mark at marka@fohonline.com
