Using mixbus as pre master

[pvannatto]

Feedback occurs when the phase-gain product is greater than unity.

I would disagree with that. I've run many a live events above "unity" with out any sign of feedback.

Simply if the sound from the loudspeaker is greater than the original sound

Which is always the case when using sound equipment.

it will continue to increase in level with each subsequent passage through the system

Sort of. Feedback typically occurs when a microphone picks up specific frequencies from the speaker system, which is then processed through the system, creating a signal loop. It can also happen when a bus is fed to itself, which is called internal feedback. Feedback is reduced with proper speaker and mic placement and EQ (since feedback is frequency sensitive). Compression, when used correctly has little or no effect on feedback.

Say that gain is 20dB. If you use 4dB make up gain to a compressor, the input gain must be reduced by 4dB to 16dB or the system will feedback.

I think there is a misunderstanding of what gain actually is. Gain is an sensitivity adjustment, set usually during sound check to have all incoming levels relatively the same (ballpark target is -18db on the channel meter). Once that is set, it doesn't change during the live event because if it is changed, it affects all mixes, including monitor mixes, IEMs, broadcast (livestream), etc. Also, input signals are not fixed levels. They are dynamic and can range from -90 to +10 (on an X32) during a performance.

When a compressor is properly setup, it is simply a tool that can help the tech control a signal's overall level. It is an alternate (and automatic) method instead of "riding the fader". For example, when a vocalist starts to sing, I adjust the threshold to where the compression meter bounces between -2 and -6 db. Knee is at 3 (soft compression) and attack, hold and release set to auto (yeah I'm lazy), and make up gain set to 3 to 4 db. Sometimes the level is below the threshold and the compressor doesn't kick in. Other times the level is very high (vocalist belts out a section of a song) and the compressor level will go beyond the -6 db. Because the knee is soft, the intensity will be heard without being "blow dried" by the level.

Whether you leave the input gain at 20dB and increase the threshold level to achieve the desired compression, or reduce the input gain, lower the threshold, and use make up gain, the maximum overall gain can’t exceed 20dB.

Not true, sorry

In a studio situation, which is where the OP is coming from, compressors are generally used to reduce the dynamic range so that gain can be added to bring up the quieter bits, without the louder bits being too loud. In effect, the compressor is used to set the upper level and the make up gain then brings the lower level upwards.

Exactly the reason why we use compressors in a live event situation. The task of a live sound tech is to mix so that everything can be heard, but under control. Since levels will always be dynamic, he/she can either ride the faders all night long, or use compressors to automate it (to a degree).

[STEVEB58]

I would disagree with that. I've run many a live events above "unity" with out any sign of feedback.

I’m talking about the closed loop gain of the complete system. If you take the mixer in isolation, without the rest of the system there is no closed loop. You have to take into consideration all aspects, including speaker efficiency and directivity, the spacial relationship between the loudspeaker and microphone and the room acoustics, to name a few.

For example most stage microphones are directional. If the manufacturer’s literature is to be believed, the rear rejection can be in excess of 15dB compared to on axis response. With a vocal microphone, the singer’s mouth is usually just a few cm away from the microphone, while the floor monitor can be couple of metres away; so the attenuation due to distance could be another 6dB. If the spl of the singer is say 100dB, the monitor spl could easily be 120dB without causing problems. So yes, the loudspeaker can be louder than the source, but the loop gain is still less than unity.

If

gain is an sensitivity adjustment, set usually during sound check to have all incoming levels relatively the same (ballpark target is -18db on the channel meter).

then moving the microphone shouldn’t make any difference. However, if you point an otherwise stable microphone at the monitor, there is a good chance that it will feedback. This is because you have effectively added 15dB gain to the loop; or if you are being pedantic, lost the 15dB of attenuation caused by the directivity of the microphone.

Not true, sorry

Yes true.

If you assume the rest of the system to have a fixed gain, if the system goes into self oscillation with 20dB gain, that is your limit. You can adjust the knobs and faders to have a multitude of combinations, configurations and setting, but if the overall gain exceeds 20dB the system will feedback. Please note that the 20dB figure is just an example. In reality the value could be either more or less and will vary with frequency. Using eq to eliminate feedback works because the gain at the offending frequency is reduced, even though you have not touched the mixer’s gain knob. If you use a compressor with 4dB make up gain, below threshold, the system has 4dB more gain, which is the same as using no make up gain and adding 4dB gain at the input, or indeed pushing the fader up by 4dB. The total gain is the same.