This came up on social media recently, so I thought I would share it again here. The question was –
How does the audio limiter work when using AGC to control the audio level?
The limiter is in fact bypassed when the audio gain is set to AGC, so it doesn’t actually do anything when you are using AGC and can be ignored.
So – what AGC settings should I use?
So – what AGC settings should I use?
The AGC will set the overall gain profile depending on you AGC spec point. It will reduce gain when audio is present and then allow levels to creep up in extended quiet periods but not all the way to your chosen AGC spec point. The AGC spec point determines the average peak level the AGC will try to achieve for a moderately large input level (the input gain is determined by the Audio Input Reference Level for Ch1/2).
A higher Spec Point level (smaller number, eg -6db) increases the target peak level, so provides a wider recording range. During very sudden input peaks if the level reaches your chosen AGC spec point very aggressive gain reduction occurs, compressing the entire audio signal leading to undesirable pumping effects or audio distortion but typically avoiding hard clipping.
Using a higher Spec point reduces the chance of normal audio peaks triggering this very aggressive gain reduction due to the larger available range, but if you do hit the spec point the gain reduction is more aggressive and the resulting audio may sound even more distorted.
The Limiter:
The Limiter:
The limiter which only functions in manual mode kicks in when the audio hits the level you have set. The aggressiveness of the limiter depends on the set level. When using a lower level (larger number eg -17db) there is more range between the onset of compression and the clip point so it can be less aggressive but more of your normal lower peaks will be compressed (depending on the gain you are using). The limiter only acts on the louder part of the input rather than the overall gain so pumping etc is much less obvious than when using AGC.
If you like what you read here, please also visit and subscribe to my YouTube Channel by clicking here. Thanks.
Will using 32 bit float in the future solve the need for limiters
But 24 bit has more DR (144dB) than any microphone (absolute best will be under 130dB), most typical mics are less than 110dB. The issue is the preamp and ADC noise which also affects 32 bit float and 32 bit float has no more accuracy than 24 so the quantisation errors are similar with both which reduces the practical usable DR of 32 bit float to something not all that much greater than 24 bit. With a 32 bit system it is still possible to overload the input ADC, but you rarely hear this because of the microphones limited DR. The 24 bit issue is that most people are setting the levels far too high, based on old analog levels where recording noise was an issue. There is no additional recording noise with digital so the input gain can be set lower and you won’t clip, the microphone will saturate before the recording clips.
Would you have a recommended level for the limiter , for interviews . -12 db ? Thanks
Rather than using a low limiter I prefer to bring the input gain down so you don’t run into the limiter with my normal audio peaks at around -24dB rather than the normal EBU reference of -18dB as this doubles the headroom. I might leave the limiter at -6dB but generally my mics are going to overload before hitting 0dB.
Thanks sir