This article is part of a series; see Speech Reinforcement. This is a stub; more content may be added later.
The room greatly affects the audio signal. This can be fixed, in part, by installing acoustic treatment. It can be accommodated, in part, by applying to the signal an inverse of the room’s frequency response. Room acoustics, modes, and responses are perhaps the most complicated part of the signal path.
Acoustics
In a reverberant room (acoustically “live”), a voice can be easily heard throughout the room, but it is difficult to understand. This is because the sound bounces around in the room, so there are many sound waves present based on the original voice, but the timing is all different and the sounds are all colored by the surfaces off which they have reflected.
In a non-reverberant room (acoustically “dead”), a voice is more difficult to hear throughout the room, but what is heard is easier to understand. In this case there are much fewer reflections, but most of what is heard has come directly from the source.
Sources:
Notes
Feedback suppression using frequency shifting: shift singing up about 2 Hz (or not at all), shift speech down 6 Hz, or shift everything up 3 Hz
reverberation/RT60: Studios designed for recording speech typically have a reverberation time of about 0.3 seconds (Source). For the human voice, the most critical frequency range with regard to reverberation is 500 Hz to 2 kHz, and carpet on the walls is good at absorbing energy in this region.
RTA
acoustic tiles: type, placement, frequency ranges
ringout correction
pink noise correction with target curves
Increasing delay
Target 65 dB SPL, reasonable range 60-70 dB SPL