Suppose you could measure the room response, and use a "smart" speaker that included a correction for that. You send a digital signal, a chipset uses a programmable correction template and produces an output that would sound OK for that room. People will measure your room response for you. Invert that distribution (with cutoffs) and download it into a smart speaker, and you should get clearer messages. The only thing missing is the smart speaker technology--doable, but is it doable cheaply, and is it durable/robust over time and voltage spikes?
Friday, March 22, 2024
Muddy speakers
I have heard muddy PA announcements too often. Sometimes it has been a lousy speaker, but often these days it is the room response doing strange things with the frequency mix. High frequencies may get absorbed, low ones echo, and the distortion makes it hard to distinguish words. The room response varies from place to place, and on how many people are in the room, which complicates matters, but for the moment just consider the average.
Subscribe to:
Post Comments (Atom)
1 comment:
Many who do good PA (either system design and install or operation/mixing at a venue) use a software package https://en.wikipedia.org/wiki/Smaart which allows you to place one mic at a reference location or many mics to get an average of many listening locations, and then uses the speech or music going through the PA as the test signal so that you can see if the frequency balance at those mic positions is correct.
More often then one would think, the informed user of this software can see that the bad frequency response where the listeners are is a result of errors that can't be fixed by complementary changes at the input end.
It pains me that there is much good engineering knowledge on what is needed for good sound system intelligibility in real rooms, yet so many PA systems seem to be specified and used with little regard to what is required in the space.
Every large enclosed space has a "critical distance", beyond which the reverberant energy resulting from the sound being emitted into the room will be louder than the direct 1st arrival. The engineering response is either distributing loudspeakers so that everyone gets enough direct sound, or loudspeakers with increased 'directivity' (narrower beam°). One speaker manufacturer for big arena concert sound even had models named "floodlight" and "flashlight" to highlight the differences in directivity. I'm not sure even with today's DSP power that we could de-reverberate at a source in a way that would work for those beyond the critical distance.
Of course using speakers with narrow dispersion almost invites having people outside the "coverage angle" of the speakers, in which case they will get very little high frequency stuff in their direct sound and much more bass (big dimensions or phased-array trickery being necessary to control the directivity of sound at low frequencies -- the wavelength of 100hz being on the order of 3.5 meters) and very little highs. And if several speakers with narrow coverage are combined to widen coverage, then small physical misalignment between them can cause tremendous frequency-response notches from comb-filtering in the area where 2 or more speakers overlap.
When someone is using a microphone in the same room as the speakers it feeds, it starts to get complicated. . .
Post a Comment