Best mcontroller for fast audio acquisition/processing/rendering

Hello,
I'm looking for advice on the most suitable microcontroller for an audio processing application ingesting multiple audio channels (4) and generating sound where minimal latency is crucial, say 50 micro-seconds and a sample rate above 30KHz.
i was assuming that embedded ADC/DAC is preferable.
my first experiments with R4 Uno and ESP32 make me think that it will be too slow (even using direct register handling).
is that doable? should i go for fpga or other platforms ?
many thanks
Louis

Somehow you forgot to describe exactly what you want to do.

So, no buffering, how about toggling a digital pin like how a class D amplifier works?

What I want to achieve is a device that helps record/listen birdsong using beamforming and audio source separation techniques.
short latency is not an absolute requirement but I would like also to explore earing aid scenario

hmmm, the point is that the output signal results from some calculation based on input channels ( eg delay, sum, multiply)

Your "explaniation" is just taking around the subject. Why don't you just describe what you want to do?

right, here is an example of the process i have in mind :

setup
     for channel in [1,2,3,4] : 
         buffer[channel] = Buffer(size = 20 * channel) // creating fifo

void loop() : // expecting that it would take less than 50 microsec
   for channel in [1,2,3,4] : 
       buffer[channel].append(ADC(channel))
   int s = 0
   for channel in [0,1,2,3] : 
      s += buffer[channel].pop()
   output(DAC(s))

Get a Raspberrypi and some USB microphones.