Not an issue just a helping hand request

[StuartIanNaylor]

I am struggling to get my head round the required input numpy shape and the pos shape for a 2 mic linear (96mm spacing) Ignore the below pos as was one of many. Is it because the rec.shape (3200, 2) is wrong?

Apols all but any help would be appreciated.

from arlpy import bf
import numpy as np
import sounddevice as sd

# narrowband (1 kHz) timeseries array data assumed to be loaded in x
# sensor positions assumed to be in pos

fs = 16000
rec_duration = 0.20
duration = 10
sample_rate = 16000
num_channels = 2

def sd_callback(rec, frames, time, status):
  pos = [0.0, 0.0, 0.0, 0.0, 0.096, 0.0]
  y = bf.bartlett(rec, 1000, bf.steering_plane_wave(pos, 343, np.linspace(-np.pi/2, np.pi/2, 181)))
  print(y)

# Start streaming from microphone
with sd.InputStream(channels=num_channels,
                    samplerate=fs,
                    blocksize=int(fs * rec_duration),
                    callback=sd_callback):
    sd.sleep(int(duration * 1000))

[mchitre]

What's the error you're seeing?

I'm not sure I understand the pos. Why are 5 sensors at the same location 0.0? I'll take your advice and ignore the pos, but perhaps you didn't mean to have 5 sensors? Your recording is 2-channel, based on your rec.shape, but your pos suggests 5 sensors.

[StuartIanNaylor]

def sd_callback(rec, frames, time, status):
  print(rec.shape)
  rec = rec.reshape(2, 3200)

  pos = [-0.05, 0.05]
  y = bf.bartlett(rec, 1000, bf.steering_plane_wave(pos, 343, np.linspace(-np.pi/2, np.pi/2, 181)))
  print(y)

I am not sure why as should of realised as soundevice seems to be backwards with [audio, channels] so added a reshape and hey presto. Apols as before it was just a quick copy+paste after trying various options.

PS like your photography as the focus on shape, form & colour is actually really good.