FreeDATA/test/001_highsnr_stdio_audio/test_rx.py

168 lines
5.9 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Wed Dec 23 07:04:24 2020
@author: DJ2LS
"""
import ctypes
from ctypes import *
import pathlib
import pyaudio
import sys
import logging
import time
import threading
import sys
import argparse
import numpy as np
sys.path.insert(0,'..')
import codec2
#--------------------------------------------GET PARAMETER INPUTS
parser = argparse.ArgumentParser(description='Simons TEST TNC')
parser.add_argument('--bursts', dest="N_BURSTS", default=1, type=int)
parser.add_argument('--framesperburst', dest="N_FRAMES_PER_BURST", default=1, type=int)
parser.add_argument('--mode', dest="FREEDV_MODE", type=str, choices=['datac0', 'datac1', 'datac3'])
2021-12-14 22:56:55 +00:00
parser.add_argument('--audiodev', dest="AUDIO_INPUT_DEVICE", default=-1, type=int,
help="audio device number to use, use -2 to automatically select a loopback device")
parser.add_argument('--debug', dest="DEBUGGING_MODE", action="store_true")
2021-12-14 22:07:01 +00:00
parser.add_argument('--timeout', dest="TIMEOUT", default=10, type=int, help="Timeout (seconds) before test ends")
parser.add_argument('--list', dest="LIST", action="store_true", help="list audio devices by number and exit")
args = parser.parse_args()
if args.LIST:
p = pyaudio.PyAudio()
for dev in range(0,p.get_device_count()):
print("audiodev: ", dev, p.get_device_info_by_index(dev)["name"])
quit()
N_BURSTS = args.N_BURSTS
N_FRAMES_PER_BURST = args.N_FRAMES_PER_BURST
AUDIO_INPUT_DEVICE = args.AUDIO_INPUT_DEVICE
MODE = codec2.FREEDV_MODE[args.FREEDV_MODE].value
DEBUGGING_MODE = args.DEBUGGING_MODE
2021-12-14 22:07:01 +00:00
TIMEOUT = args.TIMEOUT
2021-12-12 12:54:23 +00:00
# AUDIO PARAMETERS
AUDIO_FRAMES_PER_BUFFER = 2400 # seems to be best if >=1024
MODEM_SAMPLE_RATE = codec2.api.FREEDV_FS_8000
2021-12-16 01:26:39 +00:00
AUDIO_SAMPLE_RATE_RX = 48000
# make sure our resampler will work
assert (AUDIO_SAMPLE_RATE_RX / MODEM_SAMPLE_RATE) == codec2.api.FDMDV_OS_48
2021-12-12 12:54:23 +00:00
# check if we want to use an audio device then do an pyaudio init
if AUDIO_INPUT_DEVICE != -1:
p = pyaudio.PyAudio()
2021-12-14 22:07:01 +00:00
# auto search for loopback devices
if AUDIO_INPUT_DEVICE == -2:
loopback_list = []
for dev in range(0,p.get_device_count()):
if 'Loopback: PCM' in p.get_device_info_by_index(dev)["name"]:
loopback_list.append(dev)
if len(loopback_list) >= 2:
AUDIO_INPUT_DEVICE = loopback_list[0] #0 = RX 1 = TX
print(f"loopback_list rx: {loopback_list}", file=sys.stderr)
else:
quit()
2021-12-14 22:56:55 +00:00
print(f"AUDIO INPUT DEVICE: {AUDIO_INPUT_DEVICE} DEVICE: {p.get_device_info_by_index(AUDIO_INPUT_DEVICE)['name']} \
AUDIO SAMPLE RATE: {AUDIO_SAMPLE_RATE_RX}", file=sys.stderr)
stream_rx = p.open(format=pyaudio.paInt16,
channels=1,
rate=AUDIO_SAMPLE_RATE_RX,
frames_per_buffer=AUDIO_FRAMES_PER_BUFFER,
input=True,
input_device_index=AUDIO_INPUT_DEVICE
)
2021-12-12 12:54:23 +00:00
# ----------------------------------------------------------------
# DATA CHANNEL INITIALISATION
2021-12-12 12:54:23 +00:00
# open codec2 instance
2021-12-12 20:52:03 +00:00
freedv = cast(codec2.api.freedv_open(MODE), c_void_p)
2021-12-12 12:54:23 +00:00
# get number of bytes per frame for mode
2021-12-12 20:52:03 +00:00
bytes_per_frame = int(codec2.api.freedv_get_bits_per_modem_frame(freedv)/8)
2021-12-12 12:54:23 +00:00
payload_bytes_per_frame = bytes_per_frame -2
2021-12-12 20:52:03 +00:00
n_max_modem_samples = codec2.api.freedv_get_n_max_modem_samples(freedv)
2021-12-12 12:54:23 +00:00
bytes_out = create_string_buffer(bytes_per_frame * 2)
2021-12-12 20:52:03 +00:00
codec2.api.freedv_set_frames_per_burst(freedv,N_FRAMES_PER_BURST)
total_n_bytes = 0
rx_total_frames = 0
rx_frames = 0
rx_bursts = 0
2021-12-14 20:30:20 +00:00
rx_errors = 0
2021-12-14 22:07:01 +00:00
timeout = time.time() + TIMEOUT
receive = True
2021-12-15 04:08:29 +00:00
audio_buffer = codec2.audio_buffer(codec2.api.freedv_get_n_max_modem_samples(freedv))
AUDIO_FRAMES_PER_BUFFER_8K = int(AUDIO_FRAMES_PER_BUFFER/codec2.api.FDMDV_OS_48)
resampler = codec2.resampler(AUDIO_FRAMES_PER_BUFFER,AUDIO_FRAMES_PER_BUFFER_8K)
# initial number of samples we need
nin = codec2.api.freedv_nin(freedv)
while receive and time.time() < timeout:
if AUDIO_INPUT_DEVICE != -1:
2021-12-16 01:26:39 +00:00
data_in48k = stream_rx.read(AUDIO_FRAMES_PER_BUFFER, exception_on_overflow = False)
else:
2021-12-16 01:26:39 +00:00
data_in48k = sys.stdin.buffer.read(AUDIO_FRAMES_PER_BUFFER*2)
# insert samples in buffer
x = np.frombuffer(data_in48k, dtype=np.int16)
if len(x) != AUDIO_FRAMES_PER_BUFFER:
break
x = resampler.resample48_to_8(x)
2021-12-15 04:08:29 +00:00
audio_buffer.push(x)
# when we have enough samples call FreeDV Rx
2021-12-15 04:08:29 +00:00
while audio_buffer.nbuffer >= nin:
2021-12-14 22:07:01 +00:00
# demodulate audio
2021-12-15 04:08:29 +00:00
nbytes = codec2.api.freedv_rawdatarx(freedv, bytes_out, audio_buffer.buffer.ctypes)
audio_buffer.pop(nin)
# call me on every loop!
nin = codec2.api.freedv_nin(freedv)
rx_status = codec2.api.freedv_get_rx_status(freedv)
if rx_status & codec2.api.FREEDV_RX_BIT_ERRORS:
rx_errors = rx_errors + 1
if DEBUGGING_MODE:
rx_status = codec2.api.rx_sync_flags_to_text[rx_status]
2021-12-15 04:08:29 +00:00
print("nin: %5d rx_status: %4s naudio_buffer: %4d" % \
(nin,rx_status,audio_buffer.nbuffer), file=sys.stderr)
2021-12-14 22:07:01 +00:00
if nbytes:
total_n_bytes = total_n_bytes + nbytes
2021-12-14 22:07:01 +00:00
if nbytes == bytes_per_frame:
rx_total_frames = rx_total_frames + 1
rx_frames = rx_frames + 1
if rx_frames == N_FRAMES_PER_BURST:
rx_frames = 0
rx_bursts = rx_bursts + 1
if rx_bursts == N_BURSTS:
receive = False
if time.time() >= timeout:
print("TIMEOUT REACHED")
2021-12-14 20:30:20 +00:00
print(f"RECEIVED BURSTS: {rx_bursts} RECEIVED FRAMES: {rx_total_frames} RX_ERRORS: {rx_errors}", file=sys.stderr)
2021-12-12 12:54:23 +00:00
# and at last check if we had an openend pyaudio instance and close it
if AUDIO_INPUT_DEVICE != -1:
stream_rx.close()
2021-12-12 12:54:23 +00:00
p.terminate()