Files
stafur 584587d814 Resolve "crackling" sounds in analog_audio_app (#3068)
* Updated PRALINE audio processing methods to account for potential slip in 25MHz reference clock. Moved all clocks to PLLB except for CLK0 to isolate the ADC from potential EMI and clock spills from the other clocks to CLK0. Added PLL reference to Clock ui_debug display and added debug view for PLLA and PLLB. Updated PRO menu in analog_audio_app to allow for 100Hz resolution in sample rate for finer control of samplerate debugging of crackling noise artifacts. Updated ppb method and set it for account for a potential 49 Hz shift in the PRALINE crystal. Started instatiating methods for synchronized DMA transfers for the spectrogram since the FFT calculations and the LCD data updates appear to be related to the crackling noise.

* Added firmware/flash.h to .gitignore.

* Changed reference clock for start_audio_pll to 40MHz instead of ~12MHz xtal. Moved clocks around to reflect best audio quality in analog_audio_app at 3.072 MHz sample frequency.

* Ran format-code.sh.

* Removed ppb in set_sampling_frequency since it created aliasing in analog_audio_app.

* Expanded baseband_bandwidth back to 1750000 after testing since audio quality was retained.

* Updated comments to address copilot recommendations where acceptable.

* Removed firmware/flashsize.h from .gitignore as coordinated with @gullradriel.
2026-03-04 02:11:19 -05:00

263 lines
11 KiB
C++

/*
* Copyright (C) 2014 Jared Boone, ShareBrained Technology, Inc.
* Copyright (C) 2016 Furrtek
*
* This file is part of PortaPack.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2, or (at your option)
* any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; see the file COPYING. If not, write to
* the Free Software Foundation, Inc., 51 Franklin Street,
* Boston, MA 02110-1301, USA.
*/
#include "proc_wfm_audio.hpp"
#include "portapack_shared_memory.hpp"
#include "audio_output.hpp"
#include "dsp_fft.hpp"
#include "event_m4.hpp"
#include "audio_dma.hpp"
#include <cstdint>
void WidebandFMAudio::execute(const buffer_c8_t& buffer) {
if (!configured) {
return;
}
const auto decim_0_out = decim_0.execute(buffer, dst_buffer);
const auto channel = decim_1.execute(decim_0_out, dst_buffer);
// TODO: Feed channel_stats post-decimation data?
feed_channel_stats(channel);
spectrum_samples += channel.count;
if (spectrum_samples >= spectrum_interval_samples) {
spectrum_samples -= spectrum_interval_samples;
channel_spectrum.feed(channel, channel_filter_low_f, channel_filter_high_f, channel_filter_transition);
}
/* 384kHz complex<int16_t>[256] for wfm
* -> FM demodulation
* -> 384kHz int16_t[256] */
/* TODO: To improve adjacent channel rejection, implement complex channel filter:
* pass < +/- 100kHz, stop > +/- 200kHz
*/
/* 96kHz complex<int16_t>[64] for wfmam NOAA
* -> FM demodulation
* -> 96kHz int16_t[64] */
auto audio_oversampled = demod.execute(channel, work_audio_buffer); // fs 384khz wfm , 96khz wfmam for NOAA
/* 384kHz int16_t[256] for wfm
* -> 4th order CIC decimation by 2, gain of 1
* -> 192kHz int16_t[128] */
/* 96kHz int16_t[64] for wfam
* -> 4th order CIC decimation by 2, gain of 1
* -> 48kHz int16_t[32] */
auto audio_4fs = audio_dec_1.execute(audio_oversampled, work_audio_buffer);
/* 192kHz int16_t[128] for wfm
* -> 4th order CIC decimation by 2, gain of 1
* -> 96kHz int16_t[64] */
/* 48kHz int16_t[32] for wfman
* -> 4th order CIC decimation by 2, gain of 1
* -> 24kHz int16_t[16] */
auto audio_2fs = audio_dec_2.execute(audio_4fs, work_audio_buffer);
// Input: 96kHz int16_t[64]
// audio_spectrum_decimator piles up 256 samples before doing FFT computation
// This sends an AudioSpectrum every: sample rate/buffer size/refresh period = 3072000/2048/50 = 30 Hz
// When audio_spectrum_timer expires, the audio spectrum computation is triggered
// 0~3: feed continuous audio
// 4~31: ignore, wrap at 31
audio_spectrum_timer++;
if (audio_spectrum_timer == 50) {
audio_spectrum_timer = 0;
audio_spectrum_state = FEED;
}
switch (audio_spectrum_state) {
case FEED:
// Convert audio to "complex" just so the FFT can be done :/
for (size_t i = 0; i < 64; i++) {
complex_audio[i] = {(int16_t)(work_audio_buffer.p[i] / 32), (int16_t)0};
}
audio_spectrum_decimator.feed(
complex_audio_buffer,
[this](const buffer_c16_t& data) {
this->post_message(data);
});
break;
case FFT:
// Spread the FFT workload in time to avoid making the audio skip
// "8" comes from the log2() of the size of audio_spectrum: log2(256) = 8
if (fft_step < 8) {
fft_c_preswapped(audio_spectrum, fft_step, fft_step + 1);
fft_step++;
} else {
constexpr float mag_scale = 5.0f;
const size_t spectrum_end = spectrum.db.size();
for (size_t i = 0; i < spectrum_end; i++) {
// const auto corrected_sample = spectrum_window_hamming_3(audio_spectrum, i);
const auto corrected_sample = audio_spectrum[i];
const auto mag2 = magnitude_squared(corrected_sample * (1.0f / 32768.0f));
const float db = mag2_to_dbv_norm(mag2);
const unsigned int v = (db * mag_scale) + 255.0f;
spectrum.db[i] = std::max(0U, std::min(255U, v));
}
AudioSpectrumMessage message{&spectrum};
shared_memory.application_queue.push(message);
audio_spectrum_state = IDLE;
}
break;
default:
break;
}
/* 96kHz int16_t[64] for wfm
* -> FIR filter, <15kHz (0.156fs) pass, >19kHz (0.198fs) stop, gain of 1
* -> 48kHz int16_t[32] */
/* 24kHz int16_t[16] for wfmam
* -> FIR filter, <4.5kHz (0.1875fs) pass, >5.2kHz (0.2166fs) stop, gain of 1
* -> 12kHz int16_t[8] */
auto audio = audio_filter.execute(audio_2fs, work_audio_buffer);
/* -> 48kHz int16_t[32] for wfm , */
/* -> 12kHz int16_t[8] for wfmam , */
if (decim_1.decimation_factor() == 2) {
audio_output.write(audio); // we are in original wfm , decim_1.decimation_factor == 2
} else {
audio_output.apt_write(audio); // we are in added wfmam (noaa), decim_1.decimation_factor == 8
}
}
void WidebandFMAudio::post_message(const buffer_c16_t& data) {
// This is called when audio_spectrum_decimator is filled up to 256 samples
fft_swap(data, audio_spectrum);
audio_spectrum_state = FFT;
fft_step = 0;
}
void WidebandFMAudio::on_message(const Message* const message) {
switch (message->id) {
case Message::ID::UpdateSpectrum:
case Message::ID::SpectrumStreamingConfig:
channel_spectrum.on_message(message);
break;
case Message::ID::WFMConfigure:
configure_wfm(*reinterpret_cast<const WFMConfigureMessage*>(message));
break;
case Message::ID::WFMAMConfigure:
configure_wfmam(*reinterpret_cast<const WFMAMConfigureMessage*>(message));
break;
case Message::ID::CaptureConfig:
capture_config(*reinterpret_cast<const CaptureConfigMessage*>(message));
break;
default:
break;
}
}
void WidebandFMAudio::configure_wfm(const WFMConfigureMessage& message) {
constexpr size_t decim_0_input_fs = baseband_fs;
constexpr size_t decim_0_output_fs = decim_0_input_fs / decim_0.decimation_factor;
constexpr size_t decim_1_input_fs = decim_0_output_fs;
decim_0.configure(message.decim_0_filter.taps);
// decim_1.configure(message.decim_1_filter.taps); // Original .
// TODO dynamic decim1 , with decimation 2 / 8 and 16 x taps , / 32 taps .
// Temptatively , I splitted, in two WidebandFMAudio::configure_wfm / WidebandFMAudio::configure_wfmam and dynamically /2, /8 (here /2)
// decim_1.set<dsp::decimate::FIRC16xR16x16Decim2>().configure(message.decim_1_filter.taps); // for wfm
// decim_1.set<dsp::decimate::FIRC16xR16x32Decim8>().configure(taps_84k_wfmam_decim_1.taps); // for wfmam
decim_1.set<dsp::decimate::FIRC16xR16x16Decim2>().configure(message.decim_1_filter.taps); // for wfm
size_t decim_1_output_fs = decim_1_input_fs / decim_1.decimation_factor(); // wfm , decim_1.decimation_factor() = /2 , if applied after the line : decim_1.set<dsp::decimate::FIRC16xR16x16Decim2>().configure(message.decim_1_filter.taps);
size_t demod_input_fs = decim_1_output_fs;
spectrum_interval_samples = decim_1_output_fs / spectrum_rate_hz;
spectrum_samples = 0;
channel_filter_low_f = message.decim_1_filter.low_frequency_normalized * decim_1_input_fs;
channel_filter_high_f = message.decim_1_filter.high_frequency_normalized * decim_1_input_fs;
channel_filter_transition = message.decim_1_filter.transition_normalized * decim_1_input_fs;
demod.configure(demod_input_fs, message.deviation);
audio_filter.configure(message.audio_filter.taps);
audio_output.configure(message.audio_hpf_config, message.audio_deemph_config);
channel_spectrum.set_decimation_factor(1);
configured = true;
}
void WidebandFMAudio::configure_wfmam(const WFMAMConfigureMessage& message) {
constexpr size_t decim_0_input_fs = baseband_fs;
constexpr size_t decim_0_output_fs = decim_0_input_fs / decim_0.decimation_factor;
constexpr size_t decim_1_input_fs = decim_0_output_fs;
decim_0.configure(message.decim_0_filter.taps);
// decim_1.configure(message.decim_1_filter.taps); // Original .
// TODO dynamic decim1 , with decimation 2 / 8 and 16 x taps , / 32 taps .
// Temptatively , I splitted, in two WidebandFMAudio::configure_wfm / WidebandFMAudio::configure_wfmam and dynamically /2, /8 . (here /8)
// decim_1.set<dsp::decimate::FIRC16xR16x16Decim2>().configure(message.decim_1_filter.taps); // for wfm
// decim_1.set<dsp::decimate::FIRC16xR16x32Decim8>().configure(message.decim_1_filter.taps); // for wfmam
decim_1.set<dsp::decimate::FIRC16xR16x32Decim8>().configure(message.decim_1_filter.taps); // for wfmam
size_t decim_1_output_fs = decim_1_input_fs / decim_1.decimation_factor(); // wfmam, decim_1.decimation_factor() = /8 ,if applied after the line, decim_1.set<dsp::decimate::FIRC16xR16x16Decim2>().configure(message.decim_1_filter.taps);
size_t demod_input_fs = decim_1_output_fs;
spectrum_interval_samples = decim_1_output_fs / spectrum_rate_hz;
spectrum_samples = 0;
channel_filter_low_f = message.decim_1_filter.low_frequency_normalized * decim_1_input_fs;
channel_filter_high_f = message.decim_1_filter.high_frequency_normalized * decim_1_input_fs;
channel_filter_transition = message.decim_1_filter.transition_normalized * decim_1_input_fs;
demod.configure(demod_input_fs, message.deviation);
audio_filter.configure(message.audio_filter.taps);
audio_output.configure(message.audio_hpf_config, message.audio_deemph_config);
channel_spectrum.set_decimation_factor(1);
configured = true;
}
void WidebandFMAudio::capture_config(const CaptureConfigMessage& message) {
if (message.config) {
audio_output.set_stream(std::make_unique<StreamInput>(message.config));
} else {
audio_output.set_stream(nullptr);
}
}
int main() {
audio::dma::init_audio_out();
EventDispatcher event_dispatcher{std::make_unique<WidebandFMAudio>()};
event_dispatcher.run();
return 0;
}