kopia lustrzana https://github.com/N0BOY/FT8CN
151 wiersze
5.9 KiB
C
151 wiersze
5.9 KiB
C
//
|
||
// Created by jmsmf on 2022/6/1.
|
||
//
|
||
|
||
#include "ft8Encoder.h"
|
||
|
||
//#define LOG_LEVEL LOG_INFO
|
||
|
||
#define FT8_SYMBOL_BT 2.0f /// 符号平滑滤波器带宽因子（BT）
|
||
#define FT4_SYMBOL_BT 1.0f /// 符号平滑滤波器带宽因子（BT）
|
||
#define GFSK_CONST_K 5.336446f ///< == pi * sqrt(2 / log(2))
|
||
|
||
|
||
/// 生成高斯平滑脉冲
|
||
/// 脉冲理论上是无限长的，然而，这里它被截断为符号长度的3倍。
|
||
/// 这意味着脉冲阵列必须有空间容纳3*n_spsym元素。
|
||
/// @param[in] n_spsym 每个符号的样本数 Number of samples per symbol
|
||
/// @param[in] b 形状参数（为FT8/FT4定义的值）
|
||
/// @param[out] pulse 脉冲采样输出阵列
|
||
///
|
||
void gfsk_pulse(int n_spsym, float symbol_bt, float *pulse) {
|
||
for (int i = 0; i < 3 * n_spsym; ++i) {
|
||
float t = i / (float) n_spsym - 1.5f;
|
||
float arg1 = GFSK_CONST_K * symbol_bt * (t + 0.5f);
|
||
float arg2 = GFSK_CONST_K * symbol_bt * (t - 0.5f);
|
||
pulse[i] = (erff(arg1) - erff(arg2)) / 2;
|
||
}
|
||
}
|
||
|
||
|
||
/// 使用GFSK相位整形合成波形数据。
|
||
/// 输出波形将包含n_sym个符号。
|
||
/// @param[in] symbols 符号（音调）数组 (0-7 for FT8)
|
||
/// @param[in] n_sym 符号数组中的符号数
|
||
/// @param[in] f0 符号0的音频频率（赫兹） (载波频率)
|
||
/// @param[in] symbol_bt 符号平滑滤波器带宽 (2 for FT8, 1 for FT4)
|
||
/// @param[in] symbol_period 符号周期（持续时间），秒
|
||
/// @param[in] signal_rate 合成信号的采样率，赫兹
|
||
/// @param[out] signal 信号波形样本的输出阵列（应为n_sym*n_spsym样本留出空间）
|
||
///
|
||
void synth_gfsk(const uint8_t *symbols, int n_sym, float f0, float symbol_bt, float symbol_period,
|
||
int signal_rate, float *signal) {
|
||
int n_spsym = (int) (0.5f + (float)signal_rate * symbol_period); // 每个符号的样本数12000*0.16=1920
|
||
int n_wave = n_sym * n_spsym; // 输出样本数79*1920=151680
|
||
float hmod = 1.0f;
|
||
|
||
|
||
// 计算平滑的频率波形。
|
||
// Length = (nsym+2)*n_spsym samples, 首个和最后一个扩展符号
|
||
float dphi_peak = 2 * M_PI * hmod / n_spsym;
|
||
|
||
//此处是与采样率有关，采样率提高后，可能会有闪退的问题
|
||
float *dphi;//此处使用内存申请的方式，而不是原来数组的方式，因为是采样率过高时，会报内存出错。
|
||
dphi = malloc(sizeof(float) * (n_wave + 2 * n_spsym));
|
||
if (dphi==0) return;//内存申请失败
|
||
//float dphi[n_wave + 2 * n_spsym];//原来的方式
|
||
|
||
// 频率上移f0
|
||
for (int i = 0; i < n_wave + 2 * n_spsym; ++i) {
|
||
dphi[i] = 2 * M_PI * f0 / signal_rate;
|
||
}
|
||
|
||
//float pulse[3 * n_spsym];
|
||
float *pulse=(float *) malloc(sizeof(float)*3 * n_spsym);
|
||
gfsk_pulse(n_spsym, symbol_bt, pulse);
|
||
|
||
for (int i = 0; i < n_sym; ++i) {
|
||
int ib = i * n_spsym;
|
||
for (int j = 0; j < 3 * n_spsym; ++j) {
|
||
dphi[j + ib] += dphi_peak * symbols[i] * pulse[j];
|
||
}
|
||
}
|
||
// 在开头和结尾添加伪符号，音调值分别等于第一个符号和最后一个符号
|
||
for (int j = 0; j < 2 * n_spsym; ++j) {
|
||
dphi[j] += dphi_peak * pulse[j + n_spsym] * symbols[0];
|
||
dphi[j + n_sym * n_spsym] += dphi_peak * pulse[j] * symbols[n_sym - 1];
|
||
}
|
||
// 计算并插入音频波形
|
||
float phi = 0;
|
||
for (int k = 0; k < n_wave; ++k) { // 不包括虚拟符号
|
||
signal[k] = sinf(phi);
|
||
phi = fmodf(phi + dphi[k + n_spsym], 2 * M_PI);
|
||
}
|
||
// 对第一个和最后一个符号应用封套成形，此处是前后增加斜坡函数，
|
||
int n_ramp = n_spsym / 8;//240个样本，20毫秒，T/8
|
||
for (int i = 0; i < n_ramp; ++i) {
|
||
float env = (1 - cosf(2 * M_PI * i / (2 * n_ramp))) / 2;
|
||
signal[i] *= env;
|
||
signal[n_wave - 1 - i] *= env;
|
||
}
|
||
free(pulse);
|
||
free(dphi);//要释放掉内存
|
||
}
|
||
|
||
//此代码已经弃用
|
||
void generateFt8ToBuffer(char *message, float frequency, short *buffer) {
|
||
// 首先，将文本数据打包为二进制消息
|
||
uint8_t packed[FTX_LDPC_K_BYTES];//91位，包括CRC。
|
||
int rc = pack77(message, packed);//生成数据
|
||
if (rc < 0) {
|
||
//LOGE("Cannot parse message!\n");
|
||
//LOGE("RC = %d\n", rc);
|
||
return;
|
||
}
|
||
|
||
|
||
//int num_tones = FT8_NN;//符号数量：FT8是79个，FT4是105个。
|
||
//float symbol_period = FT8_SYMBOL_PERIOD;//FT8_SYMBOL_PERIOD=0.160f
|
||
float symbol_bt = FT8_SYMBOL_BT;//FT8_SYMBOL_BT=2.0f
|
||
float slot_time = FT8_SLOT_TIME;//FT8_SLOT_TIME=15f
|
||
|
||
// 其次，将二进制消息编码为FSK音调序列
|
||
uint8_t tones[FT8_NN]; // 79音调（符号）数组
|
||
ft8_encode(packed, tones);
|
||
|
||
|
||
|
||
// 第三，将FSK音调转换为音频信号b
|
||
//int sample_rate = FT8_SAMPLE_RATE;//采样率
|
||
int num_samples = (int) (0.5f + FT8_NN * FT8_SYMBOL_PERIOD *
|
||
FT8_SAMPLE_RATE); // 数据信号中的采样数0.5+79*0.16*12000
|
||
//int num_silence = (slot_time * sample_rate - num_samples) / 2; // 两端填充静音到15秒（15*12000-num_samples）/2（1.18秒的样本数）
|
||
int num_silence = 20;//把前面的静音时长缩短为20毫秒，留出时间给解码
|
||
//int num_total_samples = num_silence + num_samples + num_silence; // 填充信号中的样本数2.36秒+12.64秒=15秒的样本数
|
||
float signal[Ft8num_samples];
|
||
//Ft8num_sampleFT8声音的总采样数，不是字节数。15*12000
|
||
for (int i = 0; i < Ft8num_samples; i++)//把数据全部静音。
|
||
{
|
||
signal[i] = 0;
|
||
//buffer[i + num_samples + num_silence] = 0;
|
||
}
|
||
|
||
// 合成波形数据（信号）并将其保存为WAV文件
|
||
synth_gfsk(tones, FT8_NN, frequency, symbol_bt, FT8_SYMBOL_PERIOD, FT8_SAMPLE_RATE,
|
||
signal + num_silence);
|
||
|
||
|
||
for (int i = 0; i < Ft8num_samples; i++) {
|
||
float x = signal[i];
|
||
if (x > 1.0)
|
||
x = 1.0;
|
||
else if (x < -1.0)
|
||
x = -1.0;
|
||
buffer[i] = (short) (0.5 + (x * 32767.0));
|
||
}
|
||
|
||
|
||
//save_wav(signal, num_total_samples, sample_rate, wav_path);
|
||
|
||
}
|