From c2dee5e3c29eba79e591d9e9caebaef15ea4e56b Mon Sep 17 00:00:00 2001
From: hnluo <haoneng.lhn@alibaba-inc.com>
Date: 星期四, 29 六月 2023 11:09:28 +0800
Subject: [PATCH] Merge pull request #687 from alibaba-damo-academy/dev_lhn
---
funasr/runtime/onnxruntime/include/audio.h | 51 +++++++++++++++++++++++++++++----------------------
1 files changed, 29 insertions(+), 22 deletions(-)
diff --git a/funasr/runtime/onnxruntime/include/audio.h b/funasr/runtime/onnxruntime/include/audio.h
index 9e25aac..d2100a4 100644
--- a/funasr/runtime/onnxruntime/include/audio.h
+++ b/funasr/runtime/onnxruntime/include/audio.h
@@ -1,16 +1,17 @@
-
#ifndef AUDIO_H
#define AUDIO_H
#include <queue>
#include <stdint.h>
-#include "model.h"
+#include "vad-model.h"
+#include "offline-stream.h"
#ifndef WAV_HEADER_SIZE
#define WAV_HEADER_SIZE 44
#endif
using namespace std;
+namespace funasr {
class AudioFrame {
private:
@@ -23,17 +24,18 @@
AudioFrame(int len);
~AudioFrame();
- int set_start(int val);
- int set_end(int val);
- int get_start();
- int get_len();
- int disp();
+ int SetStart(int val);
+ int SetEnd(int val);
+ int GetStart();
+ int GetLen();
+ int Disp();
};
class Audio {
private:
- float *speech_data;
- int16_t *speech_buff;
+ float *speech_data=nullptr;
+ int16_t *speech_buff=nullptr;
+ char* speech_char=nullptr;
int speech_len;
int speech_align_len;
int offset;
@@ -45,19 +47,24 @@
Audio(int data_type);
Audio(int data_type, int size);
~Audio();
- void disp();
- bool loadwav(const char* filename, int32_t* sampling_rate);
- void wavResample(int32_t sampling_rate, const float *waveform, int32_t n);
- bool loadwav(const char* buf, int nLen, int32_t* sampling_rate);
- bool loadpcmwav(const char* buf, int nFileLen, int32_t* sampling_rate);
- bool loadpcmwav(const char* filename, int32_t* sampling_rate);
- int fetch_chunck(float *&dout, int len);
- int fetch(float *&dout, int &len, int &flag);
- void padding();
- void split(Model* pRecogObj);
- float get_time_len();
-
- int get_queue_size() { return (int)frame_queue.size(); }
+ void Disp();
+ void WavResample(int32_t sampling_rate, const float *waveform, int32_t n);
+ bool LoadWav(const char* buf, int n_len, int32_t* sampling_rate);
+ bool LoadWav(const char* filename, int32_t* sampling_rate);
+ bool LoadWav2Char(const char* filename, int32_t* sampling_rate);
+ bool LoadPcmwav(const char* buf, int n_file_len, int32_t* sampling_rate);
+ bool LoadPcmwav(const char* filename, int32_t* sampling_rate);
+ bool LoadPcmwav2Char(const char* filename, int32_t* sampling_rate);
+ int FetchChunck(float *&dout, int len);
+ int Fetch(float *&dout, int &len, int &flag);
+ void Padding();
+ void Split(OfflineStream* offline_streamj);
+ void Split(VadModel* vad_obj, vector<std::vector<int>>& vad_segments, bool input_finished=true);
+ float GetTimeLen();
+ int GetQueueSize() { return (int)frame_queue.size(); }
+ char* GetSpeechChar(){return speech_char;}
+ int GetSpeechLen(){return speech_len;}
};
+} // namespace funasr
#endif
--
Gitblit v1.9.1