使用 OpenSL 进行通话录音

本文介绍了使用 OpenSL 进行通话录音的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！

问题描述

自从我的 Galaxy S5 中的棒棒糖更新以来，我尝试修复我的应用程序中的通话录音.作为基础，我从这里使用谷歌示例项目:示例.

I try to fix call recording in my app since lolipop update in my Galaxy S5. As a base I am using google sample project from here: Sample.

这是代码的主要部分:

SLresult result;
sampleInfo_ = *sampleFormat;
SLAndroidDataFormat_PCM_EX format_pcm;
ConvertToSLSampleFormat(&format_pcm, &sampleInfo_);

gFp = fopen("/storage/emulated/0/file.pcm", "w");

// configure audio source
SLDataLocator_IODevice loc_dev = {SL_DATALOCATOR_IODEVICE, SL_IODEVICE_AUDIOINPUT,
                                  SL_DEFAULTDEVICEID_AUDIOINPUT, NULL};
SLDataSource audioSrc = {&loc_dev, NULL};


// configure audio sink
SLDataLocator_AndroidSimpleBufferQueue loc_bq = {
        SL_DATALOCATOR_ANDROIDSIMPLEBUFFERQUEUE,
        DEVICE_SHADOW_BUFFER_QUEUE_LEN};

SLDataSink audioSnk = {&loc_bq, &format_pcm};

// create audio recorder
// (requires the RECORD_AUDIO permission)
const SLInterfaceID id[2] = {SL_IID_ANDROIDSIMPLEBUFFERQUEUE,
                             SL_IID_ANDROIDCONFIGURATION};

const SLboolean req[2] = {SL_BOOLEAN_FALSE, SL_BOOLEAN_FALSE};

result = (*slEngine)->CreateAudioRecorder(slEngine,
                                          &recObjectItf_,
                                          &audioSrc,
                                          &audioSnk,
                                          2,
                                          id, req);
SLASSERT(result);

// Configure the voice recognition preset which has no
// signal processing for lower latency.
SLAndroidConfigurationItf inputConfig;
result = (*recObjectItf_)->GetInterface(recObjectItf_,
                                        SL_IID_ANDROIDCONFIGURATION,
                                        &inputConfig);

SLuint32 presetValue = SL_ANDROID_RECORDING_PRESET_VOICE_COMMUNICATION;
result = (*inputConfig)->SetConfiguration(inputConfig,
                                          SL_ANDROID_KEY_RECORDING_PRESET,
                                          &presetValue,
                                          sizeof(SLint32));
SLASSERT(result);


result = (*recObjectItf_)->Realize(recObjectItf_, SL_BOOLEAN_FALSE);
SLASSERT(result);

result = (*recObjectItf_)->GetInterface(recObjectItf_, SL_IID_RECORD, &recItf_);
SLASSERT(result);

result = (*recObjectItf_)->GetInterface(recObjectItf_, SL_IID_ANDROIDSIMPLEBUFFERQUEUE,
                                        &recBufQueueItf_);
SLASSERT(result);

result = (*recBufQueueItf_)->RegisterCallback(recBufQueueItf_, bqRecorderCallback, this);
SLASSERT(result);

devShadowQueue_ = new AudioQueue(DEVICE_SHADOW_BUFFER_QUEUE_LEN);
assert(devShadowQueue_);

这是我的问题，这段代码没有记录通话的另一端，在输出文件中我只能听到麦克风的声音.我尝试更改参数但结果相同.有人知道我做错了什么吗?

And here is my problem this code do not record other side of the call, in output file I can only hear voice from microphone. I tried changing parameters but with same result. Anyone know what I am doing wrong?

推荐答案

我找到了修复 Galaxy S5 通话录音的解决方案.

I found the solution to fix Galaxy S5 call recording.

主要是调用这个:status_t AudioSystem::setParameters(audio_io_handle_t ioHandle, const String8& keyValuePairs) 在调用开始时循环.

Main thing is to call this:status_t AudioSystem::setParameters(audio_io_handle_t ioHandle, const String8& keyValuePairs) in the loop when the call is started.

首先得到想要的函数:

open_media = dlopen("/system/lib/libmedia.so", RTLD_LAZY);

set_parameters = (int (*)(int, void *)) dlsym(open_media,
                                                  "_ZN7android11AudioSystem13setParametersEiRKNS_7String8E");

接下来我们需要 audio_io_handle_t 和 String8& 对象:

Next we need audio_io_handle_t and String8& object:

audio_io_handle_t- 是audio session id加1，你可以从AudioRecord.getAudioSessionId
String8& 这个比较难:

audio_io_handle_t- is audio session id increased by 1, You can get it from AudioRecord.getAudioSessionId
String8& this is more difficult:

//First inicialize function
create_string = (void (*)(void *, const char *)) dlsym(open_util,
"_ZN7android7String8C2EPKc");

//next call this function to convert string to required object
create_string(&str8, str);

当我们拥有所有需要的部分时，我们可以调用 setParameters 函数:

When we have all needed parts we can call setParameters function:

//remember to call this in loop when recording is starting
set_parameters(id + 1, &str8);

变量声明的样子:

int (*set_parameters)(int, void *);

void (*create_string)(void *, const char *);

void *str8 = 0;
const char *str = "input_source=4";

@ChanchalShelar

@Peter @AkshatVajpayee

这是我的 .cpp 文件的外观:

This is how my .cpp file looks:

void *open_media;
void *open_util;

int (*set_parameters)(int, void *);

void (*create_string)(void *, const char *);

void *str8 = 0;
const char *str = "input_source=4";


extern "C" {
    JNIEXPORT bool JNICALL
    Java_com_sample_NativeAudio_init(JNIEnv *env, jclass);
    JNIEXPORT int JNICALL
    Java_com_sample_NativeAudio_setParameters(JNIEnv *env, jclass, int id);
}


void get_string8() {
    create_string = (void (*)(void *, const char *)) dlsym(open_util, "_ZN7android7String8C2EPKc");

    if (!create_string) {
        LOGD("There is no create_string function");
    } else {
        LOGD("create_string function OK");
    }

    create_string(&str8, str);
    if (!str8) {
        LOGD("Filed to create str8");
    } else {
        LOGD("create str8 success");
    }

}

JNIEXPORT int JNICALL Java_com_sample_NativeAudio_setParameters(JNIEnv *env,
                                                                  jclass     type, int id) {
    if (set_parameters) {
        return set_parameters(id + 1, &str8);
    }

    return 0;
}


JNIEXPORT bool JNICALL Java_com_sample_NativeAudio_init(JNIEnv *env, jclass type) {

    open_util = dlopen("/system/lib/libutils.so", RTLD_LAZY);

    if (open_util) {
        get_string8();
    } else {
        return false;
    }

    open_media = dlopen("/system/lib/libmedia.so", RTLD_LAZY);

    if (open_media) {
        set_parameters = (int (*)(int, void *)) dlsym(open_media,
                                              "_ZN7android11AudioSystem13setParametersEiRKNS_7String8E");
    } else {
        return false;
    }

    return true;
}

这篇关于使用 OpenSL 进行通话录音的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！