一个使用FFmpeg库读取3gp视频的例子-Android中使用FFmpeg媒体库(三)

简介:

原文:http://doandroid.info/?p=497

在续系列文章在32位的Ubuntu 11.04中为Android NDK r6编译FFmpeg0.8.1版-Android中使用FFmpeg媒体库(一)在Android中通过jni方式使用编译好的FFmpeg库-Android中使用FFmpeg媒体库(二)文章后,本文将根据github中churnlabs的一个开源项目,来深入展开说明如何使用FFmpeg库进行多媒体的开发。
本文中的代码来自于https://github.com/churnlabs/android-ffmpeg-sample,更多的可以参考这个项目代码。我会在代码中加一些自己的注释。感谢作者churnlabs给我们提供这么好的例子以供我们学习。
在Android的一些系统层应用开发大多数是采用jni的方式调用,另外对于一些比较吃CPU或者处理逻辑比较复杂的程序,也可以考虑使用jni方式来封装。可以提高程序的执行效率。

本文涉及到以下几个方面:
1 将3gp文件push到模拟机器的sdcard中
2 写jni代码,内部调用ffmpeg库的方法,编译jni库
3 loadLibrary生成的库,然后撰写相应的java代码
4 执行程序,并查看最终运行结果。
最终程序的显示效果如下:


1 使用eclipse的DDMS工具,将vid.3pg push到sdcard中

2 撰写相应的jni文件

/*
 * Copyright 2011 - Churn Labs, LLC
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
 
/*
 * This is mostly based off of the FFMPEG tutorial:
 * http://dranger.com/ffmpeg/
 * With a few updates to support Android output mechanisms and to update
 * places where the APIs have shifted.
 */
 
#include <jni.h>
#include <string.h>
#include <stdio.h>
#include <android/log.h>
#include <android/bitmap.h>
 
//包含ffmpeg库头文件,这些文件都直接方案jni目录下
#include <libavcodec/avcodec.h>
#include <libavformat/avformat.h>
#include <libswscale/swscale.h>
 
#define  LOG_TAG    "FFMPEGSample"
#define  LOGI(...)  __android_log_print(ANDROID_LOG_INFO,LOG_TAG,__VA_ARGS__)
#define  LOGE(...)  __android_log_print(ANDROID_LOG_ERROR,LOG_TAG,__VA_ARGS__)
 
/* Cheat to keep things simple and just use some globals. */
//全局对象
AVFormatContext *pFormatCtx;
AVCodecContext *pCodecCtx;
AVFrame *pFrame;
AVFrame *pFrameRGB;
int videoStream;
 
/*
 * Write a frame worth of video (in pFrame) into the Android bitmap
 * described by info using the raw pixel buffer.  It's a very inefficient
 * draw routine, but it's easy to read. Relies on the format of the
 * bitmap being 8bits per color component plus an 8bit alpha channel.
 */
 
//定义的静态方法,将某帧AVFrame在Android的Bitmap中绘制
static void fill_bitmap(AndroidBitmapInfo*  info, void *pixels, AVFrame *pFrame)
{
    uint8_t *frameLine;
 
    int  yy;
    for (yy = 0; yy < info->height; yy++) {
        uint8_t*  line = (uint8_t*)pixels;
        frameLine = (uint8_t *)pFrame->data[0] + (yy * pFrame->linesize[0]);
 
        int xx;
        for (xx = 0; xx < info->width; xx++) {
            int out_offset = xx * 4;
            int in_offset = xx * 3;
 
            line[out_offset] = frameLine[in_offset];
            line[out_offset+1] = frameLine[in_offset+1];
            line[out_offset+2] = frameLine[in_offset+2];
            line[out_offset+3] = 0;
        }
        pixels = (char*)pixels + info->stride;
    }
}
 
//定义java回调函数,相当与 com.churnlabs中的ffmpegsample中的MainActivity类中的openFile方法。
void Java_com_churnlabs_ffmpegsample_MainActivity_openFile(JNIEnv * env, jobject this)
{
    int ret;
    int err;
    int i;
    AVCodec *pCodec;
    uint8_t *buffer;
    int numBytes;
    //注册所有的函数
    av_register_all();
    LOGE("Registered formats");
    //打开sdcard中的vid.3gp文件
    err = av_open_input_file(&pFormatCtx, "file:/sdcard/vid.3gp", NULL, 0, NULL);
    LOGE("Called open file");
    if(err!=0) {
        LOGE("Couldn't open file");
        return;
    }
    LOGE("Opened file");
 
    if(av_find_stream_info(pFormatCtx)<0) {
        LOGE("Unable to get stream info");
        return;
    }
 
    videoStream = -1;
    //定义设置videoStream
    for (i=0; i<pFormatCtx->nb_streams; i++) {
        if(pFormatCtx->streams[i]->codec->codec_type==CODEC_TYPE_VIDEO) {
            videoStream = i;
            break;
        }
    }
    if(videoStream==-1) {
        LOGE("Unable to find video stream");
        return;
    }
 
    LOGI("Video stream is [%d]", videoStream);
    //定义编码类型
    pCodecCtx=pFormatCtx->streams[videoStream]->codec;
    //获取解码器
    pCodec=avcodec_find_decoder(pCodecCtx->codec_id);
    if(pCodec==NULL) {
        LOGE("Unsupported codec");
        return;
    }
    //使用特定的解码器打开
    if(avcodec_open(pCodecCtx, pCodec)<0) {
        LOGE("Unable to open codec");
        return;
    }
    //分配帧空间
    pFrame=avcodec_alloc_frame();
    //分配RGB帧空间
    pFrameRGB=avcodec_alloc_frame();
    LOGI("Video size is [%d x %d]", pCodecCtx->width, pCodecCtx->height);
    //获取大小
    numBytes=avpicture_get_size(PIX_FMT_RGB24, pCodecCtx->width, pCodecCtx->height);
    分配空间
    buffer=(uint8_t *)av_malloc(numBytes*sizeof(uint8_t));
 
    avpicture_fill((AVPicture *)pFrameRGB, buffer, PIX_FMT_RGB24,
                            pCodecCtx->width, pCodecCtx->height);
}
//定义java回调函数,相当与 com.churnlabs中的ffmpegsample中的MainActivity类中的drawFrame方法。
void Java_com_churnlabs_ffmpegsample_MainActivity_drawFrame(JNIEnv * env, jobject this, jstring bitmap)
{
    AndroidBitmapInfo  info;
    void*              pixels;
    int                ret;
 
    int err;
    int i;
    int frameFinished = 0;
    AVPacket packet;
    static struct SwsContext *img_convert_ctx;
    int64_t seek_target;
 
    if ((ret = AndroidBitmap_getInfo(env, bitmap, &info)) < 0) {
        LOGE("AndroidBitmap_getInfo() failed ! error=%d", ret);
        return;
    }
    LOGE("Checked on the bitmap");
 
    if ((ret = AndroidBitmap_lockPixels(env, bitmap, &pixels)) < 0) {
        LOGE("AndroidBitmap_lockPixels() failed ! error=%d", ret);
    }
    LOGE("Grabbed the pixels");
 
    i = 0;
    while((i==0) && (av_read_frame(pFormatCtx, &packet)>=0)) {
        if(packet.stream_index==videoStream) {
            avcodec_decode_video2(pCodecCtx, pFrame, &frameFinished, &packet);
 
            if(frameFinished) {
                LOGE("packet pts %llu", packet.pts);
                // This is much different than the tutorial, sws_scale
                // replaces img_convert, but it's not a complete drop in.
                // This version keeps the image the same size but swaps to
                // RGB24 format, which works perfect for PPM output.
                int target_width = 320;
                int target_height = 240;
                img_convert_ctx = sws_getContext(pCodecCtx->width, pCodecCtx->height,
                       pCodecCtx->pix_fmt,
                       target_width, target_height, PIX_FMT_RGB24, SWS_BICUBIC,
                       NULL, NULL, NULL);
                if(img_convert_ctx == NULL) {
                    LOGE("could not initialize conversion context\n");
                    return;
                }
                sws_scale(img_convert_ctx, (const uint8_t* const*)pFrame->data, pFrame->linesize, 0, pCodecCtx->height, pFrameRGB->data, pFrameRGB->linesize);
 
                // save_frame(pFrameRGB, target_width, target_height, i);
                fill_bitmap(&info, pixels, pFrameRGB);
                i = 1;
            }
        }
        av_free_packet(&packet);
    }
 
    AndroidBitmap_unlockPixels(env, bitmap);
}
 
//内部调用函数,不对外,用来查找帧
int seek_frame(int tsms)
{
    int64_t frame;
 
    frame = av_rescale(tsms,pFormatCtx->streams[videoStream]->time_base.den,pFormatCtx->streams[videoStream]->time_base.num);
    frame/=1000;
 
    if(avformat_seek_file(pFormatCtx,videoStream,0,frame,frame,AVSEEK_FLAG_FRAME)<0) {
        return 0;
    }
 
    avcodec_flush_buffers(pCodecCtx);
 
    return 1;
}
//定义java回调函数,相当与 com.churnlabs中的ffmpegsample中的MainActivity类中的drawFrameAt方法。
void Java_com_churnlabs_ffmpegsample_MainActivity_drawFrameAt(JNIEnv * env, jobject this, jstring bitmap, jint secs)
{
    AndroidBitmapInfo  info;
    void*              pixels;
    int                ret;
 
    int err;
    int i;
    int frameFinished = 0;
    AVPacket packet;
    static struct SwsContext *img_convert_ctx;
    int64_t seek_target;
 
    if ((ret = AndroidBitmap_getInfo(env, bitmap, &info)) < 0) {
        LOGE("AndroidBitmap_getInfo() failed ! error=%d", ret);
        return;
    }
    LOGE("Checked on the bitmap");
 
    if ((ret = AndroidBitmap_lockPixels(env, bitmap, &pixels)) < 0) {
        LOGE("AndroidBitmap_lockPixels() failed ! error=%d", ret);
    }
    LOGE("Grabbed the pixels");
 
    seek_frame(secs * 1000);
 
    i = 0;
    while ((i== 0) && (av_read_frame(pFormatCtx, &packet)>=0)) {
        if(packet.stream_index==videoStream) {
            avcodec_decode_video2(pCodecCtx, pFrame, &frameFinished, &packet);
 
            if(frameFinished) {
                // This is much different than the tutorial, sws_scale
                // replaces img_convert, but it's not a complete drop in.
                // This version keeps the image the same size but swaps to
                // RGB24 format, which works perfect for PPM output.
                int target_width = 320;
                int target_height = 240;
                img_convert_ctx = sws_getContext(pCodecCtx->width, pCodecCtx->height,
                       pCodecCtx->pix_fmt,
                       target_width, target_height, PIX_FMT_RGB24, SWS_BICUBIC,
                       NULL, NULL, NULL);
                if(img_convert_ctx == NULL) {
                    LOGE("could not initialize conversion context\n");
                    return;
                }
                sws_scale(img_convert_ctx, (const uint8_t* const*)pFrame->data, pFrame->linesize, 0, pCodecCtx->height, pFrameRGB->data, pFrameRGB->linesize);
 
                // save_frame(pFrameRGB, target_width, target_height, i);
                fill_bitmap(&info, pixels, pFrameRGB);
                i = 1;
            }
        }
        av_free_packet(&packet);
    }
 
    AndroidBitmap_unlockPixels(env, bitmap);
}

3 撰写相应的Android.mk文件

LOCAL_PATH := $(call my-dir)
 
include $(CLEAR_VARS)
 
LOCAL_MODULE    := ffmpegutils
LOCAL_SRC_FILES := native.c
 
LOCAL_C_INCLUDES := $(LOCAL_PATH)/include
LOCAL_LDLIBS := -L$(NDK_PLATFORMS_ROOT)/$(TARGET_PLATFORM)/arch-arm/usr/lib -L$(LOCAL_PATH) -lavformat -lavcodec -lavdevice -lavfilter -lavcore -lavutil -lswscale -llog -ljnigraphics -lz -ldl -lgcc
 
include $(BUILD_SHARED_LIBRARY)

这里需要注意一下文件的目录情况,我截图说明一下。

在Android.mk中有意个LOCAL_C_INCLUDES :=$(LOCAL_PATH)/include指明了相应的FFmpeg的头文件路径。故在代码中包含

#include <libavcodec/avcodec.h>
#include <libavformat/avformat.h>
#include <libswscale/swscale.h>

就可以。

4 调用ndk-build,生成libffmpegutils.so文件,将这个文件拷贝到/root/develop/android-ndk-r6/platforms/android-8/arch-arm/usr/lib目录,使得我们在下面使用Android AVD2.2的时候,可以加载到这个so文件。

5 撰写相应的Eclipse项目代码,由于在native.c文件中指明了项目的工程名词以及类名词还有函数名词,故我们的项目为com.churnlabs.ffmpegsample下面的MainActivity.java文件

package com.churnlabs.ffmpegsample;
 
import android.app.Activity;
import android.graphics.Bitmap;
import android.os.Bundle;
import android.view.View;
import android.view.View.OnClickListener;
import android.widget.Button;
import android.widget.ImageView;
 
public class MainActivity extends Activity {
    private static native void openFile();
    private static native void drawFrame(Bitmap bitmap);
    private static native void drawFrameAt(Bitmap bitmap, int secs);
 
    private Bitmap mBitmap;
    private int mSecs = 0;
 
    static {
        System.loadLibrary("ffmpegutils");
    }
 
    /** Called when the activity is first created. */
    @Override
    public void onCreate(Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);
        //setContentView(new VideoView(this));
        setContentView(R.layout.main);
 
        mBitmap = Bitmap.createBitmap(320, 240, Bitmap.Config.ARGB_8888);
        openFile();
 
        Button btn = (Button)findViewById(R.id.frame_adv);
        btn.setOnClickListener(new OnClickListener() {
            public void onClick(View v) {
                drawFrame(mBitmap);
                ImageView i = (ImageView)findViewById(R.id.frame);
                i.setImageBitmap(mBitmap);
            }
        });
 
        Button btn_fwd = (Button)findViewById(R.id.frame_fwd);
        btn_fwd.setOnClickListener(new OnClickListener() {
            public void onClick(View v) {
                mSecs += 5;
                drawFrameAt(mBitmap, mSecs);
                ImageView i = (ImageView)findViewById(R.id.frame);
                i.setImageBitmap(mBitmap);
            }
        });
 
        Button btn_back = (Button)findViewById(R.id.frame_back);
        btn_back.setOnClickListener(new OnClickListener() {
            public void onClick(View v) {
                mSecs -= 5;
                drawFrameAt(mBitmap, mSecs);
                ImageView i = (ImageView)findViewById(R.id.frame);
                i.setImageBitmap(mBitmap);
            }
        });
    }
}

6 编译运行即可,最终效果图

7 项目代码下载:

https://github.com/churnlabs/android-ffmpeg-sample/zipball/master

参考:
1 https://github.com/churnlabs/android-ffmpeg-sample
2 http://www.360doc.com/content/10/1216/17/474846_78726683.shtml
3 https://github.com/prajnashi






相关文章
|
24天前
|
算法 数据处理 开发者
FFmpeg库的使用与深度解析:解码音频流流程
FFmpeg库的使用与深度解析:解码音频流流程
35 0
|
25天前
|
存储 编解码 算法
探索FFmpeg复用:深入理解媒体数据的组织与封装(三)
探索FFmpeg复用:深入理解媒体数据的组织与封装
32 0
|
29天前
|
存储 编解码 算法
深度探索:使用FFmpeg实现视频Logo的添加与移除(二)
深度探索:使用FFmpeg实现视频Logo的添加与移除
39 0
|
1月前
|
编解码 监控 计算机视觉
QT5.14.2 视频分帧:QT与FFmpeg的高效结合
QT5.14.2 视频分帧:QT与FFmpeg的高效结合
|
1月前
|
编译器 开发工具 Android开发
Android 引入FFmpeg
Android 引入FFmpeg
20 0
|
25天前
|
存储 编解码 数据处理
【FFmpeg 视频基本格式】深入理解FFmpeg:从YUV到PCM,解码到编码(三)
【FFmpeg 视频基本格式】深入理解FFmpeg:从YUV到PCM,解码到编码
26 0
|
25天前
|
存储 编解码 数据处理
【FFmpeg 视频基本格式】深入理解FFmpeg:从YUV到PCM,解码到编码(二)
【FFmpeg 视频基本格式】深入理解FFmpeg:从YUV到PCM,解码到编码
33 0
|
29天前
|
存储 编解码 调度
剖析ffmpeg视频解码播放:时间戳的处理
剖析ffmpeg视频解码播放:时间戳的处理
42 0
|
30天前
|
存储 编解码 vr&ar
用C++实现视频编码器:FFmpeg与SDL技术结合,轻松编写高效编解码器
用C++实现视频编码器:FFmpeg与SDL技术结合,轻松编写高效编解码器
68 0
|
29天前
|
编解码 算法 vr&ar
深度剖析FFmpeg视频解码后的帧处理到Qt显示 从AVFrame到QImage的转换(二)
深度剖析FFmpeg视频解码后的帧处理到Qt显示 从AVFrame到QImage的转换
24 1