android - Compress Videos using FFMPEG and JNI

Question

I want to create an android application which can locate a video file (which is more than 300 mb) and compress it to lower size mp4 file.

i already tried to do it with this

This tutorial is a very effective since you 're compressing a small size video (below than 100 mb)

So i tried to implement it using JNI .

i managed to build ffmpeg using this

But currently what I want to do is to compress videos . I don't have very good knowledge on JNI. But i tried to understand it using following link

If some one can guide me the steps to compress video after open file it using JNI that whould really great , thanks

Answer 1

Assuming you've got the String path of the input file, we can accomplish your task fairly easily. I'll assume you have an understanding of the NDK basics: How to connect a native .c file to native methods in a corresponding .java file (Let me know if that's part of your question). Instead I'll focus on how to use FFmpeg within the context of Android / JNI.

High-Level Overview:

#include <jni.h>
#include <android/log.h>
#include <string.h>
#include "libavcodec/avcodec.h"
#include "libavformat/avformat.h"

#define LOG_TAG "FFmpegWrapper"
#define LOGI(...)  __android_log_print(ANDROID_LOG_INFO, LOG_TAG, __VA_ARGS__)
#define LOGE(...)  __android_log_print(ANDROID_LOG_ERROR, LOG_TAG, __VA_ARGS__)


void Java_com_example_yourapp_yourJavaClass_compressFile(JNIEnv *env, jobject obj, jstring jInputPath, jstring jInputFormat, jstring jOutputPath, jstring JOutputFormat){
  // One-time FFmpeg initialization
  av_register_all();
  avformat_network_init();
  avcodec_register_all();

  const char* inputPath = (*env)->GetStringUTFChars(env, jInputPath, NULL);
  const char* outputPath = (*env)->GetStringUTFChars(env, jOutputPath, NULL);
  // format names are hints. See available options on your host machine via $ ffmpeg -formats
  const char* inputFormat = (*env)->GetStringUTFChars(env, jInputFormat, NULL);
  const char* outputFormat = (*env)->GetStringUTFChars(env, jOutputFormat, NULL);

  AVFormatContext *outputFormatContext = avFormatContextForOutputPath(outputPath, outputFormat);
  AVFormatContext *inputFormatContext = avFormatContextForInputPath(inputPath, inputFormat /* not necessary since file can be inspected */);

  copyAVFormatContext(&outputFormatContext, &inputFormatContext);
  // Modify outputFormatContext->codec parameters per your liking
  // See http://ffmpeg.org/doxygen/trunk/structAVCodecContext.html

  int result = openFileForWriting(outputFormatContext, outputPath);
  if(result < 0){
      LOGE("openFileForWriting error: %d", result);
  }

  writeFileHeader(outputFormatContext);

  // Copy input to output frame by frame
  AVPacket *inputPacket;
  inputPacket = av_malloc(sizeof(AVPacket));

  int continueRecording = 1;
  int avReadResult = 0;
  int writeFrameResult = 0;
  int frameCount = 0;
  while(continueRecording == 1){
      avReadResult = av_read_frame(inputFormatContext, inputPacket);
      frameCount++;
      if(avReadResult != 0){
        if (avReadResult != AVERROR_EOF) {
            LOGE("av_read_frame error: %s", stringForAVErrorNumber(avReadResult));
        }else{
            LOGI("End of input file");
        }
        continueRecording = 0;
      }

      AVStream *outStream = outputFormatContext->streams[inputPacket->stream_index];
      writeFrameResult = av_interleaved_write_frame(outputFormatContext, inputPacket);
      if(writeFrameResult < 0){
          LOGE("av_interleaved_write_frame error: %s", stringForAVErrorNumber(avReadResult));
      }
  }

  // Finalize the output file
  int writeTrailerResult = writeFileTrailer(outputFormatContext);
  if(writeTrailerResult < 0){
      LOGE("av_write_trailer error: %s", stringForAVErrorNumber(writeTrailerResult));
  }
  LOGI("Wrote trailer");
}

For the full content of all the auxillary functions (the ones in camelCase), see my full project on Github. Got questions? I'm happy to elaborate.