Recherche avancée

Sur d’autres sites (9466)

• Building a shared library from static libraries for ffmpeg 2.5.2

  7 janvier 2015, par abijinx

  I am currently building a shared library for ffmpeg as myffmpeg.so using ffmpeg 0.8.6. I am achieving by combining static libs of individual modules of ffmpeg, along with a few additional static libraries. The main makefile used is as follows :

  LOCAL_PATH := $(call my-dir)
  
  
  
  include $(LOCAL_PATH)/Android_Files.mk
  
  
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE:= libavcodec
  
  LOCAL_SRC_FILES:= $(MY_AVCODEC_FILES)
  
  include $(LOCAL_PATH)/Android_Common.mk
  
  include $(BUILD_STATIC_LIBRARY)
  
  
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE:= libavfilter
  
  LOCAL_SRC_FILES:= $(MY_AVFILTER_FILES)
  
  include $(LOCAL_PATH)/Android_Common.mk
  
  include $(BUILD_STATIC_LIBRARY)
  
  
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE:= libavformat
  
  LOCAL_SRC_FILES:= $(MY_AVFORMAT_FILES)
  
  include $(LOCAL_PATH)/Android_Common.mk
  
  include $(BUILD_STATIC_LIBRARY)
  
  
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE:= libavutil
  
  LOCAL_SRC_FILES:= $(MY_AVUTIL_FILES)
  
  include $(LOCAL_PATH)/Android_Common.mk
  
  include $(BUILD_STATIC_LIBRARY)
  
  
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE:= libswscale
  
  LOCAL_SRC_FILES:= $(MY_SWSCALE_FILES)
  
  include $(LOCAL_PATH)/Android_Common.mk
  
  include $(BUILD_STATIC_LIBRARY)
  
  
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE:= ffmpeg_myc
  
  LOCAL_SRC_FILES:= ffmpeg.c cmdutils.c
  
  include $(LOCAL_PATH)/Android_Common.mk
  
  include $(BUILD_STATIC_LIBRARY)
  
  
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE := myffmpeg
  
  LOCAL_WHOLE_STATIC_LIBRARIES := libavcodec libavfilter libavformat libavutil libswscale ffmpeg_c
  
  LOCAL_LDFLAGS += -lz -lm -llog
  
  ifeq ($(FF_ENABLE_AMR),yes)
  
  LOCAL_WHOLE_STATIC_LIBRARIES += opencore-amrnb opencore-amrwb
  
  endif
  
  ifeq ($(FF_ENABLE_AAC),yes)
  
  LOCAL_WHOLE_STATIC_LIBRARIES += vo-aacenc
  
  endif
  
  
  
  include $(BUILD_SHARED_LIBRARY)

  The above makefile uses only few selected source files to get the individual static libs as mentioned in the FF_SOURCE_FILES parameter.Then combines all of those with the additional libs to create a shared library myffmpeg.so.

  Now I am trying to get a similar output with ffmpeg 2.5.2 for 64-bit arm architecture. This time I have built the latest ffmpeg using the following configure script and generated the static libraries of different ffmpeg modules like libavcodec.a, libavfilter.a, etc

  #!/bin/bash
  
  ABI=aarch64-linux-android
  
  NDK=
  
  SYSROOT=$NDK/platforms/android-21/arch-arm64/
  
  TOOLCHAIN=$NDK/toolchains/aarch64-linux-android-4.9/prebuilt/linux-x86_64
  
  CPU=arm64
  
  PREFIX=$(pwd)/android/$CPU
  
  ./configure \
  
  --prefix=$PREFIX \
  
  --cc=$TOOLCHAIN/bin/$ABI-gcc \
  
  --enable-static \
  
  --disable-doc \
  
  --disable-ffmpeg \
  
  --disable-ffplay \
  
  --disable-ffprobe \
  
  --disable-network \
  
  --disable-ffserver \
  
  --disable-devices \
  
  --disable-avdevice \
  
  --disable-swscale-alpha \
  
  --disable-doc \
  
  --disable-symver \
  
  --disable-neon \
  
  --enable-optimizations \
  
  --cross-prefix=$TOOLCHAIN/bin/aarch64-linux-android- \
  
  --target-os=linux \
  
  --arch=arm64 \
  
  --enable-cross-compile \
  
  --sysroot=$SYSROOT 
  
  
  
      --enable-libopencore-amrnb \
  
      --enable-libopencore-amrwb \
  
      --enable-libvo-aacenc \
  
      $ADDITIONAL_CONFIGURE_FLAG
  
  make clean
  
  make
  
  make install 

  Now I am trying to combine the generated static libraries i.e

  • libavcodec.a
    libavfilter.a
    libavformat.a
    libavutil.a
    libswresample.a
    libswscale.a
    with

  • ffmpeg_myc opencore-amrnb opencore-amrwb

  and generate a myffmpeg.so shared library by giving ndk-build.

  I tried to paste the generated ffmpeg static libraries from ffmpeg/android folder to obj/local/arm64-v8a (path where are all the static libs were previously created in 0.8.6 according the makefile). I also made changes to the makefile thinking that it will use the already generated ffmpeg libraries. But I found that the size of the generated shared library was too low and that it has not included the generated ffmpeg libs.

  LOCAL_PATH := $(call my-dir)
  
  include $(CLEAR_VARS)
  
  LOCAL_MODULE := myffmpeg
  
  LOCAL_WHOLE_STATIC_LIBRARIES := libavcodec libavfilter libavformat libavutil libswscale ffmpeg_c
  
  LOCAL_LDFLAGS += -lz -lm -llog
  
  ifeq ($(FF_ENABLE_AMR),yes)
  
  LOCAL_WHOLE_STATIC_LIBRARIES += opencore-amrnb opencore-amrwb
  
  endif
  
  ifeq ($(FF_ENABLE_AAC),yes)
  
  LOCAL_WHOLE_STATIC_LIBRARIES += vo-aacenc
  
  endif
  
  
  
  include $(BUILD_SHARED_LIBRARY)

  I would like to know whether I am going in a right way and have to do some minor changes. Or if I should start with an entirely different approach to achieve this. Any suggestions would be helpful.

  Thanks in advance

• FFmpeg corrupted output from concatenation

  25 juillet 2021, par Manumit

  EDIT/UPDATE : current solution is to run this code on all files ffmpeg -i up.mp4 -vf scale=1920:1080 -crf 22 reUP.mp4 so resolution, but also frames per second match.

  


  

  


  I used to be able to open CMD, type CD C:\Users\... and then ffmpeg -f concat -safe 0 -i xmylist.txt -crf 22 -c copy x1.mp4

  


  

  


  The recording are from same phone, some from front cam, some from back cam. I understand this could cause non-identical problems ? but it worked fine until recently.

  


  Now the output is corrupted with this log : 
The individual video files play just fine, so does most of the concatted output, but some sections become like this and freeze :

  


• Fastest way to extract moving dynamic crop from video using ffmpeg

  5 avril 2024, par NateW

  I'm working on an AI project that involves object detection and action recognition on roughly 30 minute videos.

  


  My pipeline is the following :

  


  

  • determine crops using object detection model

  


  • extract crops using Python and write them to disk as individual images for each frame.

  


  • use action recognition model by inputting a series of the crops.

  


  


  The models are fast but actual writing of the crops to disk is slow. Sure, using an SSD would speed it up but I'm sure ffmpeg would greatly speed it up.

  


  Some of the challenges with the crops :

  


  

  • the output size is always 128x128

  


  • the input size is variable

  


  • the crop moves on every frame

  


  


  My process for extracting crops is simple using cv2.imwrite(output_crop_path, crop) in a for loop.

  


  I've done experiments trying to use sndcommand and filter_complex. I tried this https://stackoverflow.com/a/67508233/4447761 but it outputs an image with black below the crop and the image gets wrapped around on the x axis.