Video optimization for android - Part I
Download as DOC, PDF0 likes1,525 views
This document provides an overview of Android video optimization specifically for C/C++ developers. It outlines steps for setting up the development environment, using Android development tools and ADB, and working with camera and optimization techniques for ARM. Key tasks include compiling JNI code, managing YUV video files, and utilizing OpenCV for image processing.
Video optimization for android - Part I
- 1. Intro to Android Video Optimization Overview This short summary is intended for C/C++ developers which haven’t worked with Android before and would like to use Android for Signal, Image and Video processing Applications. Read ALL the instruction page before starting installations Environment Setup Goal: Understand how to use the Android development tools and Eclipse IDE. Task: Install the tools and run an SDK sample from this list. Choose a development OS (Windows / Linux /MAC) All builds and development should be done on this OS. Instructions for installation in http://developer.android.com/sdk/index.html download both SDK and NDK (include installation of Eclipse, ADT and Phone USB driver). Detailed instructions in the presentation. Alternatively try performing installation according to OpenCV instructions below. Working with ADB Goal: Understand how to work with the ADB. Task: Check the device is connected, pull and push files to/from the device and reset the device. ADB is the Android Debug Bridge, it assist developers in connecting and controlling the device. More on ADB commands in this wiki. 20 Giborey Israel, Sappir Industrial Park, Netanya, Israel | www.mobilevideotech.com e-mail:[email protected] | Telephone:+972.545.313092
- 2. Basic JNI compilation Goal: Understand JNI and how to work with native C++ code. Task: Compile and run a simple native application from the NDK samples, like the “Hello-JNI”. Building NDK samples requires separate compilation of the C/C++ part and the Java part. Understand the JNI mechanism. Working with Camera Goal: Understand how to work with the camera. Task: Compile Run and understand “Sample 0” Understand how to work with the camera try OpenCV samples for Android Read more at OpenCV Android Pages . OpenCV Android samples are run with C++ API while FastCV uses C API Optimization for ARM Goal: Understand optimization options for Android There are several methods to optimize code for ARM: 1. Using already optimized functions distributed by SoC vendors, for example: FastCV. More on FastCV Optimization in this slide. 2. Proprietary functions which don’t appear in such libraries require optimization for VFU / NEON. Task: earn more about NEON optimization review the “Hello-NEON” NDK sample. 20 Giborey Israel, Sappir Industrial Park, Netanya, Israel | www.mobilevideotech.com e-mail:[email protected] | Telephone:+972.545.313092
- 3. Working with SDCard Prepare a RAW YUV video which is used to test the algorithm. Copy the YUV input file to the SDCard using: adb push <FileName.YUV> </SDCard/Path> Get the output file from the device using: adb pull “<Path>/<RawYUVOutFileName.yuv>” . Use standard file operations to read frames from the file and pass them to image/video processing mechanism: FILE* MyInFile; FILE* MyOutFile; MyFile = fopen(“<Path>/<RawYUVInFileName.yuv>”, “rb”); MyFile = fopen(“<Path>/<RawYUVOutFileName.yuv>”, “wb”); Use standard fread / fwrite commands to manipulate the file 20 Giborey Israel, Sappir Industrial Park, Netanya, Israel | www.mobilevideotech.com e-mail:[email protected] | Telephone:+972.545.313092