Recherche avancée

Sur d’autres sites (9041)

• FFmpeg sws_scale on changed area

  1er octobre 2016, par useprxf

  I was using sws_scale to convert a group of RGB32 images to YUV420 format. Each image is very similar to the previous one and they only differ on a rectangle region Q.

  My question is how to utilize Q to speed up the conversion process ? An additional parameter should be added to sws_scale function.

  sws_scale( ctx, in_plane, in_stride, sliceY, height, out_plane, out_stride, Q){
  
      // parameter out_plane stores the YUV420 data of previous image
  
      Instead of scanning the whole image, scan through rectangle Q{
  
          Do conversion
  
      }
  
  }

• Struggling getting FFMPEG to work on Lambda function (Serverless Framework)

  3 juin 2022, par Red Vic

  I've been trying to get FFMPEG to work on my serverless framework for hours and I can't get to grasp how all this should work.

  


  This is my JavaScript code in the handler (just for testing purposes) :

  


  require("dotenv").config();

const ffmpegSync = (url) => {
  const ffmpeg = require("fluent-ffmpeg");
  ffmpeg.setFfprobePath("/opt/ffmpeg-layer/ffprobe");
  ffmpeg.setFfmpegPath("/opt/ffmpeg-layer/ffmpeg")

  return new Promise((resolve, reject) => {
    ffmpeg.ffprobe(url, async (err, metadata) => {
      if (err) {
        resolve(err);
      }
      console.log(metadata);
      resolve(metadata);
    });
  });
};

const ffmpegTeller = async (e, context) => {
  const AWS = require("aws-sdk");

  const s3 = new AWS.S3();
  AWS.config.update({
    accessKeyId: process.env.AWS_ACCESS_KEY,
    secretAccessKey: process.env.AWS_SECRET_KEY,
    region: "us-east-1",
  });

  const sourceBucket = e["Records"][0]["s3"]["bucket"]["name"];
  const sourceKey = e["Records"][0]["s3"]["object"]["key"];
  const signedUrlExpireSeconds = 60 * 5;

  const url = s3.getSignedUrl("getObject", {
    Bucket: sourceBucket,
    Key: sourceKey,
    Expires: signedUrlExpireSeconds,
  });

  await ffmpegSync(url).then((data) => {
    console.log(data);
    return {
      body: JSON.stringify(data),
      statusCode: 200,
    };
  });
};

module.exports = {
  ffmpegTeller,
};


  


  serverless.yaml :

  


  service: my-ffmpeg-api
frameworkVersion: '3'

provider:
  name: aws
  runtime: nodejs14.x
  region: us-east-1

functions:
  ffmpegTeller:
    handler: handler.ffmpegTeller
    events: 
      - s3:
          bucket: sourceBucket
          event: s3:ObjectCreated:*


  


  package.json

  


  {
  "dependencies": {
    "dotenv": "^16.0.1",
    "fluent-ffmpeg": "^2.1.2"
  }
}


  


  The error I'm getting on CloudWatch when trying to use FFprobe


  


  This is the view on the Lambda function page on AWS :

  


  

  


  I'd really appreciate some help. I'm really going mad. Thanks in advance.

  


• Optimize x264 based remote desktop by dirty regions

  23 novembre 2016, par useprxf

  I was using x264 to achieve remote desktop, but had some problems on handling P_SKIP detection.

  Dirty regions indicate changed areas. I would like to encode those macroblocks which don’t intersect any dirty region as P_SKIP types.

  I inserted the following code into x264_macroblock_prob_skip_internal function :

  if (! h->isdirty[h->mb.i_mb_x][h->mb.i_mb_y] && ! M32(h->mb.cache.pskip_mv))
  
      return 1;

  but there is almost no speed-up. I think it may be the information preparation for the macroblock analysis that takes influence.

  My question is how to speed up x264 by considering dirty regions ?