Recherche avancée

Sur d’autres sites (11930)

• Rebuilding Website for sharing videos

  22 novembre 2015, par Léo Le Gall

  Some friends and I run a sport forum with a decent user base. A lot of users wanted the ability to share video clips of their tricks. We didn’t really think anyone would use a video website that we made, so I built a really simple one just to see if the users would really use it. I hosted the video site on a $10 VPS, and it got blown away. The site was literally garbage, it was not visually appealing and the performance was just sad. Just as expected really, since this was just a test site. Since our test project was a success, we want to create a new and more polished site for the videos.

  The website is really simple, and probably not optimized in any way at the moment. I will try to explain in details what the website does. The user uploads some video files in format X, the website combines them and converts (using ffmpeg) the final video to mp4(so it can be served with a html5 video player). The users gets a link (example.com/randomvideo) where he can see the video through a html5 player serving the mp4 file(just default html5, nothing fancy). The videos only contains highlights, and the final video is always under 1 minute, most are around 30 seconds.

  Currently everything happens on the same server, both the video processing and the video serving. I can try to show how it works :

  1. User uploads some videos
  2. Servers stores the videos in a new random folder
  3. Combine videos and convert to mp4 (ffmpeg)
  4. Move final video(random name) to directory containing processed videos
  5. Store the name of the video file in the database (for website to serve it)
  6. Delete directory used to process the videos

  I want to rebuild the website’s architecture to make it able to scale and handle heavy load. I have never done this before and I am currently making a plan how to do this. My plan is at the moment :

  1. Seperate servers for processing videos and viewing videos (all VPS in the start)
  2. Utilize content delivery network for serving static files
  3. Utilize load balancers both for servers proccesing and serving

  I don’t really know what I should do with database(s). Can I do with one database or should use more ? They do not need to store sensitive information, since the auth in done through an API. They only need store information about the videos. I have experience with postgresql, mysql and redis, but I am not limited to those. What would you recommend in terms of scaleability ?

  I will appriciate all the feedback I can get regarding my plan and what to do about databases. I know this might be a bit vague, so please ask me if I have forgotten anything imporant. Thanks for reading.

• FFMPEG and JNI - pass AVFrame data to Java and Back

  17 octobre 2015, par tishu

  I have some C code that decodes a video frame by frame. I get to a point where i have an AVFrame in BGR32 and would like to send it back to Java for editing.

  I have a ByteBuffer object in my C code that was created in Java using AllocateDirect but i struggle to write the content of the AVFrame->data[0] (of uint8_t type) to it and read it back. I have tried memcpy with no luck. Does anyone have an idea how to achieve this ?

  UPDATE
  Followed Will’s comment below and wrote this in C

  char *buf = (*pEnv)->GetDirectBufferAddress(pEnv, byteBuffer);
  
  memcpy(buf, rgb_frame->data[0], output_width*output_height*4);

  The buffer does contain some data in Java but doing the following returns a null bitmap

  BufferedImage frame = ImageIO.read(bitmapStream);

  Where bitmapStream is a ByteBufferInputStream defined here :
  https://code.google.com/p/kryo/source/browse/trunk/src/com/esotericsoftware/kryo/io/ByteBufferInputStream.java?r=205

  Not sure if I am not writing things correctly in this buffer

  UPDATE 2

  Got pretty close now thanks to the latest snippet. I am using BGR32 format in my C code ie 4 bytes per pixel. So I modified things a bit in Java :

  final byte[] dataArray = new byte[width*height*4];
  
  imageData.get(dataArray);
  
  final BufferedImage image = new BufferedImage(width, height, BufferedImage.TYPE_INT_BGR);
  
  final DataBuffer buffer = new DataBufferByte(dataArray, dataArray.length);
  
  Raster raster = Raster.createRaster(sampleModel, buffer, null);
  
  image.setData(raster);

  I get the image correctly but there seems to be an issue with color channels
  

  Tried different formats with no luck

• delay in lyrics shown when concat video with lyrics with an intro video using ffmpeg

  19 mars 2018, par no name

  I have a video with lyrics I add the lyrics used ffmpeg too. I have the lyrics in lrc formate and i change it to ass lyrics formate then I start to add it to my video using this command

  ffmpeg -i video.mp4 -vf "ass=out.ass,fade=in:0:30" -y amr.mp4

  I use fade=in:0:30 to make fade in the start of the video so when I start merge intro video my original video not start at once but to have some fade in, in the start to have a good view.
  then I start to add the intro video using a text file which has this

  file intro.mp4
  
  file amr.mp4

  then I merge them using this command

  ffmpeg -f concat -safe 0 -i out.txt -c copy -y final.mp4

  i got these message

  [mov,mp4,m4a,3gp,3g2,mj2 @ 03eadb80] Auto-inserting h264_mp4toannexb bitstream filter
  
  [mp4 @ 06419580] Non-monotonous DTS in output stream 0:1; previous: 310272,     current: 285626; changing to 310273. This may result in incorrect timestamps in   the output file.

  repeated many times.
  when i watch the final video the lyrics have some delay from the sound the problem appeared when I merge the intro video but without it the lyrics have no problem so what I made wrong make the lyrics shown in this way and not in the right timing as it should have no problems as I merge only the intro video and after add the lyrics to the original one
  Thanks in advance