Recherche avancée

Sur d’autres sites (2406)

• x86 : Serialize rdtsc in read_time()

  8 juillet 2015, par Henrik Gramner

  x86 : Serialize rdtsc in read_time()
  Improves the accuracy of measurements, especially in short sections.
  To quote the Intel 64 and IA-32 Architectures Software Developer’s Manual :
  
  "The RDTSC instruction is not a serializing instruction. It does not necessarily
  
  wait until all previous instructions have been executed before reading the counter.
  
  Similarly, subsequent instructions may begin execution before the read operation
  
  is performed. If software requires RDTSC to be executed only after all previous
  
  instructions have completed locally, it can either use RDTSCP (if the processor
  
  supports that instruction) or execute the sequence LFENCE ;RDTSC."
  SSE2 is a requirement for lfence so only use it on SSE2-capable systems.
  
  Prefer lfence ;rdtsc over rdtscp since rdtscp is supported on fewer systems.
  Signed-off-by : Luca Barbato <lu_zero@gentoo.org>
  

  • [DBH] libavutil/x86/timer.h

• Trimming and batch converting arbitrary video files to WebMs with python and ffmpeg

  22 juin 2018, par Romtromon

  I’ve got a whole lot of video files that I want to break up into multiple WebMs, each containing a trimmed portion of a specific video.

  I plan on having a csv file attached to each video file (having the same filename as the video) with a column structure similar to :

  Start Time | End Time | Rotation | Output Filename

  And I want to parse the csv files using python to execute ffmpeg, and I did a quick search and found the ffmpeg-python library and thought it might do the trick. The problem is, I don’t know the first thing about ffmpeg or video encoding. I’ve tried reading through the ffmpeg documentation and trying to replicate stuff using ffmpeg-python but the furthest I can get is an output video of the same format as the input file, trimmed (but mpv still shows the duration as the duration of the original file but cuts off playback when the end of the trim is reached) and without audio.

  As a side note, I currently use a software named "WebM for Retards" (excuse the offensive title) which uses ffmpeg and I’m happy with its output but it’s very tedious for my requirement. However I noticed that these are the arguments passed by using the software :

  -f nut -i pipe:0   -c:v libvpx -pix_fmt yuv420p -threads 8 -slices 4 -metadata title="This is a title" -ac 2 -c:a libvorbis -qmin 28 -crf 30 -qmax 32 -qcomp 1 -b:v 0 -qscale:a 3 -f webm -y "C:\Output.webm"

  I’d be happy if I could replicate output similar to what this provides. Thanks in advance for any help !

• Best practice to play segments of larger video files in HTML5 video format

  8 octobre 2015, par zx98

  I have to migrate a legacy project from Windows Media to HTML5 video/audio. The system should provide video/audio content to the web user meeting following criteria :

  1. Playout : Segments (TcIn to TcOut) of large video/audio files should be played "on the fly" to the web user. (Cutting the source material into segments does not work because there is a high overlap rate between segments, so 3-4 times more HDD space and a lot of extra programming would be needed. Markered playout of the large source files is not an option either as there are forbidden parts of the source material which should never be available to the user).
  2. Cutting : In general, users are permitted to make copies of the material. Indeed, by using the dowload function, users can download selected segments. I guess, this is not a problem : a service starts a cutting software (probably ffmpeg) that cuts the video, and the user is notified by email to download the product. (That’s the way I do it now but there might be a more simple option).
  3. Formats : what is the preferred video/audio format that most browsers (Chrome, Firefox, Opera, Android) support ?
  4. Server : If possible, I would keep Windows Server because the database behind is MS SQL Server.

  (Windows Media Server and WMEncoder currently works fine for all the above, however WM is not suppored by browsers any more. That is the reason for reengineering).

  I would appreciate any of your suggestions on how to rebuild this video/audio system. (What software elements to use and how).

  Thank you.

  -------- addition -------
  Dear commenters,
  Thank you for your comments. Please note that the above is one single question, the rest is just a detailed description of the issue/conditions, as recommended by stackoverflow.com. I guess, the answer is also very simple, I try to write it myself :

  1. You should keep the large video/audio files in xxx format because...
  2. You create a playlist file like this example... for each of the the segments.
  3. You install the xxx streaming software component on the web/media server to playout the segments.
  4. You can address and play the segments from the client side web browser like this :...
  5. For asynchronous cutting, yes, I recommend ffmpeg, try these profile settings.

  Just a few parts in bold are missing. Thanks to anyone who can fill them.