Recherche avancée
Médias (1)
-
La conservation du net art au musée. Les stratégies à l’œuvre
26 mai 2011
Mis à jour : Juillet 2013
Langue : français
Type : Texte
Autres articles (97)
-
Multilang : améliorer l’interface pour les blocs multilingues
18 février 2011, parMultilang est un plugin supplémentaire qui n’est pas activé par défaut lors de l’initialisation de MediaSPIP.
Après son activation, une préconfiguration est mise en place automatiquement par MediaSPIP init permettant à la nouvelle fonctionnalité d’être automatiquement opérationnelle. Il n’est donc pas obligatoire de passer par une étape de configuration pour cela. -
Gestion des droits de création et d’édition des objets
8 février 2011, parPar défaut, beaucoup de fonctionnalités sont limitées aux administrateurs mais restent configurables indépendamment pour modifier leur statut minimal d’utilisation notamment : la rédaction de contenus sur le site modifiables dans la gestion des templates de formulaires ; l’ajout de notes aux articles ; l’ajout de légendes et d’annotations sur les images ;
-
Dépôt de média et thèmes par FTP
31 mai 2013, parL’outil MédiaSPIP traite aussi les média transférés par la voie FTP. Si vous préférez déposer par cette voie, récupérez les identifiants d’accès vers votre site MédiaSPIP et utilisez votre client FTP favori.
Vous trouverez dès le départ les dossiers suivants dans votre espace FTP : config/ : dossier de configuration du site IMG/ : dossier des média déjà traités et en ligne sur le site local/ : répertoire cache du site web themes/ : les thèmes ou les feuilles de style personnalisées tmp/ : dossier de travail (...)
Sur d’autres sites (9398)
-
Is there a way to eliminate seek time when decoding part of a video using ffmpeg ?
17 décembre 2019, par BabisI’ve got some MKV videos encoded with FFV1. For each of the frames, I want to run some complex and time-intensive python or matlab code, so I’m using multithreading, where each thread works on an individual image.
I’ve tried extracting a single frame from the video using -ss, but it’s terribly inefficient.
The most efficient way is to decompress everything into images in one go, but then I’m writing to disk, and then I’ll be reading from disk, therefore it’s not ideal either.
I’ve tried using a ram disk to export images to, and reading them from python/matlab, but it’s not great performance-wise either. Also, I have to split the export into several batches, as the video file is 20GB and all of the exported images will not fit into memory
Is there a way to rapidly extract individual frames from ffmpeg directly into RAM (or ram disk), so that they can be used by another program ? For example using something like a lookup-table.
For reference, each video is about 20GB, comprised of 50000 frames, and they are all keyframes (it’s for archival purposes)
-
FFMpeg C++ encoder fails with hardware acceleration quick sync intel
28 août 2022, par tealsI am trying to debug an issue I have with a simple C++ encoder using FFMmpeg. The following code below works correctly on these other hardware acceleration systems/platforms :


- 

- Mac/VideoToolbox,
- Cuda/NVEnc
- Raspberry pi/h264_v4l2m2m.








It fails on intel quick sync at
avcodec_open2. I'm using Ubuntu 22.04. I built FFMpeg from source with the Intel Media SDK installed and configured. I verified that the QuickSync install works. I also have a corresponding decoder implementation that works correctly using intel quick sync. So I know everything properly installed. I am also running this on an 11th Gen Intel CPU.

[hevc_qsv @ 0x564b0d34f0c0] Low power mode is unsupported (This sometimes shows up)
[hevc_qsv @ 0x563691234000] some encoding parameters are not supported by the QSV runtime. Please double check the input parameters.
Error initializing output stream 0:0 -- Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height


This is the only error that I get and I haven't found any helpful resources. I tried playing around with different parameters but nothing seems to work. Any help or insight would be extremely helpful.


bool VideoEncoder::create() {
 char const* outfile = filePath.c_str();

 // open output format context
 int ret = avformat_alloc_output_context2(&ofctx, nullptr, nullptr, outfile);
 if (ret < 0) {
 std::cerr << "fail to avformat_alloc_output_context2(" << outfile << "): ret=" << ret;
 return false;
 }

 // create new video stream
 const AVCodec* codec = avcodec_find_encoder_by_name(encoderName.c_str());
 vstrm = avformat_new_stream(ofctx, codec);
 if (!vstrm) {
 std::cerr << "fail to avformat_new_stream";
 return false;
 }

 // open video encoder
 cctx = avcodec_alloc_context3(codec);
 if (!vstrm) {
 std::cerr << "fail to avcodec_alloc_context3";
 return false;
 }
 const AVRational dst_fps = {fps, 1};
 cctx->width = width;
 cctx->height = height;
 if(pixel_format != AV_PIX_FMT_NONE) {
 cctx->pix_fmt = pixel_format;
 }
 else {
 cctx->pix_fmt = codec->pix_fmts[0];
 }
 cctx->time_base = av_inv_q(dst_fps);
 cctx->framerate = dst_fps;
 cctx->bit_rate = bitrate * 1000000;


 AVDictionary* options = nullptr;
 if(gpu_id >= 0) {
 av_dict_set_int(&options, "gpu", gpu_id, 0);
 }

 
 if (ofctx->oformat->flags & AVFMT_GLOBALHEADER)
 cctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
 ret = avcodec_open2(cctx, codec, &options);
 if (ret < 0) {
 std::cerr << "fail to avcodec_open2: ret=" << ret;
 return false;
 }
 avcodec_parameters_from_context(vstrm->codecpar, cctx);

 //For mac/quicktime we need to add hvc1 tag.
 if (encoderName.find("hevc") != std::string::npos) {
 vstrm->codecpar->codec_tag = MKTAG('h', 'v', 'c', '1');
 }

 /*
 std::cout
 << "outfile: " << outfile << "\n"
 << "format: " << ofctx->oformat->name << "\n"
 << "vcodec: " << codec->name << "\n"
 << "size: " << width << 'x' << height << "\n"
 << "fps: " << av_q2d(cctx->framerate) << "\n"
 << "pixfmt: " << av_get_pix_fmt_name(cctx->pix_fmt) << "\n"
 << std::flush;*/

 // initialize sample scaler
 swsCtx = sws_getContext(
 width, height, AV_PIX_FMT_BGR24,
 width, height, cctx->pix_fmt,
 SWS_BILINEAR, nullptr, nullptr, nullptr);
 if (!swsCtx) {
 std::cerr << "fail to sws_getContext";
 return false;
 }

 // allocate frame buffer for encoding
 frame = av_frame_alloc();
 frame->width = width;
 frame->height = height;
 frame->format = static_cast<int>(cctx->pix_fmt);
 ret = av_frame_get_buffer(frame, 32);
 if (ret < 0) {
 std::cerr << "fail to av_frame_get_buffer: ret=" << ret;
 return false;
 }

 // allocate packet to retrive encoded frame
 pkt = av_packet_alloc();
 // open output IO context
 ret = avio_open2(&ofctx->pb, outfile, AVIO_FLAG_WRITE, nullptr, nullptr);
 if (ret < 0) {
 std::cerr << "fail to avio_open2: ret=" << ret;
 return false;
 }

 // write media container header (if any)
 ret = avformat_write_header(ofctx, nullptr);
 if (ret < 0) {
 std::cerr << "fail to avformat_write_header: ret=" << ret;
 return false;
 }

 return true;
}
</int>

-
Trolls in trouble
6 juin 2013, par Mans — Law and libertyLife as a patent troll is hopefully set to get more difficult. In a memo describing patent trolls as a “drain on the American economy,” the White House this week outlined a number of steps it is taking to stem this evil tide. Chiming in, the Chief Judge of the Court of Appeals for the Federal Circuit (where patent cases are heard) in a New York Times op-ed laments the toll patent trolling is taking on the industry, and urges judges to use powers already at their disposal to make the practice less attractive. However, while certainly a step in the right direction, these measures all fail to address the more fundamental properties of the patent system allowing trolls to exist in the first place.
System and method for patent trolling
Most patent trolling operations comprise the same basic elements :
- One or more patents with broad claims.
- The patents of (1) acquired by an otherwise non-practising entity (troll).
- The entity of (2) filing numerous lawsuits alleging infringement of the patents of (1).
- The lawsuits of (3) targeting end users or retailers.
- The lawsuits of (3) listing as plaintiffs difficult to trace shell companies.
The recent legislative actions all take aim at the latter entries in this list. In so doing, they will no doubt cripple the trolls, but the trolls will remain alive, ready to resume their wicked ways once a new loophole is found in the system.
To kill a patent troll
As Judge Rader and his co-authors point out in the New York Times, “the problem stems largely from the fact that, [...] trolls have an important strategic advantage over their adversaries : they don’t make anything.” This is the heart of the troll, and this is where the blow should be struck. Our weapon shall be the mightiest judicial sword of all, the Constitution.
The United States Constitution contains (in Article I, Section 8) the foundation for the patent system (emphasis mine) :
The Congress shall have Power [...] To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.
Patent trolls are typically not inventors. They are merely hoarders of other people’s discarded inventions, and that allowing others to reap the benefits of an inventor’s work would somehow promote progress should be a tough argument. Indeed, it is the dissociation between investment and reward which has allowed the patent trolls to rise and prosper.
In light of the above, the solution to the troll menace is actually strikingly simple : make patents non-transferable.
Having the inventor retain the rights to his or her inventions (works for hire still being recognised), would render the establishment of non-practising entities, which most trolls are, virtually impossible. The original purpose of patents, to protect the investment of inventors, would remain unaffected, if not strengthened, by such a change.
Links
Syndiquer tout le site
