Recherche avancée
Médias (1)
-
Rennes Emotion Map 2010-11
19 octobre 2011, par
Mis à jour : Juillet 2013
Langue : français
Type : Texte
Autres articles (18)
-
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. -
Librairies et logiciels spécifiques aux médias
10 décembre 2010, parPour un fonctionnement correct et optimal, plusieurs choses sont à prendre en considération.
Il est important, après avoir installé apache2, mysql et php5, d’installer d’autres logiciels nécessaires dont les installations sont décrites dans les liens afférants. Un ensemble de librairies multimedias (x264, libtheora, libvpx) utilisées pour l’encodage et le décodage des vidéos et sons afin de supporter le plus grand nombre de fichiers possibles. Cf. : ce tutoriel ; FFMpeg avec le maximum de décodeurs et (...) -
Les autorisations surchargées par les plugins
27 avril 2010, parMediaspip core
autoriser_auteur_modifier() afin que les visiteurs soient capables de modifier leurs informations sur la page d’auteurs
Sur d’autres sites (4643)
-
x264 encoding severe quality loss
1er septembre 2020, par SolskGaerI used this repo to encode a mjpeg stream to a h264 one, but the output is not so good. The stream is a series of screenshot of an iPhone. In the output stream, even the lines between two items in the settings app are gone. How do I improve the output stream quality ?
Here is the code snippet that x264-go use to init an encoder


func NewEncoder(w io.Writer, opts *Options) (e *Encoder, err error) {
 e = &Encoder{}

 e.w = w
 e.pts = 0
 e.opts = opts

 e.csp = x264c.CspI420

 e.nals = make([]*x264c.Nal, 3)
 e.img = NewYCbCr(image.Rect(0, 0, e.opts.Width, e.opts.Height))

 param := x264c.Param{}

 if e.opts.Preset != "" && e.opts.Profile != "" {
 ret := x264c.ParamDefaultPreset(&param, e.opts.Preset, e.opts.Tune)
 if ret < 0 {
 err = fmt.Errorf("x264: invalid preset/tune name")
 return
 }
 } else {
 x264c.ParamDefault(&param)
 }

 param.IWidth = int32(e.opts.Width)
 param.IHeight = int32(e.opts.Height)

 param.ICsp = e.csp
 param.BVfrInput = 0
 param.BRepeatHeaders = 1
 param.BAnnexb = 1

 param.ILogLevel = e.opts.LogLevel

 if e.opts.FrameRate > 0 {
 param.IFpsNum = uint32(e.opts.FrameRate)
 param.IFpsDen = 1

 param.IKeyintMax = int32(e.opts.FrameRate)
 param.BIntraRefresh = 1
 }

 if e.opts.Profile != "" {
 ret := x264c.ParamApplyProfile(&param, e.opts.Profile)
 if ret < 0 {
 err = fmt.Errorf("x264: invalid profile name")
 return
 }
 }

 // Allocate on create instead while encoding
 var picIn x264c.Picture
 ret := x264c.PictureAlloc(&picIn, e.csp, int32(e.opts.Width), int32(e.opts.Height))
 if ret < 0 {
 err = fmt.Errorf("x264: cannot allocate picture")
 return
 }
 e.picIn = picIn
 defer func() {
 // Cleanup if intialization fail
 if err != nil {
 x264c.PictureClean(&picIn)
 }
 }()

 e.e = x264c.EncoderOpen(&param)
 if e.e == nil {
 err = fmt.Errorf("x264: cannot open the encoder")
 return
 }

 ret = x264c.EncoderHeaders(e.e, e.nals, &e.nnals)
 if ret < 0 {
 err = fmt.Errorf("x264: cannot encode headers")
 return
 }

 if ret > 0 {
 b := C.GoBytes(e.nals[0].PPayload, C.int(ret))
 n, er := e.w.Write(b)
 if er != nil {
 err = er
 return
 }

 if int(ret) != n {
 err = fmt.Errorf("x264: error writing headers, size=%d, n=%d", ret, n)
 }
 }

 return
}


and the encoder option is defined as


opts := &x264.Options{
 Width: int(width)/2*2,
 Height: int(height)/2*2,
 FrameRate: 15,
 Tune: "zerolatency",
 Preset: "medium",
 Profile: "baseline",
 LogLevel: x264.LogNone,
 }


And I also used ffmpeg(though the api is pretty old)


package screencast

import (

 /*
 #include 
 #include 
 #include 
 #include 
 #include <libavcodec></libavcodec>avcodec.h>
 #include <libavformat></libavformat>avformat.h>
 #include <libavutil></libavutil>avutil.h>
 #include <libavutil></libavutil>opt.h>
 #include <libavutil></libavutil>channel_layout.h>
 #include <libavutil></libavutil>common.h>
 #include <libavutil></libavutil>imgutils.h>
 #include <libavutil></libavutil>mathematics.h>
 #include <libavutil></libavutil>samplefmt.h>

 typedef struct {
 int w, h;
 int pixfmt;
 char *preset[2];
 char *profile;
 int bitrate;
 int got;
 AVCodec *c;
 AVCodecContext *ctx;
 AVFrame *f;
 AVPacket pkt;
 } h264enc_t;

 static int h264enc_new(h264enc_t *m) {
 m->c = avcodec_find_encoder(AV_CODEC_ID_H264);
 m->ctx = avcodec_alloc_context3(m->c);
 m->ctx->width = m->w;
 m->ctx->height = m->h;
 m->ctx->pix_fmt = m->pixfmt;
 m->ctx->time_base = (AVRational){1,10};
 av_opt_set(m->ctx->priv_data, "preset", "slow", 0);
 av_opt_set(m->ctx->priv_data, "tune", "zerolatency", 0);
 av_opt_set(m->ctx->priv_data, "profile", "baseline", 0);
 av_opt_set(m->ctx->priv_data, "crf", "18.0.", 0);
 m->f = av_frame_alloc();
 m->f->format = m->ctx->pix_fmt;
 m->f->width = m->ctx->width;
 m->f->height = m->ctx->height;
 avcodec_open2(m->ctx, m->c, NULL);
 return av_image_alloc(m->f->data, m->f->linesize, m->ctx->width, m->ctx->height, m->ctx->pix_fmt, 32);
 }

 */
 "C"
 "errors"
 "image"
 "unsafe"
 //"log"
)

type H264Encoder struct {
 m C.h264enc_t
 Header []byte
 Pixfmt image.YCbCrSubsampleRatio
 W, H int
 pts int
}

func NewH264Encoder(w, h int) (m *H264Encoder, err error) {
 m = &H264Encoder{}
 m.m.w = (C.int)(w)
 m.m.h = (C.int)(h)
 m.W = w
 m.H = h
 m.Pixfmt = image.YCbCrSubsampleRatio420
 m.m.pixfmt = C.AV_PIX_FMT_YUV420P
 r := C.h264enc_new(&m.m)
 if int(r) < 0 {
 err = errors.New("open encoder failed")
 return
 }
 return
}

func (m *H264Encoder) Encode(img *image.YCbCr) (data []byte, err error) {
 var f *C.AVFrame
 if img == nil {
 f = nil
 } else {
 if img.SubsampleRatio != m.Pixfmt {
 err = errors.New("image pixfmt not match")
 return
 }
 if img.Rect.Dx() != m.W || img.Rect.Dy() != m.H {
 err = errors.New("image size not match")
 return
 }
 f = m.m.f
 f.data[0] = (*C.uint8_t)(unsafe.Pointer(&img.Y[0]))
 f.data[1] = (*C.uint8_t)(unsafe.Pointer(&img.Cb[0]))
 f.data[2] = (*C.uint8_t)(unsafe.Pointer(&img.Cr[0]))
 f.linesize[0] = (C.int)(img.YStride)
 f.linesize[1] = (C.int)(img.CStride)
 f.linesize[2] = (C.int)(img.CStride)
 }

 C.av_init_packet(&m.m.pkt)
 m.m.pkt.data = nil
 m.m.pkt.size = 0
 f.pts = (C.longlong)(m.pts)
 m.pts++
 r := C.avcodec_encode_video2(m.m.ctx, &m.m.pkt, f, &m.m.got)
 defer C.av_packet_unref(&m.m.pkt)
 if int(r) < 0 {
 err = errors.New("encode failed")
 return
 }
 if m.m.got == 0 {
 err = errors.New("no picture")
 return
 }
 if m.m.pkt.size == 0 {
 err = errors.New("packet size == 0")
 return
 }

 data = make([]byte, m.m.pkt.size)
 C.memcpy(
 unsafe.Pointer(&data[0]),
 unsafe.Pointer(m.m.pkt.data),
 (C.size_t)(m.m.pkt.size),
 )
 return data, nil
}


but got the same output. However, when I use the ffmpeg binary, the result was pretty good, so I guess I set wrong parameters, but I don't know which.
Any suggestion would be appreciated. If you have better way to achieve this, I'll appreciate it.


For your information : I must do this using golang.




-
ffmpeg fix video orientation
22 juillet 2020, par PiTheNumberA video can contain a meta info about the camera orientation. For example iPhone and other phones set this flag if you turn the device. Problem is while some player read this info and rotate the video accordingly, other players do not.



To fix this the video has to be rotated and the meta info needs to be set correctly.



Does ffmpeg provide a fix for this or do I have to go the hard way (Read rotation, rotate, set meta data)


-
ffmpeg - Remove slow motion from MOV
16 juillet 2020, par KalemKutuI recorded a short movie with an iPhone.
No idea how I did, but a part of the video is in slow motion.


I want to play but also store the video on my computer MAC in "normal" speed,
if possible without losing quality & without encoding.


I tried following parameters in ffmpeg, which works.
But I have the feeling the outcome quality has reduced "little-bit".


ffmpeg -i input.MOV -r 30 output.MOV


Is there a better way ?


Thanks, Kalem


Syndiquer tout le site
