mirror of
https://github.com/nzp-team/fteqw.git
synced 2024-11-29 07:02:12 +00:00
1bb752b582
hacky rendertarget stuff. not polished. don't use except for testing. feedback desired. switched file system to use a qofs_t type instead. define FS_64BIT to make it 64bit (standard on 64bit cpus). rewrote zip support, ditching unzip.c. this provided zip64 support, and unicode in zips. changed local address enumeration to not be so stupid. updated ode support a little to match some dp features. changed fs_cache scheme, to not rebuild needlessly. git-svn-id: https://svn.code.sf.net/p/fteqw/code/trunk@4596 fc73d0e0-1445-4013-8a0c-d673dee63da5
470 lines
12 KiB
C
470 lines
12 KiB
C
#include "../plugin.h"
|
|
#include "../engine.h"
|
|
|
|
#include "libavcodec/avcodec.h"
|
|
#include "libavformat/avformat.h"
|
|
#include "libswscale/swscale.h"
|
|
|
|
//between av 52.31 and 54.35, lots of constants etc got renamed to gain an extra AV_ prefix.
|
|
/*
|
|
#define AV_PIX_FMT_BGRA PIX_FMT_BGRA
|
|
#define AVMEDIA_TYPE_AUDIO CODEC_TYPE_AUDIO
|
|
#define AVMEDIA_TYPE_VIDEO CODEC_TYPE_VIDEO
|
|
#define AV_PIX_FMT_BGRA PIX_FMT_BGRA
|
|
#define AV_SAMPLE_FMT_U8 SAMPLE_FMT_U8
|
|
#define AV_SAMPLE_FMT_S16 SAMPLE_FMT_S16
|
|
#define AV_SAMPLE_FMT_FLT SAMPLE_FMT_FLT
|
|
#define AVIOContext ByteIOContext
|
|
#define avio_alloc_context av_alloc_put_byte
|
|
*/
|
|
|
|
#define PASSFLOAT(f) *(int*)&(f)
|
|
|
|
#define ARGNAMES ,sourceid, data, speed, samples, channels, width, PASSFLOAT(volume)
|
|
BUILTIN(void, S_RawAudio, (int sourceid, void *data, int speed, int samples, int channels, int width, float volume));
|
|
#undef ARGNAMES
|
|
|
|
/*should probably try threading this*/
|
|
/*timing is based upon the start time. this means overflow issues with rtsp etc*/
|
|
|
|
struct decctx
|
|
{
|
|
unsigned int width, height;
|
|
|
|
qhandle_t file;
|
|
int64_t fileofs;
|
|
int64_t filelen;
|
|
AVFormatContext *pFormatCtx;
|
|
|
|
int audioStream;
|
|
AVCodecContext *pACodecCtx;
|
|
AVFrame *pAFrame;
|
|
|
|
int videoStream;
|
|
AVCodecContext *pVCodecCtx;
|
|
AVFrame *pVFrame;
|
|
int64_t num, denum;
|
|
|
|
AVPicture pFrameRGB;
|
|
struct SwsContext *pScaleCtx;
|
|
|
|
unsigned int starttime;
|
|
unsigned int lastframe;
|
|
};
|
|
|
|
static qboolean AVDec_SetSize (void *vctx, int width, int height)
|
|
{
|
|
struct decctx *ctx = (struct decctx*)vctx;
|
|
AVPicture newscaled;
|
|
|
|
//colourspace conversions will be fastest if we
|
|
// if (width > ctx->pCodecCtx->width)
|
|
width = ctx->pVCodecCtx->width;
|
|
// if (height > ctx->pCodecCtx->height)
|
|
height = ctx->pVCodecCtx->height;
|
|
|
|
//is this a no-op?
|
|
if (width == ctx->width && height == ctx->height && ctx->pScaleCtx)
|
|
return true;
|
|
|
|
if (avpicture_alloc(&newscaled, AV_PIX_FMT_BGRA, width, height) >= 0)
|
|
{
|
|
//update the scale context as required
|
|
//clear the old stuff out
|
|
avpicture_free(&ctx->pFrameRGB);
|
|
|
|
ctx->width = width;
|
|
ctx->height = height;
|
|
ctx->pFrameRGB = newscaled;
|
|
return qtrue;
|
|
}
|
|
return qfalse; //unsupported
|
|
}
|
|
|
|
static int AVIO_Read(void *opaque, uint8_t *buf, int buf_size)
|
|
{
|
|
struct decctx *ctx = opaque;
|
|
int ammount;
|
|
ammount = pFS_Read(ctx->file, buf, buf_size);
|
|
if (ammount > 0)
|
|
ctx->fileofs += ammount;
|
|
return ammount;
|
|
}
|
|
static int64_t AVIO_Seek(void *opaque, int64_t offset, int whence)
|
|
{
|
|
struct decctx *ctx = opaque;
|
|
int64_t ret = ctx->fileofs;
|
|
whence &= ~AVSEEK_FORCE;
|
|
switch(whence)
|
|
{
|
|
case SEEK_SET:
|
|
default:
|
|
ctx->fileofs = offset;
|
|
break;
|
|
case SEEK_CUR:
|
|
ctx->fileofs += offset;
|
|
break;
|
|
case SEEK_END:
|
|
ctx->fileofs = ctx->filelen + offset;
|
|
break;
|
|
case AVSEEK_SIZE:
|
|
return ctx->filelen;
|
|
}
|
|
pFS_Seek(ctx->file, ctx->fileofs & 0xffffffff, ctx->fileofs>>32);
|
|
return ctx->fileofs;
|
|
}
|
|
|
|
static void AVDec_Destroy(void *vctx)
|
|
{
|
|
struct decctx *ctx = (struct decctx*)vctx;
|
|
|
|
// Free the video stuff
|
|
avpicture_free(&ctx->pFrameRGB);
|
|
if (ctx->pVCodecCtx)
|
|
avcodec_close(ctx->pVCodecCtx);
|
|
av_free(ctx->pVFrame);
|
|
|
|
// Free the audio decoder
|
|
if (ctx->pACodecCtx)
|
|
avcodec_close(ctx->pACodecCtx);
|
|
av_free(ctx->pAFrame);
|
|
|
|
// Close the video file
|
|
avformat_close_input(&ctx->pFormatCtx);
|
|
|
|
if (ctx->file >= 0)
|
|
pFS_Close(ctx->file);
|
|
|
|
free(ctx);
|
|
}
|
|
|
|
static void *AVDec_Create(char *medianame)
|
|
{
|
|
struct decctx *ctx;
|
|
|
|
unsigned int i;
|
|
AVCodec *pCodec;
|
|
qboolean useioctx = false;
|
|
|
|
/*only respond to av: media prefixes*/
|
|
if (!strncmp(medianame, "av:", 3))
|
|
{
|
|
medianame = medianame + 3;
|
|
useioctx = true;
|
|
}
|
|
else if (!strncmp(medianame, "avs:", 4))
|
|
{
|
|
medianame = medianame + 4;
|
|
//let avformat do its own avio context stuff
|
|
}
|
|
else
|
|
return NULL;
|
|
|
|
ctx = malloc(sizeof(*ctx));
|
|
memset(ctx, 0, sizeof(*ctx));
|
|
|
|
//so we always decode the first frame instantly.
|
|
|
|
ctx->starttime = pSys_Milliseconds();
|
|
|
|
ctx->file = -1;
|
|
if (useioctx)
|
|
{
|
|
// Create internal Buffer for FFmpeg:
|
|
const int iBufSize = 32 * 1024;
|
|
char *pBuffer = malloc(iBufSize);
|
|
AVIOContext *ioctx;
|
|
|
|
ctx->filelen = pFS_Open(medianame, &ctx->file, 1);
|
|
if (ctx->filelen < 0)
|
|
{
|
|
Con_Printf("Unable to open %s\n", medianame);
|
|
free(ctx);
|
|
free(pBuffer);
|
|
return NULL;
|
|
}
|
|
|
|
ioctx = avio_alloc_context(pBuffer, iBufSize, 0, ctx, AVIO_Read, 0, AVIO_Seek);
|
|
ctx->pFormatCtx = avformat_alloc_context();
|
|
|
|
ctx->pFormatCtx->pb = ioctx;
|
|
}
|
|
/*
|
|
small how-to note for if I ever try to add support for voice-and-video rtp decoding.
|
|
this stuff is presumably needed to handle ICE+stun+ports etc.
|
|
I prolly need to hack around with adding rtcp too. :s
|
|
|
|
rtsp: Add support for depacketizing RTP data via custom IO
|
|
|
|
To use this, set sdpflags=custom_io to the sdp demuxer. During
|
|
the avformat_open_input call, the SDP is read from the AVFormatContext
|
|
AVIOContext (ctx->pb) - after the avformat_open_input call,
|
|
during the av_read_frame() calls, the same ctx->pb is used for reading
|
|
packets (and sending back RTCP RR packets).
|
|
|
|
Normally, one would use this with a read-only AVIOContext for the
|
|
SDP during the avformat_open_input call, then close that one and
|
|
replace it with a read-write one for the packets after the
|
|
avformat_open_input call has returned.
|
|
|
|
This allows using the RTP depacketizers as "pure" demuxers, without
|
|
having them tied to the libavformat network IO.
|
|
*/
|
|
|
|
|
|
// Open video file
|
|
if(avformat_open_input(&ctx->pFormatCtx, medianame, NULL, NULL)==0)
|
|
{
|
|
// Retrieve stream information
|
|
if(avformat_find_stream_info(ctx->pFormatCtx, NULL)>=0)
|
|
{
|
|
ctx->audioStream=-1;
|
|
for(i=0; i<ctx->pFormatCtx->nb_streams; i++)
|
|
if(ctx->pFormatCtx->streams[i]->codec->codec_type==AVMEDIA_TYPE_AUDIO)
|
|
{
|
|
ctx->audioStream=i;
|
|
break;
|
|
}
|
|
if(ctx->audioStream!=-1)
|
|
{
|
|
ctx->pACodecCtx=ctx->pFormatCtx->streams[ctx->audioStream]->codec;
|
|
pCodec=avcodec_find_decoder(ctx->pACodecCtx->codec_id);
|
|
|
|
ctx->pAFrame=avcodec_alloc_frame();
|
|
if(pCodec!=NULL && ctx->pAFrame && avcodec_open2(ctx->pACodecCtx, pCodec, NULL) >= 0)
|
|
{
|
|
|
|
}
|
|
else
|
|
ctx->audioStream = -1;
|
|
}
|
|
|
|
ctx->videoStream=-1;
|
|
for(i=0; i<ctx->pFormatCtx->nb_streams; i++)
|
|
if(ctx->pFormatCtx->streams[i]->codec->codec_type==AVMEDIA_TYPE_VIDEO)
|
|
{
|
|
ctx->videoStream=i;
|
|
break;
|
|
}
|
|
if(ctx->videoStream!=-1)
|
|
{
|
|
// Get a pointer to the codec context for the video stream
|
|
ctx->pVCodecCtx=ctx->pFormatCtx->streams[ctx->videoStream]->codec;
|
|
ctx->num = ctx->pFormatCtx->streams[ctx->videoStream]->time_base.num;
|
|
ctx->denum = ctx->pFormatCtx->streams[ctx->videoStream]->time_base.den;
|
|
|
|
// Find the decoder for the video stream
|
|
pCodec=avcodec_find_decoder(ctx->pVCodecCtx->codec_id);
|
|
|
|
// Open codec
|
|
if(pCodec!=NULL && avcodec_open2(ctx->pVCodecCtx, pCodec, NULL) >= 0)
|
|
{
|
|
// Allocate video frame
|
|
ctx->pVFrame=avcodec_alloc_frame();
|
|
if(ctx->pVFrame!=NULL)
|
|
{
|
|
if (AVDec_SetSize(ctx, ctx->pVCodecCtx->width, ctx->pVCodecCtx->height))
|
|
{
|
|
return ctx;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
AVDec_Destroy(ctx);
|
|
return NULL;
|
|
}
|
|
|
|
static void *AVDec_DisplayFrame(void *vctx, qboolean nosound, uploadfmt_t *fmt, int *width, int *height)
|
|
{
|
|
struct decctx *ctx = (struct decctx*)vctx;
|
|
AVPacket packet;
|
|
int frameFinished;
|
|
qboolean repainted = false;
|
|
int64_t curtime, lasttime;
|
|
|
|
curtime = ((pSys_Milliseconds() - ctx->starttime) * ctx->denum);
|
|
curtime /= (ctx->num * 1000);
|
|
|
|
*fmt = TF_BGRA32;
|
|
while (1)
|
|
{
|
|
lasttime = av_frame_get_best_effort_timestamp(ctx->pVFrame);
|
|
|
|
if (lasttime > curtime)
|
|
break;
|
|
|
|
// We're ahead of the previous frame. try and read the next.
|
|
if (av_read_frame(ctx->pFormatCtx, &packet) < 0)
|
|
{
|
|
*fmt = TF_INVALID;
|
|
break;
|
|
}
|
|
|
|
// Is this a packet from the video stream?
|
|
if(packet.stream_index==ctx->videoStream)
|
|
{
|
|
// Decode video frame
|
|
avcodec_decode_video2(ctx->pVCodecCtx, ctx->pVFrame, &frameFinished, &packet);
|
|
|
|
// Did we get a video frame?
|
|
if(frameFinished)
|
|
{
|
|
ctx->pScaleCtx = sws_getCachedContext(ctx->pScaleCtx, ctx->pVCodecCtx->width, ctx->pVCodecCtx->height, ctx->pVCodecCtx->pix_fmt, ctx->width, ctx->height, AV_PIX_FMT_BGRA, SWS_POINT, 0, 0, 0);
|
|
|
|
// Convert the image from its native format to RGB
|
|
sws_scale(ctx->pScaleCtx, (void*)ctx->pVFrame->data, ctx->pVFrame->linesize, 0, ctx->pVCodecCtx->height, ctx->pFrameRGB.data, ctx->pFrameRGB.linesize);
|
|
|
|
repainted = true;
|
|
}
|
|
}
|
|
else if(packet.stream_index==ctx->audioStream && !nosound)
|
|
{
|
|
int okay;
|
|
int len;
|
|
void *odata = packet.data;
|
|
while (packet.size > 0)
|
|
{
|
|
okay = false;
|
|
len = avcodec_decode_audio4(ctx->pACodecCtx, ctx->pAFrame, &okay, &packet);
|
|
if (len < 0)
|
|
break;
|
|
packet.size -= len;
|
|
packet.data += len;
|
|
if (okay)
|
|
{
|
|
int width = 2;
|
|
unsigned int auddatasize = av_samples_get_buffer_size(NULL, ctx->pACodecCtx->channels, ctx->pAFrame->nb_samples, ctx->pACodecCtx->sample_fmt, 1);
|
|
void *auddata = ctx->pAFrame->data[0];
|
|
switch(ctx->pACodecCtx->sample_fmt)
|
|
{
|
|
default:
|
|
auddatasize = 0;
|
|
break;
|
|
case AV_SAMPLE_FMT_U8:
|
|
width = 1;
|
|
break;
|
|
case AV_SAMPLE_FMT_S16:
|
|
width = 2;
|
|
break;
|
|
case AV_SAMPLE_FMT_FLT:
|
|
{
|
|
float *in = (void*)auddata;
|
|
signed short *out = (void*)auddata;
|
|
int v;
|
|
unsigned int i;
|
|
for (i = 0; i < auddatasize/sizeof(*in); i++)
|
|
{
|
|
v = (short)(in[i]*32767);
|
|
if (v < -32767)
|
|
v = -32767;
|
|
else if (v > 32767)
|
|
v = 32767;
|
|
out[i] = v;
|
|
}
|
|
auddatasize/=2;
|
|
width = 2;
|
|
}
|
|
break;
|
|
}
|
|
pS_RawAudio(-1, auddata, ctx->pACodecCtx->sample_rate, auddatasize/(ctx->pACodecCtx->channels*width), ctx->pACodecCtx->channels, width, 1);
|
|
}
|
|
}
|
|
packet.data = odata;
|
|
}
|
|
|
|
// Free the packet that was allocated by av_read_frame
|
|
av_free_packet(&packet);
|
|
}
|
|
|
|
*width = ctx->width;
|
|
*height = ctx->height;
|
|
if (!repainted)
|
|
return NULL;
|
|
return ctx->pFrameRGB.data[0];
|
|
}
|
|
static void AVDec_GetSize (void *vctx, int *width, int *height)
|
|
{
|
|
struct decctx *ctx = (struct decctx*)vctx;
|
|
*width = ctx->width;
|
|
*height = ctx->height;
|
|
}
|
|
|
|
/*static void AVDec_CursorMove (void *vctx, float posx, float posy)
|
|
{
|
|
//its a video, dumbass
|
|
}
|
|
static void AVDec_Key (void *vctx, int code, int unicode, int isup)
|
|
{
|
|
//its a video, dumbass
|
|
}
|
|
static void AVDec_ChangeStream(void *vctx, char *newstream)
|
|
{
|
|
}
|
|
*/
|
|
static void AVDec_Rewind(void *vctx)
|
|
{
|
|
struct decctx *ctx = (struct decctx*)vctx;
|
|
if (ctx->videoStream >= 0)
|
|
av_seek_frame(ctx->pFormatCtx, ctx->videoStream, 0, AVSEEK_FLAG_BACKWARD);
|
|
if (ctx->audioStream >= 0)
|
|
av_seek_frame(ctx->pFormatCtx, ctx->audioStream, 0, AVSEEK_FLAG_BACKWARD);
|
|
|
|
ctx->starttime = pSys_Milliseconds();
|
|
}
|
|
|
|
/*
|
|
//avcodec has no way to shut down properly.
|
|
static qintptr_t AVDec_Shutdown(qintptr_t *args)
|
|
{
|
|
return 0;
|
|
}
|
|
*/
|
|
|
|
static media_decoder_funcs_t decoderfuncs =
|
|
{
|
|
"avplug",
|
|
AVDec_Create,
|
|
AVDec_DisplayFrame,
|
|
NULL,//doneframe
|
|
AVDec_Destroy,
|
|
AVDec_Rewind,
|
|
|
|
NULL,//AVDec_CursorMove,
|
|
NULL,//AVDec_Key,
|
|
NULL,//AVDec_SetSize,
|
|
AVDec_GetSize,
|
|
NULL,//AVDec_ChangeStream
|
|
};
|
|
|
|
static qboolean AVDec_Init(void)
|
|
{
|
|
if (!pPlug_ExportNative("Media_VideoDecoder", &decoderfuncs))
|
|
{
|
|
Con_Printf("avplug: Engine doesn't support media decoder plugins\n");
|
|
return false;
|
|
}
|
|
|
|
CHECKBUILTIN(S_RawAudio);
|
|
CHECKBUILTIN(FS_Seek);
|
|
|
|
return true;
|
|
}
|
|
|
|
//get the encoder/decoders to register themselves with the engine, then make sure avformat/avcodec have registered all they have to give.
|
|
qboolean AVEnc_Init(void);
|
|
qintptr_t Plug_Init(qintptr_t *args)
|
|
{
|
|
qboolean okay = false;
|
|
|
|
okay |= AVDec_Init();
|
|
okay |= AVEnc_Init();
|
|
if (okay)
|
|
{
|
|
av_register_all();
|
|
avcodec_register_all();
|
|
}
|
|
return okay;
|
|
}
|
|
|