/*
* RoQ Video Encoder.
*
- * Copyright (C) 2007 Vitor <vitor1001@gmail.com>
+ * Copyright (C) 2007 Vitor Sessak <vitor1001@gmail.com>
* Copyright (C) 2004-2007 Eric Lasota
* Based on RoQ specs (C) 2001 Tim Ferguson
*
- * This file is part of FFmpeg.
+ * This file is part of Libav.
*
- * FFmpeg is free software; you can redistribute it and/or
+ * Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
- * FFmpeg is distributed in the hope that it will be useful,
+ * Libav is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
- * License along with FFmpeg; if not, write to the Free Software
+ * License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
- * @file roqvideoenc.c
- * Id RoQ encoder by Vitor. Based on the Switchblade3 library and the
- * Switchblade3 FFmpeg glue by Eric Lasota.
+ * @file
+ * id RoQ encoder by Vitor. Based on the Switchblade3 library and the
+ * Switchblade3 Libav glue by Eric Lasota.
*/
/*
*/
#include <string.h>
-#include <unistd.h>
#include "roqvideo.h"
#include "bytestream.h"
#include "elbg.h"
+#include "mathops.h"
#define CHROMA_BIAS 1
// FIXME Could use DSPContext.sse, but it is not so speed critical (used
// just for motion estimation).
static int block_sse(uint8_t **buf1, uint8_t **buf2, int x1, int y1, int x2,
- int y2, int stride, int size)
+ int y2, int *stride1, int *stride2, int size)
{
int i, k;
int sse=0;
for (k=0; k<3; k++) {
int bias = (k ? CHROMA_BIAS : 4);
for (i=0; i<size; i++)
- sse += bias*eval_sse(buf1[k] + (y1+i)*stride + x1,
- buf2[k] + (y2+i)*stride + x2, size);
+ sse += bias*eval_sse(buf1[k] + (y1+i)*stride1[k] + x1,
+ buf2[k] + (y2+i)*stride2[k] + x2, size);
}
return sse;
return INT_MAX;
return block_sse(enc->frame_to_enc->data, enc->last_frame->data, x, y,
- mx, my, enc->y_stride, size);
+ mx, my,
+ enc->frame_to_enc->linesize, enc->last_frame->linesize,
+ size);
}
/**
- * Returns distortion between two macroblocks
+ * @return distortion between two macroblocks
*/
static inline int squared_diff_macroblock(uint8_t a[], uint8_t b[], int size)
{
int subCels[4];
motion_vect motion;
int cbEntry;
-} subcel_evaluation_t;
+} SubcelEvaluation;
typedef struct
{
int eval_dist[4];
int best_coding;
- subcel_evaluation_t subCels[4];
+ SubcelEvaluation subCels[4];
motion_vect motion;
int cbEntry;
int sourceX, sourceY;
-} cel_evaluation_t;
+} CelEvaluation;
typedef struct
{
uint8_t unpacked_cb2[MAX_CBS_2x2*2*2*3];
uint8_t unpacked_cb4[MAX_CBS_4x4*4*4*3];
uint8_t unpacked_cb4_enlarged[MAX_CBS_4x4*8*8*3];
-} roq_codebooks_t;
+} RoqCodebooks;
/**
* Temporary vars
*/
-typedef struct
+typedef struct RoqTempData
{
- cel_evaluation_t *cel_evals;
+ CelEvaluation *cel_evals;
int f2i4[MAX_CBS_4x4];
int i2f4[MAX_CBS_4x4];
int numCB4;
int numCB2;
- roq_codebooks_t codebooks;
+ RoqCodebooks codebooks;
int *closest_cb2;
int used_option[4];
-} roq_tempdata_t;
+} RoqTempdata;
/**
- * Initializes cel evaluators and sets their source coordinates
+ * Initialize cel evaluators and set their source coordinates
*/
-static void create_cel_evals(RoqContext *enc, roq_tempdata_t *tempData)
+static void create_cel_evals(RoqContext *enc, RoqTempdata *tempData)
{
int n=0, x, y, i;
- tempData->cel_evals = av_malloc(enc->width*enc->height/64 * sizeof(cel_evaluation_t));
+ tempData->cel_evals = av_malloc(enc->width*enc->height/64 * sizeof(CelEvaluation));
/* Map to the ROQ quadtree order */
for (y=0; y<enc->height; y+=16)
static void get_frame_mb(AVFrame *frame, int x, int y, uint8_t mb[], int dim)
{
int i, j, cp;
- int stride = frame->linesize[0];
- for (cp=0; cp<3; cp++)
+ for (cp=0; cp<3; cp++) {
+ int stride = frame->linesize[cp];
for (i=0; i<dim; i++)
for (j=0; j<dim; j++)
*mb++ = frame->data[cp][(y+i)*stride + x + j];
+ }
}
/**
}
/**
- * Gets distortion for all options available to a subcel
+ * Get distortion for all options available to a subcel
*/
-static void gather_data_for_subcel(subcel_evaluation_t *subcel, int x,
- int y, RoqContext *enc, roq_tempdata_t *tempData)
+static void gather_data_for_subcel(SubcelEvaluation *subcel, int x,
+ int y, RoqContext *enc, RoqTempdata *tempData)
{
uint8_t mb4[4*4*3];
uint8_t mb2[2*2*3];
if (enc->framesSinceKeyframe >= 2)
subcel->eval_dist[RoQ_ID_MOT] = block_sse(enc->frame_to_enc->data,
enc->current_frame->data, x,
- y, x, y, enc->y_stride, 4);
+ y, x, y,
+ enc->frame_to_enc->linesize,
+ enc->current_frame->linesize,
+ 4);
else
subcel->eval_dist[RoQ_ID_MOT] = INT_MAX;
}
/**
- * Gets distortion for all options available to a cel
+ * Get distortion for all options available to a cel
*/
-static void gather_data_for_cel(cel_evaluation_t *cel, RoqContext *enc,
- roq_tempdata_t *tempData)
+static void gather_data_for_cel(CelEvaluation *cel, RoqContext *enc,
+ RoqTempdata *tempData)
{
uint8_t mb8[8*8*3];
int index = cel->sourceY*enc->width/64 + cel->sourceX/8;
enc->current_frame->data,
cel->sourceX, cel->sourceY,
cel->sourceX, cel->sourceY,
- enc->y_stride, 8);
+ enc->frame_to_enc->linesize,
+ enc->current_frame->linesize,8);
else
cel->eval_dist[RoQ_ID_MOT] = INT_MAX;
}
}
-static void remap_codebooks(RoqContext *enc, roq_tempdata_t *tempData)
+static void remap_codebooks(RoqContext *enc, RoqTempdata *tempData)
{
int i, j, idx=0;
/**
* Write codebook chunk
*/
-static void write_codebooks(RoqContext *enc, roq_tempdata_t *tempData)
+static void write_codebooks(RoqContext *enc, RoqTempdata *tempData)
{
int i, j;
uint8_t **outp= &enc->out_buf;
/* NOTE: Typecodes must be spooled AFTER arguments!! */
static void write_typecode(CodingSpool *s, uint8_t type)
{
- s->typeSpool |= ((type) & 3) << (14 - s->typeSpoolLength);
+ s->typeSpool |= (type & 3) << (14 - s->typeSpoolLength);
s->typeSpoolLength += 2;
if (s->typeSpoolLength == 16) {
bytestream_put_le16(s->pout, s->typeSpool);
}
}
-static void reconstruct_and_encode_image(RoqContext *enc, roq_tempdata_t *tempData, int w, int h, int numBlocks)
+static void reconstruct_and_encode_image(RoqContext *enc, RoqTempdata *tempData, int w, int h, int numBlocks)
{
int i, j, k;
int x, y;
int dist=0;
roq_qcell *qcell;
- cel_evaluation_t *eval;
+ CelEvaluation *eval;
CodingSpool spool;
av_log(enc->avctx, AV_LOG_ERROR, "Expected distortion: %i Actual: %i\n",
dist,
block_sse(fdata, cdata, 0, 0, 0, 0,
- enc->y_stride, enc->width)); //WARNING: Square dimensions implied...
+ enc->frame_to_enc->linesize,
+ enc->current_frame->linesize,
+ enc->width)); //WARNING: Square dimensions implied...
#endif
}
for (j=0; j<2; j++) {
int x = (top+i)*stride[0] + left + j;
*block++ = data[0][x];
+ x = (top+i)*stride[1] + left + j;
u += data[1][x];
v += data[2][x];
}
}
/**
- * Creates YUV clusters for the entire image
+ * Create YUV clusters for the entire image
*/
static void create_clusters(AVFrame *frame, int w, int h, uint8_t *yuvClusters)
{
}
}
-static void generate_codebook(RoqContext *enc, roq_tempdata_t *tempdata,
+static void generate_codebook(RoqContext *enc, RoqTempdata *tempdata,
int *points, int inputCount, roq_cell *results,
int size, int cbsize)
{
int i, j, k;
int c_size = size*size/4;
- int *buf = points;
+ int *buf;
int *codebook = av_malloc(6*c_size*cbsize*sizeof(int));
int *closest_cb;
av_free(codebook);
}
-static void generate_new_codebooks(RoqContext *enc, roq_tempdata_t *tempData)
+static void generate_new_codebooks(RoqContext *enc, RoqTempdata *tempData)
{
int i,j;
- roq_codebooks_t *codebooks = &tempData->codebooks;
+ RoqCodebooks *codebooks = &tempData->codebooks;
int max = enc->width*enc->height/16;
uint8_t mb2[3*4];
roq_cell *results4 = av_malloc(sizeof(roq_cell)*MAX_CBS_4x4*4);
static void roq_encode_video(RoqContext *enc)
{
- roq_tempdata_t tempData;
+ RoqTempdata *tempData = enc->tmpData;
int i;
- memset(&tempData, 0, sizeof(tempData));
+ memset(tempData, 0, sizeof(*tempData));
- create_cel_evals(enc, &tempData);
+ create_cel_evals(enc, tempData);
- generate_new_codebooks(enc, &tempData);
+ generate_new_codebooks(enc, tempData);
if (enc->framesSinceKeyframe >= 1) {
motion_search(enc, 8);
retry_encode:
for (i=0; i<enc->width*enc->height/64; i++)
- gather_data_for_cel(tempData.cel_evals + i, enc, &tempData);
+ gather_data_for_cel(tempData->cel_evals + i, enc, tempData);
+
+ /* Quake 3 can't handle chunks bigger than 65535 bytes */
+ if (tempData->mainChunkSize/8 > 65535) {
+ av_log(enc->avctx, AV_LOG_ERROR,
+ "Warning, generated a frame too big (%d > 65535), "
+ "try using a smaller qscale value.\n",
+ tempData->mainChunkSize/8);
+ enc->lambda *= 1.5;
+ tempData->mainChunkSize = 0;
+ memset(tempData->used_option, 0, sizeof(tempData->used_option));
+ memset(tempData->codebooks.usedCB4, 0,
+ sizeof(tempData->codebooks.usedCB4));
+ memset(tempData->codebooks.usedCB2, 0,
+ sizeof(tempData->codebooks.usedCB2));
- /* Quake 3 can't handle chunks bigger than 65536 bytes */
- if (tempData.mainChunkSize/8 > 65536) {
- enc->lambda *= .8;
goto retry_encode;
}
- remap_codebooks(enc, &tempData);
+ remap_codebooks(enc, tempData);
- write_codebooks(enc, &tempData);
+ write_codebooks(enc, tempData);
- reconstruct_and_encode_image(enc, &tempData, enc->width, enc->height,
+ reconstruct_and_encode_image(enc, tempData, enc->width, enc->height,
enc->width*enc->height/64);
+ enc->avctx->coded_frame = enc->current_frame;
+
/* Rotate frame history */
FFSWAP(AVFrame *, enc->current_frame, enc->last_frame);
FFSWAP(motion_vect *, enc->last_motion4, enc->this_motion4);
FFSWAP(motion_vect *, enc->last_motion8, enc->this_motion8);
- av_free(tempData.cel_evals);
- av_free(tempData.closest_cb2);
+ av_free(tempData->cel_evals);
+ av_free(tempData->closest_cb2);
enc->framesSinceKeyframe++;
}
{
RoqContext *enc = avctx->priv_data;
- av_init_random(1, &enc->randctx);
+ av_lfg_init(&enc->randctx, 1);
enc->framesSinceKeyframe = 0;
if ((avctx->width & 0xf) || (avctx->height & 0xf)) {
if (((avctx->width)&(avctx->width-1))||((avctx->height)&(avctx->height-1)))
av_log(avctx, AV_LOG_ERROR, "Warning: dimensions not power of two\n");
- if (avcodec_check_dimensions(avctx, avctx->width, avctx->height)) {
- av_log(avctx, AV_LOG_ERROR, "Invalid dimensions (%dx%d)\n",
- avctx->width, avctx->height);
- return -1;
- }
-
enc->width = avctx->width;
enc->height = avctx->height;
enc->last_frame = &enc->frames[0];
enc->current_frame = &enc->frames[1];
+ enc->tmpData = av_malloc(sizeof(RoqTempdata));
+
enc->this_motion4 =
av_mallocz((enc->width*enc->height/16)*sizeof(motion_vect));
uint8_t *buf_start = buf;
enc->out_buf = buf;
- enc->y_stride = frame->linesize[0];
- enc->c_stride = frame->linesize[1];
enc->avctx = avctx;
enc->frame_to_enc = frame;
avctx->release_buffer(avctx, enc->last_frame);
avctx->release_buffer(avctx, enc->current_frame);
+ av_free(enc->tmpData);
av_free(enc->this_motion4);
av_free(enc->last_motion4);
av_free(enc->this_motion8);
return 0;
}
-AVCodec roq_encoder =
+AVCodec ff_roq_encoder =
{
"roqvideo",
- CODEC_TYPE_VIDEO,
+ AVMEDIA_TYPE_VIDEO,
CODEC_ID_ROQ,
sizeof(RoqContext),
roq_encode_init,
roq_encode_frame,
roq_encode_end,
- .supported_framerates = (AVRational[]){{30,1}, {0,0}},
- .pix_fmts = (enum PixelFormat[]){PIX_FMT_YUV444P, -1},
+ .supported_framerates = (const AVRational[]){{30,1}, {0,0}},
+ .pix_fmts = (const enum PixelFormat[]){PIX_FMT_YUV444P, PIX_FMT_NONE},
+ .long_name = NULL_IF_CONFIG_SMALL("id RoQ video"),
};