aboutsummaryrefslogtreecommitdiffstats
path: root/libavcodec/idcinvideo.c
blob: 87a10139667a38d9f3b16ff4d9423e694d4a992b (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
/*
 * id Quake II CIN Video Decoder
 * Copyright (C) 2003 the ffmpeg project
 *
 * This file is part of Libav.
 *
 * 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.
 *
 * 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 Libav; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/**
 * @file
 * id Quake II Cin Video Decoder by Dr. Tim Ferguson
 * For more information about the id CIN format, visit:
 *   http://www.csse.monash.edu.au/~timf/
 *
 * This video decoder outputs PAL8 colorspace data. Interacting with this
 * decoder is a little involved. During initialization, the demuxer must
 * transmit the 65536-byte Huffman table(s) to the decoder via extradata.
 * Then, whenever a palette change is encountered while demuxing the file,
 * the demuxer must use the same extradata space to transmit an
 * AVPaletteControl structure.
 *
 * id CIN video is purely Huffman-coded, intraframe-only codec. It achieves
 * a little more compression by exploiting the fact that adjacent pixels
 * tend to be similar.
 *
 * Note that this decoder could use libavcodec's optimized VLC facilities
 * rather than naive, tree-based Huffman decoding. However, there are 256
 * Huffman tables. Plus, the VLC bit coding order is right -> left instead
 * or left -> right, so all of the bits would have to be reversed. Further,
 * the original Quake II implementation likely used a similar naive
 * decoding algorithm and it worked fine on much lower spec machines.
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include "avcodec.h"
#include "internal.h"
#include "libavutil/internal.h"

#define HUFFMAN_TABLE_SIZE 64 * 1024
#define HUF_TOKENS 256
#define PALETTE_COUNT 256

typedef struct hnode {
  int count;
  unsigned char used;
  int children[2];
} hnode;

typedef struct IdcinContext {

    AVCodecContext *avctx;

    const unsigned char *buf;
    int size;

    hnode huff_nodes[256][HUF_TOKENS*2];
    int num_huff_nodes[256];

    uint32_t pal[256];
} IdcinContext;

/*
 * Find the lowest probability node in a Huffman table, and mark it as
 * being assigned to a higher probability.
 * @return the node index of the lowest unused node, or -1 if all nodes
 * are used.
 */
static int huff_smallest_node(hnode *hnodes, int num_hnodes) {
    int i;
    int best, best_node;

    best = 99999999;
    best_node = -1;
    for(i = 0; i < num_hnodes; i++) {
        if(hnodes[i].used)
            continue;
        if(!hnodes[i].count)
            continue;
        if(hnodes[i].count < best) {
            best = hnodes[i].count;
            best_node = i;
        }
    }

    if(best_node == -1)
        return -1;
    hnodes[best_node].used = 1;
    return best_node;
}

/*
 * Build the Huffman tree using the generated/loaded probabilities histogram.
 *
 * On completion:
 *  huff_nodes[prev][i < HUF_TOKENS] - are the nodes at the base of the tree.
 *  huff_nodes[prev][i >= HUF_TOKENS] - are used to construct the tree.
 *  num_huff_nodes[prev] - contains the index to the root node of the tree.
 *    That is: huff_nodes[prev][num_huff_nodes[prev]] is the root node.
 */
static av_cold void huff_build_tree(IdcinContext *s, int prev) {
    hnode *node, *hnodes;
     int num_hnodes, i;

    num_hnodes = HUF_TOKENS;
    hnodes = s->huff_nodes[prev];
    for(i = 0; i < HUF_TOKENS * 2; i++)
        hnodes[i].used = 0;

    while (1) {
        node = &hnodes[num_hnodes];             /* next free node */

        /* pick two lowest counts */
        node->children[0] = huff_smallest_node(hnodes, num_hnodes);
        if(node->children[0] == -1)
            break;      /* reached the root node */

        node->children[1] = huff_smallest_node(hnodes, num_hnodes);
        if(node->children[1] == -1)
            break;      /* reached the root node */

        /* combine nodes probability for new node */
        node->count = hnodes[node->children[0]].count +
        hnodes[node->children[1]].count;
        num_hnodes++;
    }

    s->num_huff_nodes[prev] = num_hnodes - 1;
}

static av_cold int idcin_decode_init(AVCodecContext *avctx)
{
    IdcinContext *s = avctx->priv_data;
    int i, j, histogram_index = 0;
    unsigned char *histograms;

    s->avctx = avctx;
    avctx->pix_fmt = AV_PIX_FMT_PAL8;

    /* make sure the Huffman tables make it */
    if (s->avctx->extradata_size != HUFFMAN_TABLE_SIZE) {
        av_log(s->avctx, AV_LOG_ERROR, "  id CIN video: expected extradata size of %d\n", HUFFMAN_TABLE_SIZE);
        return -1;
    }

    /* build the 256 Huffman decode trees */
    histograms = (unsigned char *)s->avctx->extradata;
    for (i = 0; i < 256; i++) {
        for(j = 0; j < HUF_TOKENS; j++)
            s->huff_nodes[i][j].count = histograms[histogram_index++];
        huff_build_tree(s, i);
    }

    return 0;
}

static void idcin_decode_vlcs(IdcinContext *s, AVFrame *frame)
{
    hnode *hnodes;
    long x, y;
    int prev;
    unsigned char v = 0;
    int bit_pos, node_num, dat_pos;

    prev = bit_pos = dat_pos = 0;
    for (y = 0; y < (frame->linesize[0] * s->avctx->height);
        y += frame->linesize[0]) {
        for (x = y; x < y + s->avctx->width; x++) {
            node_num = s->num_huff_nodes[prev];
            hnodes = s->huff_nodes[prev];

            while(node_num >= HUF_TOKENS) {
                if(!bit_pos) {
                    if(dat_pos >= s->size) {
                        av_log(s->avctx, AV_LOG_ERROR, "Huffman decode error.\n");
                        return;
                    }
                    bit_pos = 8;
                    v = s->buf[dat_pos++];
                }

                node_num = hnodes[node_num].children[v & 0x01];
                v = v >> 1;
                bit_pos--;
            }

            frame->data[0][x] = node_num;
            prev = node_num;
        }
    }
}

static int idcin_decode_frame(AVCodecContext *avctx,
                              void *data, int *got_frame,
                              AVPacket *avpkt)
{
    const uint8_t *buf = avpkt->data;
    int buf_size = avpkt->size;
    IdcinContext *s = avctx->priv_data;
    const uint8_t *pal = av_packet_get_side_data(avpkt, AV_PKT_DATA_PALETTE, NULL);
    AVFrame *frame = data;
    int ret;

    s->buf = buf;
    s->size = buf_size;

    if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
        av_log(avctx, AV_LOG_ERROR, "  id CIN Video: get_buffer() failed\n");
        return ret;
    }

    idcin_decode_vlcs(s, frame);

    if (pal) {
        frame->palette_has_changed = 1;
        memcpy(s->pal, pal, AVPALETTE_SIZE);
    }
    /* make the palette available on the way out */
    memcpy(frame->data[1], s->pal, AVPALETTE_SIZE);

    *got_frame = 1;

    /* report that the buffer was completely consumed */
    return buf_size;
}

AVCodec ff_idcin_decoder = {
    .name           = "idcinvideo",
    .long_name      = NULL_IF_CONFIG_SMALL("id Quake II CIN video"),
    .type           = AVMEDIA_TYPE_VIDEO,
    .id             = AV_CODEC_ID_IDCIN,
    .priv_data_size = sizeof(IdcinContext),
    .init           = idcin_decode_init,
    .decode         = idcin_decode_frame,
    .capabilities   = CODEC_CAP_DR1,
};