vnc-enc-tight.h (8498B)
1/* 2 * QEMU VNC display driver: tight encoding 3 * 4 * From libvncserver/rfb/rfbproto.h 5 * Copyright (C) 2005 Rohit Kumar, Johannes E. Schindelin 6 * Copyright (C) 2000-2002 Constantin Kaplinsky. All Rights Reserved. 7 * Copyright (C) 2000 Tridia Corporation. All Rights Reserved. 8 * Copyright (C) 1999 AT&T Laboratories Cambridge. All Rights Reserved. 9 * 10 * 11 * Permission is hereby granted, free of charge, to any person obtaining a copy 12 * of this software and associated documentation files (the "Software"), to deal 13 * in the Software without restriction, including without limitation the rights 14 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 15 * copies of the Software, and to permit persons to whom the Software is 16 * furnished to do so, subject to the following conditions: 17 * 18 * The above copyright notice and this permission notice shall be included in 19 * all copies or substantial portions of the Software. 20 * 21 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 22 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 23 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL 24 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 25 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 26 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN 27 * THE SOFTWARE. 28 */ 29 30#ifndef VNC_ENC_TIGHT_H 31#define VNC_ENC_TIGHT_H 32 33/*- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 34 * Tight Encoding. 35 * 36 *-- The first byte of each Tight-encoded rectangle is a "compression control 37 * byte". Its format is as follows (bit 0 is the least significant one): 38 * 39 * bit 0: if 1, then compression stream 0 should be reset; 40 * bit 1: if 1, then compression stream 1 should be reset; 41 * bit 2: if 1, then compression stream 2 should be reset; 42 * bit 3: if 1, then compression stream 3 should be reset; 43 * bits 7-4: if 1000 (0x08), then the compression type is "fill", 44 * if 1001 (0x09), then the compression type is "jpeg", 45 * if 1010 (0x0A), then the compression type is "png", 46 * if 0xxx, then the compression type is "basic", 47 * values greater than 1010 are not valid. 48 * 49 * If the compression type is "basic", then bits 6..4 of the 50 * compression control byte (those xxx in 0xxx) specify the following: 51 * 52 * bits 5-4: decimal representation is the index of a particular zlib 53 * stream which should be used for decompressing the data; 54 * bit 6: if 1, then a "filter id" byte is following this byte. 55 * 56 *-- The data that follows after the compression control byte described 57 * above depends on the compression type ("fill", "jpeg", "png" or "basic"). 58 * 59 *-- If the compression type is "fill", then the only pixel value follows, in 60 * client pixel format (see NOTE 1). This value applies to all pixels of the 61 * rectangle. 62 * 63 *-- If the compression type is "jpeg" or "png", the following data stream 64 * looks like this: 65 * 66 * 1..3 bytes: data size (N) in compact representation; 67 * N bytes: JPEG or PNG image. 68 * 69 * Data size is compactly represented in one, two or three bytes, according 70 * to the following scheme: 71 * 72 * 0xxxxxxx (for values 0..127) 73 * 1xxxxxxx 0yyyyyyy (for values 128..16383) 74 * 1xxxxxxx 1yyyyyyy zzzzzzzz (for values 16384..4194303) 75 * 76 * Here each character denotes one bit, xxxxxxx are the least significant 7 77 * bits of the value (bits 0-6), yyyyyyy are bits 7-13, and zzzzzzzz are the 78 * most significant 8 bits (bits 14-21). For example, decimal value 10000 79 * should be represented as two bytes: binary 10010000 01001110, or 80 * hexadecimal 90 4E. 81 * 82 *-- If the compression type is "basic" and bit 6 of the compression control 83 * byte was set to 1, then the next (second) byte specifies "filter id" which 84 * tells the decoder what filter type was used by the encoder to pre-process 85 * pixel data before the compression. The "filter id" byte can be one of the 86 * following: 87 * 88 * 0: no filter ("copy" filter); 89 * 1: "palette" filter; 90 * 2: "gradient" filter. 91 * 92 *-- If bit 6 of the compression control byte is set to 0 (no "filter id" 93 * byte), or if the filter id is 0, then raw pixel values in the client 94 * format (see NOTE 1) will be compressed. See below details on the 95 * compression. 96 * 97 *-- The "gradient" filter pre-processes pixel data with a simple algorithm 98 * which converts each color component to a difference between a "predicted" 99 * intensity and the actual intensity. Such a technique does not affect 100 * uncompressed data size, but helps to compress photo-like images better. 101 * Pseudo-code for converting intensities to differences is the following: 102 * 103 * P[i,j] := V[i-1,j] + V[i,j-1] - V[i-1,j-1]; 104 * if (P[i,j] < 0) then P[i,j] := 0; 105 * if (P[i,j] > MAX) then P[i,j] := MAX; 106 * D[i,j] := V[i,j] - P[i,j]; 107 * 108 * Here V[i,j] is the intensity of a color component for a pixel at 109 * coordinates (i,j). MAX is the maximum value of intensity for a color 110 * component. 111 * 112 *-- The "palette" filter converts true-color pixel data to indexed colors 113 * and a palette which can consist of 2..256 colors. If the number of colors 114 * is 2, then each pixel is encoded in 1 bit, otherwise 8 bits is used to 115 * encode one pixel. 1-bit encoding is performed such way that the most 116 * significant bits correspond to the leftmost pixels, and each raw of pixels 117 * is aligned to the byte boundary. When "palette" filter is used, the 118 * palette is sent before the pixel data. The palette begins with an unsigned 119 * byte which value is the number of colors in the palette minus 1 (i.e. 1 120 * means 2 colors, 255 means 256 colors in the palette). Then follows the 121 * palette itself which consist of pixel values in client pixel format (see 122 * NOTE 1). 123 * 124 *-- The pixel data is compressed using the zlib library. But if the data 125 * size after applying the filter but before the compression is less then 12, 126 * then the data is sent as is, uncompressed. Four separate zlib streams 127 * (0..3) can be used and the decoder should read the actual stream id from 128 * the compression control byte (see NOTE 2). 129 * 130 * If the compression is not used, then the pixel data is sent as is, 131 * otherwise the data stream looks like this: 132 * 133 * 1..3 bytes: data size (N) in compact representation; 134 * N bytes: zlib-compressed data. 135 * 136 * Data size is compactly represented in one, two or three bytes, just like 137 * in the "jpeg" compression method (see above). 138 * 139 *-- NOTE 1. If the color depth is 24, and all three color components are 140 * 8-bit wide, then one pixel in Tight encoding is always represented by 141 * three bytes, where the first byte is red component, the second byte is 142 * green component, and the third byte is blue component of the pixel color 143 * value. This applies to colors in palettes as well. 144 * 145 *-- NOTE 2. The decoder must reset compression streams' states before 146 * decoding the rectangle, if some of bits 0,1,2,3 in the compression control 147 * byte are set to 1. Note that the decoder must reset zlib streams even if 148 * the compression type is "fill", "jpeg" or "png". 149 * 150 *-- NOTE 3. The "gradient" filter and "jpeg" compression may be used only 151 * when bits-per-pixel value is either 16 or 32, not 8. 152 * 153 *-- NOTE 4. The width of any Tight-encoded rectangle cannot exceed 2048 154 * pixels. If a rectangle is wider, it must be split into several rectangles 155 * and each one should be encoded separately. 156 * 157 */ 158 159#define VNC_TIGHT_EXPLICIT_FILTER 0x04 160#define VNC_TIGHT_FILL 0x08 161#define VNC_TIGHT_JPEG 0x09 162#define VNC_TIGHT_PNG 0x0A 163#define VNC_TIGHT_MAX_SUBENCODING 0x0A 164 165/* Filters to improve compression efficiency */ 166#define VNC_TIGHT_FILTER_COPY 0x00 167#define VNC_TIGHT_FILTER_PALETTE 0x01 168#define VNC_TIGHT_FILTER_GRADIENT 0x02 169 170/* Note: The following constant should not be changed. */ 171#define VNC_TIGHT_MIN_TO_COMPRESS 12 172 173/* The parameters below may be adjusted. */ 174#define VNC_TIGHT_MIN_SPLIT_RECT_SIZE 4096 175#define VNC_TIGHT_MIN_SOLID_SUBRECT_SIZE 2048 176#define VNC_TIGHT_MAX_SPLIT_TILE_SIZE 16 177 178#define VNC_TIGHT_JPEG_MIN_RECT_SIZE 4096 179#define VNC_TIGHT_DETECT_SUBROW_WIDTH 7 180#define VNC_TIGHT_DETECT_MIN_WIDTH 8 181#define VNC_TIGHT_DETECT_MIN_HEIGHT 8 182 183#endif /* VNC_ENC_TIGHT_H */