Home | History | Annotate | Download | only in utils
      1 // Copyright 2012 Google Inc. All Rights Reserved.
      2 //
      3 // Use of this source code is governed by a BSD-style license
      4 // that can be found in the COPYING file in the root of the source
      5 // tree. An additional intellectual property rights grant can be found
      6 // in the file PATENTS. All contributing project authors may
      7 // be found in the AUTHORS file in the root of the source tree.
      8 // -----------------------------------------------------------------------------
      9 //
     10 // Utilities for building and looking up Huffman trees.
     11 //
     12 // Author: Urvang Joshi (urvang (at) google.com)
     13 
     14 #include <assert.h>
     15 #include <stdlib.h>
     16 #include <string.h>
     17 #include "./huffman_utils.h"
     18 #include "./utils.h"
     19 #include "../webp/format_constants.h"
     20 
     21 // Huffman data read via DecodeImageStream is represented in two (red and green)
     22 // bytes.
     23 #define MAX_HTREE_GROUPS    0x10000
     24 
     25 HTreeGroup* VP8LHtreeGroupsNew(int num_htree_groups) {
     26   HTreeGroup* const htree_groups =
     27       (HTreeGroup*)WebPSafeMalloc(num_htree_groups, sizeof(*htree_groups));
     28   if (htree_groups == NULL) {
     29     return NULL;
     30   }
     31   assert(num_htree_groups <= MAX_HTREE_GROUPS);
     32   return htree_groups;
     33 }
     34 
     35 void VP8LHtreeGroupsFree(HTreeGroup* const htree_groups) {
     36   if (htree_groups != NULL) {
     37     WebPSafeFree(htree_groups);
     38   }
     39 }
     40 
     41 // Returns reverse(reverse(key, len) + 1, len), where reverse(key, len) is the
     42 // bit-wise reversal of the len least significant bits of key.
     43 static WEBP_INLINE uint32_t GetNextKey(uint32_t key, int len) {
     44   uint32_t step = 1 << (len - 1);
     45   while (key & step) {
     46     step >>= 1;
     47   }
     48   return step ? (key & (step - 1)) + step : key;
     49 }
     50 
     51 // Stores code in table[0], table[step], table[2*step], ..., table[end].
     52 // Assumes that end is an integer multiple of step.
     53 static WEBP_INLINE void ReplicateValue(HuffmanCode* table,
     54                                        int step, int end,
     55                                        HuffmanCode code) {
     56   assert(end % step == 0);
     57   do {
     58     end -= step;
     59     table[end] = code;
     60   } while (end > 0);
     61 }
     62 
     63 // Returns the table width of the next 2nd level table. count is the histogram
     64 // of bit lengths for the remaining symbols, len is the code length of the next
     65 // processed symbol
     66 static WEBP_INLINE int NextTableBitSize(const int* const count,
     67                                         int len, int root_bits) {
     68   int left = 1 << (len - root_bits);
     69   while (len < MAX_ALLOWED_CODE_LENGTH) {
     70     left -= count[len];
     71     if (left <= 0) break;
     72     ++len;
     73     left <<= 1;
     74   }
     75   return len - root_bits;
     76 }
     77 
     78 // sorted[code_lengths_size] is a pre-allocated array for sorting symbols
     79 // by code length.
     80 static int BuildHuffmanTable(HuffmanCode* const root_table, int root_bits,
     81                              const int code_lengths[], int code_lengths_size,
     82                              uint16_t sorted[]) {
     83   HuffmanCode* table = root_table;  // next available space in table
     84   int total_size = 1 << root_bits;  // total size root table + 2nd level table
     85   int len;                          // current code length
     86   int symbol;                       // symbol index in original or sorted table
     87   // number of codes of each length:
     88   int count[MAX_ALLOWED_CODE_LENGTH + 1] = { 0 };
     89   // offsets in sorted table for each length:
     90   int offset[MAX_ALLOWED_CODE_LENGTH + 1];
     91 
     92   assert(code_lengths_size != 0);
     93   assert(code_lengths != NULL);
     94   assert(root_table != NULL);
     95   assert(root_bits > 0);
     96 
     97   // Build histogram of code lengths.
     98   for (symbol = 0; symbol < code_lengths_size; ++symbol) {
     99     if (code_lengths[symbol] > MAX_ALLOWED_CODE_LENGTH) {
    100       return 0;
    101     }
    102     ++count[code_lengths[symbol]];
    103   }
    104 
    105   // Error, all code lengths are zeros.
    106   if (count[0] == code_lengths_size) {
    107     return 0;
    108   }
    109 
    110   // Generate offsets into sorted symbol table by code length.
    111   offset[1] = 0;
    112   for (len = 1; len < MAX_ALLOWED_CODE_LENGTH; ++len) {
    113     if (count[len] > (1 << len)) {
    114       return 0;
    115     }
    116     offset[len + 1] = offset[len] + count[len];
    117   }
    118 
    119   // Sort symbols by length, by symbol order within each length.
    120   for (symbol = 0; symbol < code_lengths_size; ++symbol) {
    121     const int symbol_code_length = code_lengths[symbol];
    122     if (code_lengths[symbol] > 0) {
    123       sorted[offset[symbol_code_length]++] = symbol;
    124     }
    125   }
    126 
    127   // Special case code with only one value.
    128   if (offset[MAX_ALLOWED_CODE_LENGTH] == 1) {
    129     HuffmanCode code;
    130     code.bits = 0;
    131     code.value = (uint16_t)sorted[0];
    132     ReplicateValue(table, 1, total_size, code);
    133     return total_size;
    134   }
    135 
    136   {
    137     int step;              // step size to replicate values in current table
    138     uint32_t low = -1;     // low bits for current root entry
    139     uint32_t mask = total_size - 1;    // mask for low bits
    140     uint32_t key = 0;      // reversed prefix code
    141     int num_nodes = 1;     // number of Huffman tree nodes
    142     int num_open = 1;      // number of open branches in current tree level
    143     int table_bits = root_bits;        // key length of current table
    144     int table_size = 1 << table_bits;  // size of current table
    145     symbol = 0;
    146     // Fill in root table.
    147     for (len = 1, step = 2; len <= root_bits; ++len, step <<= 1) {
    148       num_open <<= 1;
    149       num_nodes += num_open;
    150       num_open -= count[len];
    151       if (num_open < 0) {
    152         return 0;
    153       }
    154       for (; count[len] > 0; --count[len]) {
    155         HuffmanCode code;
    156         code.bits = (uint8_t)len;
    157         code.value = (uint16_t)sorted[symbol++];
    158         ReplicateValue(&table[key], step, table_size, code);
    159         key = GetNextKey(key, len);
    160       }
    161     }
    162 
    163     // Fill in 2nd level tables and add pointers to root table.
    164     for (len = root_bits + 1, step = 2; len <= MAX_ALLOWED_CODE_LENGTH;
    165          ++len, step <<= 1) {
    166       num_open <<= 1;
    167       num_nodes += num_open;
    168       num_open -= count[len];
    169       if (num_open < 0) {
    170         return 0;
    171       }
    172       for (; count[len] > 0; --count[len]) {
    173         HuffmanCode code;
    174         if ((key & mask) != low) {
    175           table += table_size;
    176           table_bits = NextTableBitSize(count, len, root_bits);
    177           table_size = 1 << table_bits;
    178           total_size += table_size;
    179           low = key & mask;
    180           root_table[low].bits = (uint8_t)(table_bits + root_bits);
    181           root_table[low].value = (uint16_t)((table - root_table) - low);
    182         }
    183         code.bits = (uint8_t)(len - root_bits);
    184         code.value = (uint16_t)sorted[symbol++];
    185         ReplicateValue(&table[key >> root_bits], step, table_size, code);
    186         key = GetNextKey(key, len);
    187       }
    188     }
    189 
    190     // Check if tree is full.
    191     if (num_nodes != 2 * offset[MAX_ALLOWED_CODE_LENGTH] - 1) {
    192       return 0;
    193     }
    194   }
    195 
    196   return total_size;
    197 }
    198 
    199 // Maximum code_lengths_size is 2328 (reached for 11-bit color_cache_bits).
    200 // More commonly, the value is around ~280.
    201 #define MAX_CODE_LENGTHS_SIZE \
    202   ((1 << MAX_CACHE_BITS) + NUM_LITERAL_CODES + NUM_LENGTH_CODES)
    203 // Cut-off value for switching between heap and stack allocation.
    204 #define SORTED_SIZE_CUTOFF 512
    205 int VP8LBuildHuffmanTable(HuffmanCode* const root_table, int root_bits,
    206                           const int code_lengths[], int code_lengths_size) {
    207   int total_size;
    208   assert(code_lengths_size <= MAX_CODE_LENGTHS_SIZE);
    209   if (code_lengths_size <= SORTED_SIZE_CUTOFF) {
    210     // use local stack-allocated array.
    211     uint16_t sorted[SORTED_SIZE_CUTOFF];
    212     total_size = BuildHuffmanTable(root_table, root_bits,
    213                                    code_lengths, code_lengths_size, sorted);
    214   } else {   // rare case. Use heap allocation.
    215     uint16_t* const sorted =
    216         (uint16_t*)WebPSafeMalloc(code_lengths_size, sizeof(*sorted));
    217     if (sorted == NULL) return 0;
    218     total_size = BuildHuffmanTable(root_table, root_bits,
    219                                    code_lengths, code_lengths_size, sorted);
    220     WebPSafeFree(sorted);
    221   }
    222   return total_size;
    223 }
    224