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
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
//! Parsing of layers and their segments, as specified in Annex B.
use alloc::boxed::Box;
use super::build::Segment;
use super::codestream::markers::{EPH, SOP};
use super::codestream::{ComponentInfo, Header};
use super::decode::{DecompositionStorage, TileDecodeContext};
use super::progression::ProgressionData;
use super::tile::{Tile, TilePart};
use crate::error::{Result, TileError, bail};
use crate::reader::BitReader;
pub(crate) const MAX_BITPLANE_COUNT: u8 = 32;
pub(crate) fn parse<'a>(
tile: &'a Tile<'a>,
mut progression_iterator: Box<dyn Iterator<Item = ProgressionData> + '_>,
tile_ctx: &mut TileDecodeContext<'a>,
header: &Header<'_>,
storage: &mut DecompositionStorage<'a>,
) -> Result<()> {
for tile_part in &tile.tile_parts {
if parse_inner(
tile_part.clone(),
&mut progression_iterator,
tile_ctx,
storage,
)
.is_none()
&& header.strict
{
bail!(TileError::Invalid);
}
}
Ok(())
}
fn parse_inner<'a>(
mut tile_part: TilePart<'a>,
progression_iterator: &mut dyn Iterator<Item = ProgressionData>,
tile_ctx: &mut TileDecodeContext<'a>,
storage: &mut DecompositionStorage<'a>,
) -> Option<()> {
while !tile_part.header().at_end() {
let progression_data = progression_iterator.next()?;
let resolution = progression_data.resolution;
let component_info = &tile_ctx.tile.component_infos[progression_data.component as usize];
let tile_decompositions =
&mut storage.tile_decompositions[progression_data.component as usize];
let sub_band_iter = tile_decompositions.sub_band_iter(resolution, &storage.decompositions);
let body_reader = tile_part.body();
if component_info.coding_style.flags.may_use_sop_markers()
&& body_reader.peek_marker() == Some(SOP)
{
body_reader.read_marker().ok()?;
body_reader.skip_bytes(4)?;
}
let header_reader = tile_part.header();
let zero_length = header_reader.read_bits_with_stuffing(1)? == 0;
// B.10.3 Zero length packet
// "The first bit in the packet header denotes whether the packet has a length of zero
// (empty packet). The value 0 indicates a zero length; no code-blocks are included in this
// case. The value 1 indicates a non-zero length."
if !zero_length {
for sub_band in sub_band_iter.clone() {
resolve_segments(
sub_band,
&progression_data,
header_reader,
storage,
component_info,
)?;
}
}
header_reader.align();
if component_info.coding_style.flags.uses_eph_marker()
&& header_reader.read_marker().ok()? != EPH
{
return None;
}
// Now read the packet body.
let body_reader = tile_part.body();
if !zero_length {
for sub_band in sub_band_iter {
let sub_band = &mut storage.sub_bands[sub_band];
let precinct = &mut storage.precincts[sub_band.precincts.clone()]
[progression_data.precinct as usize];
let code_blocks = &mut storage.code_blocks[precinct.code_blocks.clone()];
for code_block in code_blocks {
let layer = &mut storage.layers[code_block.layers.clone()]
[progression_data.layer_num as usize];
if let Some(segments) = layer.segments.clone() {
let segments = &mut storage.segments[segments.clone()];
for segment in segments {
segment.data = body_reader.read_bytes(segment.data_length as usize)?;
}
}
}
}
}
}
Some(())
}
fn resolve_segments(
sub_band_dx: usize,
progression_data: &ProgressionData,
reader: &mut BitReader<'_>,
storage: &mut DecompositionStorage<'_>,
component_info: &ComponentInfo,
) -> Option<()> {
// We don't support more than 32-bit precision.
const MAX_CODING_PASSES: u8 = 1 + 3 * (MAX_BITPLANE_COUNT - 1);
let sub_band = &storage.sub_bands[sub_band_dx];
let precincts = &mut storage.precincts[sub_band.precincts.clone()];
let Some(precinct) = precincts.get_mut(progression_data.precinct as usize) else {
// An invalid file could trigger this code path.
lwarn!("progression data yielded invalid precinct index");
return None;
};
let code_blocks = &mut storage.code_blocks[precinct.code_blocks.clone()];
for code_block in code_blocks {
// B.10.4 Code-block inclusion
let is_included = if code_block.has_been_included {
// "For code-blocks that have been included in a previous packet,
// a single bit is used to represent the information, where a 1
// means that the code-block is included in this layer and a 0 means
// that it is not."
reader.read_bits_with_stuffing(1)? == 1
} else {
// "For code-blocks that have not been previously included in any packet,
// this information is signalled with a separate tag tree code for each precinct
// as confined to a sub-band. The values in this tag tree are the number of the
// layer in which the current code-block is first included. Although the exact
// sequence of bits that represent the inclusion tag tree appears in the bit
// stream, only the bits needed for determining whether the code-block is
// included are placed in the packet header. If some of the tag tree is already
// known from previous code-blocks or previous layers, it is not repeated.
// Likewise, only as much of the tag tree as is needed to determine inclusion in
// the current layer is included. If a code-block is not included until a later
// layer, then only a partial tag tree is included at that point in the bit
// stream."
precinct.code_inclusion_tree.read(
code_block.x_idx,
code_block.y_idx,
reader,
progression_data.layer_num as u32 + 1,
&mut storage.tag_tree_nodes,
)? <= progression_data.layer_num as u32
};
ltrace!("code-block inclusion: {}", is_included);
if !is_included {
continue;
}
let layer =
&mut storage.layers[code_block.layers.clone()][progression_data.layer_num as usize];
let included_first_time = is_included && !code_block.has_been_included;
// B.10.5 Zero bit-plane information
// "If a code-block is included for the first time, the packet header contains
// information identifying the actual number of bit-planes used to represent
// coefficients from the code-block. The maximum number of bit-planes available
// for the representation of coefficients in any sub-band, b, is given by Mb as
// defined in Equation (E-2). In general, however, the
// number of actual bit-planes for which coding passes are generated is Mb – P,
// where the number of missing most significant bit-planes, P, may vary from
// code-block to code-block; these missing bit-planes are all taken to be zero. The
// value of P is coded in the packet header with a separate tag tree for every
// precinct, in the same manner as the code block inclusion information."
if included_first_time {
code_block.missing_bit_planes = precinct.zero_bitplane_tree.read(
code_block.x_idx,
code_block.y_idx,
reader,
u32::MAX,
&mut storage.tag_tree_nodes,
)? as u8;
ltrace!(
"zero bit-plane information: {}",
code_block.missing_bit_planes
);
}
code_block.has_been_included |= is_included;
// B.10.6 Number of coding passes
// "The number of coding passes included in this packet from each code-block is
// identified in the packet header using the codewords shown in Table B.4. This
// table provides for the possibility of signalling up to 164 coding passes."
let added_coding_passes = if reader.peak_bits_with_stuffing(9) == Some(0x1ff) {
reader.read_bits_with_stuffing(9)?;
reader.read_bits_with_stuffing(7)? + 37
} else if reader.peak_bits_with_stuffing(4) == Some(0x0f) {
reader.read_bits_with_stuffing(4)?;
reader.read_bits_with_stuffing(5)? + 6
} else if reader.peak_bits_with_stuffing(4) == Some(0b1110) {
reader.read_bits_with_stuffing(4)?;
5
} else if reader.peak_bits_with_stuffing(4) == Some(0b1101) {
reader.read_bits_with_stuffing(4)?;
4
} else if reader.peak_bits_with_stuffing(4) == Some(0b1100) {
reader.read_bits_with_stuffing(4)?;
3
} else if reader.peak_bits_with_stuffing(2) == Some(0b10) {
reader.read_bits_with_stuffing(2)?;
2
} else if reader.peak_bits_with_stuffing(1) == Some(0) {
reader.read_bits_with_stuffing(1)?;
1
} else {
return None;
} as u8;
ltrace!("number of coding passes: {}", added_coding_passes);
let mut k = 0;
while reader.read_bits_with_stuffing(1)? == 1 {
k += 1;
}
code_block.l_block += k;
let previous_layers_passes = code_block.number_of_coding_passes;
let cumulative_passes = previous_layers_passes.checked_add(added_coding_passes)?;
if cumulative_passes > MAX_CODING_PASSES {
return None;
}
let get_segment_idx = |pass_idx: u8| {
if component_info.code_block_style().termination_on_each_pass {
// If we terminate on each pass, the segment is just the index
// of the pass.
pass_idx
} else if component_info
.code_block_style()
.selective_arithmetic_coding_bypass
{
// Use the formula derived from the table in the spec.
segment_idx_for_bypass(pass_idx)
} else {
// If none of the above flags is activated, the number of
// segments just corresponds to the number of layers.
code_block.non_empty_layer_count
}
};
let start = storage.segments.len();
let mut push_segment = |segment: u8, coding_passes_for_segment: u8| {
let length = {
assert!(coding_passes_for_segment > 0);
// "A codeword segment is the number of bytes contributed to a packet by a
// code-block. The length of a codeword segment is represented by a binary number of length:
// bits = Lblock + floor(log_2(coding passes added))
// where Lblock is a code-block state variable. A separate Lblock is used for each
// code-block in the precinct. The value of Lblock is initially set to three. The
// number of bytes contributed by each code-block is preceded by signalling bits
// that increase the value of Lblock, as needed. A signalling bit of zero indicates
// the current value of Lblock is sufficient. If there are k ones followed by a
// zero, the value of Lblock is incremented by k. While Lblock can only increase,
// the number of bits used to signal the length of the code-block contribution can
// increase or decrease depending on the number of coding passes included."
let length_bits = code_block.l_block + coding_passes_for_segment.ilog2();
reader.read_bits_with_stuffing(length_bits as u8)
}?;
storage.segments.push(Segment {
idx: segment,
data_length: length,
coding_pases: coding_passes_for_segment,
// Will be set later.
data: &[],
});
ltrace!("length({segment}) {}", length);
Some(())
};
let mut last_segment = get_segment_idx(previous_layers_passes);
let mut coding_passes_for_segment = 0;
for coding_pass in previous_layers_passes..cumulative_passes {
let segment = get_segment_idx(coding_pass);
if segment != last_segment {
push_segment(last_segment, coding_passes_for_segment)?;
last_segment = segment;
coding_passes_for_segment = 1;
} else {
coding_passes_for_segment += 1;
}
}
// Flush the final segment if applicable.
if coding_passes_for_segment > 0 {
push_segment(last_segment, coding_passes_for_segment)?;
}
let end = storage.segments.len();
layer.segments = Some(start..end);
code_block.number_of_coding_passes += added_coding_passes;
code_block.non_empty_layer_count += 1;
}
Some(())
}
/// Calculate the segment index for the given pass in arithmetic decoder
/// bypass (see section D.6, Table D.9).
fn segment_idx_for_bypass(pass_idx: u8) -> u8 {
if pass_idx < 10 {
0
} else {
1 + (2 * ((pass_idx - 10) / 3)) + (if ((pass_idx - 10) % 3) == 2 { 1 } else { 0 })
}
}