1 # libaec - Adaptive Entropy Coding library
3 Libaec provides fast lossless compression of 1 up to 32 bit wide
4 signed or unsigned integers (samples). The library achieves best
5 results for low entropy data as often encountered in space imaging
6 instrument data or numerical model output from weather or climate
7 simulations. While floating point representations are not directly
8 supported, they can also be efficiently coded by grouping exponents
12 [Golomb-Rice](http://en.wikipedia.org/wiki/Golomb_coding) coding as
13 defined in the Space Data System Standard documents [121.0-B-2][1] and
18 In [license.txt](doc/license.txt) a clarification on potentially
19 applying intellectual property rights is given.
23 See [INSTALL](INSTALL) for details.
27 [Libaec can replace SZIP](README.SZIP).
31 In this context efficiency refers to the size of the encoded
32 data. Performance refers to the time it takes to encode data.
34 Suppose you have an array of 32 bit signed integers you want to
35 compress. The pointer pointing to the data shall be called `*source`,
36 output goes into `*dest`.
42 struct aec_stream strm;
46 /* input data is 32 bits wide */
47 strm.bits_per_sample = 32;
49 /* define a block size of 16 */
52 /* the reference sample interval is set to 128 blocks */
55 /* input data is signed and needs to be preprocessed */
56 strm.flags = AEC_DATA_SIGNED | AEC_DATA_PREPROCESS;
58 /* pointer to input */
59 strm.next_in = (unsigned char *)source;
61 /* length of input in bytes */
62 strm.avail_in = source_length * sizeof(int32_t);
64 /* pointer to output buffer */
67 /* length of output buffer in bytes */
68 strm.avail_out = dest_length;
70 /* initialize encoding */
71 if (aec_encode_init(&strm) != AEC_OK)
74 /* Perform encoding in one call and flush output. */
75 /* In this example you must be sure that the output */
76 /* buffer is large enough for all compressed output */
77 if (aec_encode(&strm, AEC_FLUSH) != AEC_OK)
80 /* free all resources used by encoder */
81 aec_encode_end(&strm);
85 `block_size` can vary from 8 to 64 samples. Smaller blocks allow the
86 compression to adapt more rapidly to changing source
87 statistics. Larger blocks create less overhead but can be less
88 efficient if source statistics change across the block.
90 `rsi` sets the reference sample interval. A large RSI will improve
91 performance and efficiency. It will also increase memory requirements
92 since internal buffering is based on RSI size. A smaller RSI may be
93 desirable in situations where each RSI will be packetized and possible
94 error propagation has to be minimized.
98 * `AEC_DATA_SIGNED`: input data are signed integers. Specifying this
99 correctly increases compression efficiency. Default is unsigned.
101 * `AEC_DATA_PREPROCESS`: preprocessing input will improve compression
102 efficiency if data samples are correlated. It will only cost
103 performance for no gain in efficiency if the data is already
106 * `AEC_DATA_MSB`: input data is stored most significant byte first
107 i.e. big endian. You have to specify `AEC_DATA_MSB` even if your host
108 architecture is big endian. Default is little endian on all
111 * `AEC_DATA_3BYTE`: the 17 to 24 bit input data is stored in three
112 bytes. This flag has no effect for other sample sizes.
114 * `AEC_RESTRICTED`: use a restricted set of code options. This option is
115 only valid for `bits_per_sample` <= 4.
117 * `AEC_PAD_RSI`: assume that the encoded RSI is padded to the next byte
118 boundary while decoding. The preprocessor macro `ENABLE_RSI_PADDING`
119 needs to be defined while compiling for the encoder to honour this
124 The following rules apply for deducing storage size from sample size
127 **sample size** | **storage size**
130 9 - 16 bits | 2 bytes
131 17 - 24 bits | 3 bytes (only if `AEC_DATA_3BYTE` is set)
132 25 - 32 bits | 4 bytes (if `AEC_DATA_3BYTE` is set)
133 17 - 32 bits | 4 bytes (if `AEC_DATA_3BYTE` is not set)
135 If a sample requires less bits than the storage size provides, then
136 you have to make sure that unused bits are not set. Libaec does not
137 check this for performance reasons and will produce undefined output
138 if unused bits are set. All input data must be a multiple of the
139 storage size in bytes. Remaining bytes which do not form a complete
140 sample will be ignored.
142 Libaec accesses `next_in` and `next_out` buffers only bytewise. There
143 are no alignment requirements for these buffers.
147 `aec_encode` can be used in a streaming fashion by chunking input and
148 output and specifying `AEC_NO_FLUSH`. The function will return if either
149 the input runs empty or the output buffer is full. The calling
150 function can check `avail_in` and `avail_out` to see what occurred. The
151 last call to `aec_encode()` must set `AEC_FLUSH` to drain all
152 output. aec.c is an example of streaming usage of encoding and
157 Encoded data will be written to the buffer submitted with
158 `next_out`. The length of the compressed data is `total_out`.
160 See libaec.h for a detailed description of all relevant structure
161 members and constants.
166 Using decoding is very similar to encoding, only the meaning of input
167 and output is reversed.
173 struct aec_stream strm;
174 /* this is now the compressed data */
175 unsigned char *source;
176 /* here goes the uncompressed result */
179 strm.bits_per_sample = 32;
180 strm.block_size = 16;
182 strm.flags = AEC_DATA_SIGNED | AEC_DATA_PREPROCESS;
183 strm.next_in = source;
184 strm.avail_in = source_length;
185 strm.next_out = (unsigned char *)dest;
186 strm.avail_out = dest_lenth * sizeof(int32_t);
187 if (aec_decode_init(&strm) != AEC_OK)
189 if (aec_decode(&strm, AEC_FLUSH) != AEC_OK)
191 aec_decode_end(&strm);
195 It is strongly recommended that the size of the output buffer
196 (`next_out`) is a multiple of the storage size in bytes. If the buffer
197 is not a multiple of the storage size and the buffer gets filled to
198 the last sample, the error code `AEC_MEM_ERROR` is returned.
200 It is essential for decoding that parameters like `bits_per_sample`,
201 `block_size`, `rsi`, and `flags` are exactly the same as they were for
202 encoding. Libaec does not store these parameters in the coded stream
203 so it is up to the calling program to keep the correct parameters
204 between encoding and decoding.
206 The actual values of coding parameters are in fact only relevant for
207 efficiency and performance. Data integrity only depends on consistency
213 [Consultative Committee for Space Data Systems. Lossless Data
214 Compression. Recommendation for Space Data System Standards, CCSDS
215 121.0-B-2. Blue Book. Issue 2. Washington, D.C.: CCSDS, May 2012.][1]
216 [1]: http://public.ccsds.org/publications/archive/121x0b2.pdf
218 [Consultative Committee for Space Data Systems. Lossless Data
219 Compression. Recommendation for Space Data System Standards, CCSDS
220 120.0-G-3. Green Book. Issue 3. Washington, D.C.: CCSDS, April 2013.][2]
221 [2]: http://public.ccsds.org/publications/archive/120x0g3.pdf