-
Notifications
You must be signed in to change notification settings - Fork 93
/
decode-detail.html
99 lines (82 loc) · 3.62 KB
/
decode-detail.html
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
<HTML><HEAD>
<TITLE> Detailed Decoding Trace Information </TITLE>
</HEAD><BODY>
<H1> Detailed Decoding Trace Information </H1>
The <B>-T</B> option to <A HREF="decoding.html#decode"><TT>decode</TT></A>
causes a detailed trace of information on the process of decoding to
be written to standard output, in a multi-column format, with the first
line containing the headers for each column. This format is suitable
for reading into S-Plus or R.
The first column is always the number of the block being decoded;
several lines may be output for each block. The other columns vary
with the decoding method used, as described below.
<H2>Enum-bit and Enum-block decoding methods</H2>
For source messages with <I>K</I> bits, 2<SUP><I>K</I></SUP> lines are output
for each block, containing the following information:
<BLOCKQUOTE>
<TABLE>
<tr align="left" valign="top">
<td> <B>block</B> </td>
<td>The number of the block, from zero</td></tr>
<tr align="left" valign="top">
<td> <B>decoding</B> </td>
<td>A possible decoding for the message bits, expressed as a hexadecimal
number. The other bits are determined by the message bits.</td></tr>
<tr align="left" valign="top">
<td> <B>likelihood</B> </td>
<td>The likelihood for this decoding (ie, the probability of obtaining
the data received if this was the message sent).</td></tr>
</TABLE>
</BLOCKQUOTE>
For these methods, the number of "iterations" (output with the
<B>-t</B> option) is always 2<SUP><I>K</I></SUP>.
<H2>Prprp decoding method</H2>
Each block results in one line of output for the initial state (based
on individual likelihood ratios), and one line for each subsequent
iteration, containing the following information:
<BLOCKQUOTE>
<TABLE>
<tr align="left" valign="top">
<td> <B>block</B> </td>
<td>The number of the block, from zero</td></tr>
<tr align="left" valign="top">
<td> <B>iter</B> </td>
<td>The number of an iteration, zero for the initial state.</td></tr>
<tr align="left" valign="top">
<td> <B>changed</B> </td>
<td>The number of bits in the decoding that differ from the bit that would
be chosen based just on the likelihood ratio for that bit. Bits whose
likelihood ratios are exactly one contribute 0.5 to this
count. (Likelihood ratios of exactly one can arise when the output
of an AWGN channel rounds to exactly 0.00.)</td></tr>
<tr align="left" valign="top">
<td> <B>perrs</B> </td>
<td>The number of parity check errors in the current tentative
decoding.</td></tr>
<tr align="left" valign="top">
<td> <B>loglik</B> </td>
<td>The log likelihood of the current tentative decoding.</td></tr>
<tr align="left" valign="top">
<td> <B>Eperrs</B> </td>
<td>The expected number of parity check errors in a decoding found
by randomly picking a value for each bit, independently, according
to the current bit probabilities.</td></tr>
<tr align="left" valign="top">
<td> <B>Eloglik</B> </td>
<td>The expected log likelihood of a decoding found by randomly picking
a value for each bit, independently, according to the current bit
probabilities.
</td></tr>
<tr align="left" valign="top">
<td> <B>entropy</B> </td>
<td>The entropy (in bits) of the distribution defined by the current bit
probablities, assumed to apply to bits independently.</td></tr>
</TABLE>
</BLOCKQUOTE>
The number of "iterations" (output with the <B>-t</B> option) is
the obvious count of probability propagation iterations. The
initial state does not count as an iteration.
<HR>
<A HREF="decoding.html">Back to decoding documentation</A><BR>
<A HREF="index.html">Back to index for LDPC software</A>
</BODY></HTML>