1 |
greg |
1.1 |
.\" RCSid "$Id$" |
2 |
|
|
.TH RPIECE 1 5/25/05 RADIANCE |
3 |
|
|
.SH NAME |
4 |
|
|
rtcontrib - compute contributions in a RADIANCE scene |
5 |
|
|
.SH SYNOPSIS |
6 |
|
|
.B rtcontrib |
7 |
|
|
[ |
8 |
|
|
.B "\-n nprocs" |
9 |
|
|
][ |
10 |
|
|
.B "\-e expr" |
11 |
|
|
][ |
12 |
|
|
.B "\-f source" |
13 |
|
|
][ |
14 |
|
|
.B "\-o fspec" |
15 |
|
|
][ |
16 |
|
|
.B "\-b binv" |
17 |
|
|
] |
18 |
|
|
[ |
19 |
|
|
.B $EVAR |
20 |
|
|
] |
21 |
|
|
[ |
22 |
|
|
.B @file |
23 |
|
|
] |
24 |
|
|
[ |
25 |
|
|
rtrace options |
26 |
|
|
] |
27 |
|
|
.B "\-m mod .." |
28 |
|
|
.B octree |
29 |
|
|
.SH DESCRIPTION |
30 |
|
|
.I Rtcontrib |
31 |
|
|
computes ray contributions (i.e., color coefficients) |
32 |
|
|
for objects whose modifiers are named in one or more |
33 |
|
|
.I \-m |
34 |
|
|
settings. |
35 |
|
|
These modifiers are usually materials associated with specific |
36 |
|
|
light sources, though they could correspond to intermediate objects as well. |
37 |
|
|
The resulting contributions may then be used in linear combination to |
38 |
|
|
reproduce any desired variation, e.g., simulating lighting controls or |
39 |
|
|
changing sky conditions via daylight coefficients. |
40 |
|
|
More generally, |
41 |
|
|
.I rtcontrib |
42 |
|
|
can compute general input-output relationships in optical |
43 |
|
|
systems, such as light pipes and shading devices. |
44 |
|
|
.PP |
45 |
|
|
.I Rtrace(1) |
46 |
|
|
is called to calculate the contributions for each input ray, |
47 |
|
|
and the output tallies are sent to one or more files according to the |
48 |
|
|
.I \-o |
49 |
|
|
specification. |
50 |
|
|
If the output file specification contains a "%s" format, this will be |
51 |
|
|
replaced by the modifier name. |
52 |
|
|
The |
53 |
|
|
.I \-b |
54 |
|
|
option may be used to further define |
55 |
|
|
a "bin number" within each object if finer resolution is desired, and |
56 |
|
|
will be applied to a "%d" format in the output file |
57 |
|
|
specification if present. |
58 |
|
|
(The actual bin number is computed at run time based on ray direction |
59 |
|
|
and surface intersection, as described below.)\0 |
60 |
|
|
The most recent |
61 |
|
|
.I \-b |
62 |
|
|
and |
63 |
|
|
.I \-o |
64 |
|
|
options (to the left) of each |
65 |
|
|
.I \-m |
66 |
|
|
setting affect only that modifier, and the ordering |
67 |
|
|
of other options is unimportant. |
68 |
|
|
.PP |
69 |
|
|
Input and output format defaults to plain text, where each ray's |
70 |
|
|
origin and direction (6 real values) must appear together per |
71 |
|
|
line of input, and one line of output is produced per output file |
72 |
|
|
file per ray. |
73 |
|
|
Alternative input and output formats may be specified using the |
74 |
|
|
.I \-f[io] |
75 |
|
|
option, which is explained in the |
76 |
|
|
.I rtrace |
77 |
|
|
man page along with the associated |
78 |
|
|
.I \-x |
79 |
|
|
and |
80 |
|
|
.I \-y |
81 |
|
|
resolution settings. |
82 |
|
|
In particular, the 'c' output setting |
83 |
|
|
together with positive dimensions for |
84 |
|
|
.I \-x |
85 |
|
|
and |
86 |
|
|
.I \-y |
87 |
|
|
will produce an uncompressed RADIANCE picture, |
88 |
|
|
suitable for manipulation with |
89 |
|
|
.I pcomb(1) |
90 |
|
|
and related tools. |
91 |
|
|
.PP |
92 |
|
|
If no |
93 |
|
|
.I \-o |
94 |
|
|
specification is given, results are written on the standard output in order |
95 |
|
|
of modifier (as given on the command line) then bin number. |
96 |
|
|
The same format is used for a simple file name specification |
97 |
|
|
without any embedded "%s" or "%d" formats. |
98 |
|
|
If a "%s" format appears but no "%d" in the |
99 |
|
|
.I \-o |
100 |
|
|
specification, then each modifier will have its own output file, with |
101 |
|
|
multiple values per record in the case of a non-zero |
102 |
|
|
.I \-b |
103 |
|
|
definition. |
104 |
|
|
If a "%d" format appears but no "%s", then each bin will get its own |
105 |
|
|
output file, with modifiers output in order in each record. |
106 |
|
|
For text output, each RGB coefficient triple is separated by a tab, |
107 |
|
|
with a newline at the end of each ray record. |
108 |
|
|
For binary output formats, there is no such delimiter to mark |
109 |
|
|
the end of each record. |
110 |
|
|
.PP |
111 |
|
|
If a |
112 |
|
|
.I \-b |
113 |
|
|
expression is defined for a particular modifier, |
114 |
|
|
the bin number will be evaluated at run time for each |
115 |
|
|
ray contribution from |
116 |
|
|
.I rtrace. |
117 |
|
|
Specifically, each ray's world intersection point will be assigned to |
118 |
|
|
the variables Px, Py, and Pz, and the normalized ray direction |
119 |
|
|
will be assigned to Dx, Dy, and Dz. |
120 |
|
|
These ray parameters may be combined with any definitions given in |
121 |
|
|
.I \-e |
122 |
|
|
options, or any files read in from |
123 |
|
|
.I \-f |
124 |
|
|
options, to compute the bin, which will be |
125 |
|
|
rounded to the closest whole number. |
126 |
|
|
This mechanism allows the user to define precise regions (or directions) |
127 |
|
|
they wish to accumulate, such as the Tregenza sky grid, which would be |
128 |
|
|
otherwise impossible to specify as a set of RADIANCE primitives. |
129 |
|
|
.PP |
130 |
|
|
If the |
131 |
|
|
.I \-n |
132 |
|
|
option is specified with a value greater than 1, multiple |
133 |
|
|
.I rtrace(1) |
134 |
|
|
processes will be used to accelerate computation on a shared |
135 |
|
|
memory machine. |
136 |
|
|
Note that there is no benefit to using more processes |
137 |
|
|
than there are local CPUs available to do the work. |
138 |
|
|
.PP |
139 |
|
|
Options may be given on the command line and/or read from the |
140 |
|
|
environment and/or read from a file. |
141 |
|
|
A command argument beginning with a dollar sign ('$') is immediately |
142 |
|
|
replaced by the contents of the given environment variable. |
143 |
|
|
A command argument beginning with an at sign ('@') is immediately |
144 |
|
|
replaced by the contents of the given file. |
145 |
|
|
.SH EXAMPLE |
146 |
|
|
First |
147 |
|
|
.I rpiece |
148 |
|
|
process is started on the machine "goober": |
149 |
|
|
.IP "" .2i |
150 |
|
|
goober% echo 1 8 > syncfile |
151 |
|
|
.br |
152 |
|
|
goober% echo -F syncfile -x 1024 -y 1024 -vf view -o picture octree > args |
153 |
|
|
.br |
154 |
|
|
goober% rpiece @args & |
155 |
|
|
.PP |
156 |
|
|
Second |
157 |
|
|
.I rpiece |
158 |
|
|
processes is started on the machine "sucker": |
159 |
|
|
.IP "" .2i |
160 |
|
|
sucker% rpiece @args & |
161 |
|
|
.SH NOTES |
162 |
|
|
Due to NFS file buffering, the network lock manager is employed to |
163 |
|
|
guarantee consistency in the output file even though non-overlapping |
164 |
|
|
writes are used. |
165 |
|
|
This would tend to slow the process down if |
166 |
|
|
.I rpiece |
167 |
|
|
were to wait for this I/O to complete before starting on the next |
168 |
|
|
piece, so |
169 |
|
|
.I rpiece |
170 |
|
|
forks separate processes to hang around waiting for I/O completion. |
171 |
|
|
The number of processes thus designated is set by the MAXFORK macro |
172 |
|
|
in the program (compiled in the src/util directory). |
173 |
|
|
If the fork call is slow on a system, it may actually be better to |
174 |
|
|
set MAXFORK to zero. |
175 |
|
|
In other cases, the network lock manager may be so slow that this |
176 |
|
|
value should be increased to get the best utilization. |
177 |
|
|
.PP |
178 |
|
|
The output picture is not run-length encoded, and can be quite |
179 |
|
|
large. |
180 |
|
|
The approximate size (in kilobytes) can be computed by the simple |
181 |
|
|
formula: |
182 |
|
|
.IP "" .2i |
183 |
|
|
filesize = xres*yres/256 |
184 |
|
|
.PP |
185 |
|
|
Make sure that there is enough space on the filesystem to hold the |
186 |
|
|
entire picture before beginning. |
187 |
|
|
Once the picture is finished, the |
188 |
|
|
.I ra_rgbe(1) |
189 |
|
|
program with the -r option may be used to convert to a run-length |
190 |
|
|
encoded picture for more efficient storage, although |
191 |
|
|
.I pfilt(1) |
192 |
|
|
or any of the other Radiance picture filters will do the same |
193 |
|
|
thing. |
194 |
|
|
.PP |
195 |
|
|
The ALRM signal may be used to gracefully terminate an |
196 |
|
|
.I rpiece |
197 |
|
|
process after it finishes the current piece. |
198 |
|
|
This permits other currently running or subsequently started |
199 |
|
|
.I rpiece |
200 |
|
|
process(es) to continue rendering the picture without loss. |
201 |
|
|
The |
202 |
|
|
.I \-T |
203 |
|
|
option will send the ALRM signal to |
204 |
|
|
.I rpiece |
205 |
|
|
after the specified number of (decimal) hours. |
206 |
|
|
This is the best way to force a time limit on the computation, |
207 |
|
|
since information will not be lost, though the process may continue |
208 |
|
|
for some time afterwards to finish its current piece. |
209 |
|
|
.SH BUGS |
210 |
|
|
This program may not work on some systems whose NFS lock manager is |
211 |
|
|
unreliable. |
212 |
|
|
In particular, some System V derivative UNIX systems often have |
213 |
|
|
problems with the network lock manager. |
214 |
|
|
If the output is scrambled or rpict aborts with some ambient file |
215 |
|
|
related problem, you should just remove the ambient file and go |
216 |
|
|
back to normal rendering. |
217 |
|
|
.SH AUTHOR |
218 |
|
|
Greg Ward |
219 |
|
|
.SH "SEE ALSO" |
220 |
|
|
cnt(1), getinfo(1), pcomb(1), pfilt(1), ra_rgbe(1), |
221 |
|
|
rcalc(1), rpict(1), rtrace(1), vwrays(1), ximage(1) |