| 53 |
|
.I \-s |
| 54 |
|
options follow the last input matrix, output results will be transformed |
| 55 |
|
and/or scaled accordingly. |
| 56 |
< |
These operations are discussed in greater detail further on. |
| 56 |
> |
These operations are discussed in greater detail below. |
| 57 |
|
A single concatenation matrix may be applied after element operations |
| 58 |
|
using the |
| 59 |
|
.I \-m |
| 71 |
|
.fi |
| 72 |
|
.sp |
| 73 |
|
The number of components indicates that each matrix element is actually |
| 74 |
< |
composed of multiple elements, most commonly an RGB triple. |
| 74 |
> |
composed of multiple channels, most commonly an RGB triple. |
| 75 |
|
This is essentially dividing the matrix into planes, where each component |
| 76 |
|
participates in a separate calculation. |
| 77 |
|
If an appropriate header is not present, it may be added with a call to |
| 188 |
|
or |
| 189 |
|
.I \-f |
| 190 |
|
options are used to define a "co" variable or "co(p)" function, |
| 191 |
< |
this will be evaluated at each output |
| 192 |
< |
component for the current element. |
| 191 |
> |
which will be evaluated for each output |
| 192 |
> |
component from the current element. |
| 193 |
|
The "co" variable defines identical operations for all components, |
| 194 |
|
whereas "co(p)" may specify different operations for each component. |
| 195 |
|
The element position is defined |
| 281 |
|
.I pcomb, |
| 282 |
|
whose capabilities somewhat overlap. |
| 283 |
|
The former loads each matrix into memory before operations, |
| 284 |
< |
and element components take 8 bytes apiece, adding up quickly. |
| 284 |
> |
and element components are stored as double-precision. |
| 285 |
|
Very large matrices therefore present a problem with that tool. |
| 286 |
|
Furthermore, |
| 287 |
|
.I rmtxop |
| 314 |
|
.I rcomb |
| 315 |
|
should be preferred over |
| 316 |
|
.I rmtxop |
| 317 |
< |
for any operations in can handle, which is everything except |
| 317 |
> |
for any operations it can handle, which is everything except |
| 318 |
|
multiple matrix concatenations and transpose |
| 319 |
< |
operations, which are handled more efficiently by |
| 320 |
< |
.I rcollate(1) |
| 321 |
< |
in any case. |
| 319 |
> |
operations. |
| 320 |
> |
The latter may be handled more efficiently by |
| 321 |
> |
.I rcollate(1). |
| 322 |
|
That said, there is no significant difference for |
| 323 |
< |
simple operations on smallish matrices, and note that only |
| 323 |
> |
simple operations on small matrices, and only |
| 324 |
|
.I rmtxop |
| 325 |
|
and |
| 326 |
|
.I dctimestep(1) |
| 327 |
< |
currently accept XML files as inputs. |
| 328 |
< |
Also, the resizing function of |
| 327 |
> |
accept XML files as inputs. |
| 328 |
> |
Also note that the resizing function of |
| 329 |
|
.I pcomb |
| 330 |
|
is not supported in |
| 331 |
|
.I rcomb, |
