| 1 |
.\" RCSid "$Id$" |
| 2 |
.TH RA_XYZE 1 5/30/96 RADIANCE |
| 3 |
.SH NAME |
| 4 |
ra_xyze - convert between RADIANCE RGBE and XYZE formats |
| 5 |
.SH SYNOPSIS |
| 6 |
.B ra_xyze |
| 7 |
[ |
| 8 |
.B \-r |
| 9 |
][ |
| 10 |
.B "\-e exposure" |
| 11 |
][ |
| 12 |
.B \-c |
| 13 |
| |
| 14 |
.B \-u |
| 15 |
][ |
| 16 |
.B "\-p xr yr xg yg xb yb xw yw" |
| 17 |
] |
| 18 |
[ |
| 19 |
.B input |
| 20 |
[ |
| 21 |
.B output |
| 22 |
] |
| 23 |
] |
| 24 |
.SH DESCRIPTION |
| 25 |
.I Ra_xyze |
| 26 |
converts between RADIANCE RGBE (red,green,blue,exponent) and XYZE |
| 27 |
(CIE X,Y,Z,exponent) formats. |
| 28 |
The |
| 29 |
.I \-e |
| 30 |
option specifies an exposure compensation, which may be given as |
| 31 |
a decimal multiplier or in f-stops (powers of two). |
| 32 |
By default, |
| 33 |
.I ra_xyze |
| 34 |
produces a flat XYZE RADIANCE picture file from any type of RADIANCE |
| 35 |
input picture. |
| 36 |
To override these defaults, the |
| 37 |
.I \-c |
| 38 |
option may be used to specify run-length encoded output, |
| 39 |
or the |
| 40 |
.I \-u |
| 41 |
option may be used to specify a flat output. |
| 42 |
.PP |
| 43 |
The |
| 44 |
.I \-r |
| 45 |
option causes |
| 46 |
.I ra_xyze |
| 47 |
to produce a run-length encoded RGBE file instead. |
| 48 |
The |
| 49 |
.I \-p |
| 50 |
option may be used to override the standard RADIANCE RGB primary |
| 51 |
colors to tailor the image for a particular output device or |
| 52 |
representation. |
| 53 |
The eight floating-point arguments to this option are the 1931 |
| 54 |
CIE (x,y) chromaticity coordinates of the three RGB primaries |
| 55 |
plus the white point, in that order. |
| 56 |
The new primaries will be recorded in the header of the output file, |
| 57 |
so that the original information may be fully recovered later. |
| 58 |
It is not necessary that the input file by in XYZE format. |
| 59 |
Th |
| 60 |
.I \-r |
| 61 |
option may therefore be used to convert from one RGB primary |
| 62 |
representation to another using the |
| 63 |
.I \-p |
| 64 |
option. |
| 65 |
.PP |
| 66 |
If the output file is missing, the standard output is used. |
| 67 |
If the input file is missing as well, the standard input is used. |
| 68 |
.SH NOTES |
| 69 |
The CIE standard used is the 1931 2-degree observer, and the |
| 70 |
correct output representation relies on the original RADIANCE |
| 71 |
input description being defined properly in terms of the standard |
| 72 |
RADIANCE RGB primaries, whose CIE (x,y) chromaticity values are |
| 73 |
defined in the header file in src/common/color.h. |
| 74 |
In this same file is a standard for the luminous efficacy of white |
| 75 |
light (WHTEFFICACY), which is used as a conversion between lumens |
| 76 |
and watts throughout RADIANCE. |
| 77 |
This same factor is applied by |
| 78 |
.I ra_xyze |
| 79 |
when converting between the radiometric units of the RGBE format |
| 80 |
and the photometric units of the XYZE format. |
| 81 |
The purpose of this factor is to ensure that the Y component of |
| 82 |
the CIE representation is luminance in units of candelas/meter^2. |
| 83 |
.PP |
| 84 |
Most of the RADIANCE picture filters should work uniformly on either |
| 85 |
RGBE or XYZE files, so it is not necessary to convert back to RGBE |
| 86 |
format except for conversion or display, in which case the correct |
| 87 |
primaries for the chosen output device should be specified with the |
| 88 |
.I \-p |
| 89 |
option if they are known. |
| 90 |
.SH EXAMPLES |
| 91 |
To convert RGBE output from |
| 92 |
.I rpict(1) |
| 93 |
into run-length encoded XYZE format: |
| 94 |
.IP "" .2i |
| 95 |
rpict [options] scene.oct | ra_xyze -c > scene_xyz.pic |
| 96 |
.PP |
| 97 |
To prepare a RADIANCE picture for display on a calibrated NTSC monitor: |
| 98 |
.IP "" .2i |
| 99 |
ra_xyze -r -p .670 .330 .210 .710 .140 .080 .333 .333 stand.pic ntsc.pic |
| 100 |
.SH AUTHOR |
| 101 |
Greg Ward |
| 102 |
.SH BUGS |
| 103 |
Any color correction applied to the original image is not removed |
| 104 |
or translated by |
| 105 |
.I ra_xyze, |
| 106 |
and it may result in color shifts in the output. |
| 107 |
If color preservation is important and the correction is unwanted, |
| 108 |
it is best to remove it with |
| 109 |
.I pfilt(1) |
| 110 |
using the |
| 111 |
.I \-er, |
| 112 |
.I \-eg |
| 113 |
and |
| 114 |
.I \-eb |
| 115 |
options first. |
| 116 |
(Simply look at the header and apply the reciprocal primaries of all |
| 117 |
COLORCORR= lines multiplied together.) |
| 118 |
Better still, get the picture before color correction is applied. |
| 119 |
.SH "SEE ALSO" |
| 120 |
pfilt(1), ra_rgbe(1), rpict(1) |
