| 1 |
\ |
| 2 |
.\" This man page was generated by the Netpbm tool 'makeman' from HTML source. |
| 3 |
.\" Do not hand-hack it! If you have bug fixes or improvements, please find |
| 4 |
.\" the corresponding HTML page on the Netpbm website, generate a patch |
| 5 |
.\" against that, and send it to the Netpbm maintainer. |
| 6 |
.TH "Pnmpsnr User Manual" 1 "06 January 2018" "netpbm documentation" |
| 7 |
|
| 8 |
.SH NAME |
| 9 |
pnmpsnr - compute the difference between two images (the PSNR) |
| 10 |
|
| 11 |
.UN synopsis |
| 12 |
.SH SYNOPSIS |
| 13 |
|
| 14 |
\fBpnmpsnr\fP |
| 15 |
|
| 16 |
[\fIpnmfile1\fP] |
| 17 |
|
| 18 |
[\fIpnmfile2\fP] |
| 19 |
|
| 20 |
[\fB-rgb\fP] |
| 21 |
[\fB-machine\fP] |
| 22 |
[\fB-max=\fP\fIn\fP] |
| 23 |
[\fB-target=\fP\fIn\fP] |
| 24 |
[\fB-target1=\fP\fIn\fP] |
| 25 |
[\fB-target2=\fP\fIn\fP] |
| 26 |
[\fB-target3=\fP\fIn\fP] |
| 27 |
.PP |
| 28 |
Minimum unique abbreviations of options are acceptable. You may use |
| 29 |
double hyphens instead of single hyphen to denote options. You may use |
| 30 |
white space in place of the equals sign to separate an option name |
| 31 |
from its value. |
| 32 |
|
| 33 |
|
| 34 |
.UN description |
| 35 |
.SH DESCRIPTION |
| 36 |
.PP |
| 37 |
This program is part of |
| 38 |
.BR "Netpbm" (1)\c |
| 39 |
\&. |
| 40 |
.PP |
| 41 |
\fBpnmpsnr\fP reads two PBM, PGM, or PPM files, or PAM equivalents, as |
| 42 |
input and computes the magnitude of difference between the two images as a peak |
| 43 |
signal-to-noise ratio (PSNR) This metric is typically used in image |
| 44 |
compression papers to rate the distortion between original and decoded image. |
| 45 |
.PP |
| 46 |
\fBpnmpsnr\fP either prints these values or compares them to thresholds |
| 47 |
you specify. |
| 48 |
|
| 49 |
.PP |
| 50 |
The PSNR of a given component is the ratio of the maximum mean square |
| 51 |
difference of component values that could exist between the two images (a |
| 52 |
measure of the information content in an image) to the actual mean square |
| 53 |
difference for the two subject images. It is expressed as a decibel value. |
| 54 |
.PP |
| 55 |
The mean square difference of a component for two images is the |
| 56 |
mean square difference of the component value, comparing each pixel |
| 57 |
with the pixel in the same position of the other image. For the |
| 58 |
purposes of this computation, components are normalized to the scale |
| 59 |
[0..1]. |
| 60 |
.PP |
| 61 |
The maximum mean square difference is identically 1. |
| 62 |
.PP |
| 63 |
So the higher the PSNR, the closer the images are. A luminance |
| 64 |
PSNR of 20 means the mean square difference of the luminances of the |
| 65 |
pixels is 100 times less than the maximum possible difference, |
| 66 |
i.e. 0.01. |
| 67 |
.PP |
| 68 |
Note that the word "peak" is a misnomer; there is no maximum involved; the |
| 69 |
metric is a mean. But "peak signal to noise ratio" is for some reason the |
| 70 |
common term for this measurement. |
| 71 |
.PP |
| 72 |
If the inputs are PBM or PGM, \fBpnmpsnr\fP computes the PSNR of the |
| 73 |
luminance only. Otherwise, it computes three separate PSNRs: either the |
| 74 |
luminance, and chrominance (Cb and Cr) components of the colors or the |
| 75 |
red, green, and blue components. |
| 76 |
.PP |
| 77 |
By default, the program prints the PSNRs to Standard Output in |
| 78 |
human-friendly form. |
| 79 |
.PP |
| 80 |
With the \fB-machine\fP option, the program prints the PSNRs, but |
| 81 |
in machine-friendly form. |
| 82 |
.PP |
| 83 |
With a \fB-target\fP[\fIx\fP] option, the program just prints |
| 84 |
\&'match' or 'nomatch', depending on whether the PSNRs |
| 85 |
exceed targets you specify. |
| 86 |
|
| 87 |
.PP |
| 88 |
\fBpnmpsnr\fP reports the PSNR either in human-friendly form or in |
| 89 |
machine-friendly form (see \fB-machine\fP). |
| 90 |
|
| 91 |
|
| 92 |
.UN options |
| 93 |
.SH OPTIONS |
| 94 |
.PP |
| 95 |
In addition to the options common to all programs based on libnetpbm |
| 96 |
(most notably \fB-quiet\fP, see |
| 97 |
.UR index.html#commonoptions |
| 98 |
Common Options |
| 99 |
.UE |
| 100 |
\&), \fBpnmpsnr\fP recognizes the following |
| 101 |
command line options: |
| 102 |
|
| 103 |
|
| 104 |
.TP |
| 105 |
\fB-rgb\fP |
| 106 |
This option causes \fBpnmpsnr\fP to compare the red, green, and blue |
| 107 |
components of the color rather than the luminance and chrominance components. |
| 108 |
It has no effect on a monotone image. |
| 109 |
.sp |
| 110 |
This option was new in Netpbm 10.71 (June 2015). |
| 111 |
|
| 112 |
.TP |
| 113 |
\fB-machine\fP |
| 114 |
This option causes \fBpnmpsnr\fP to report the PSNRs in machine-friendly |
| 115 |
form, so another program can easily use the information. |
| 116 |
.sp |
| 117 |
The output is a single line. It contains one floating point decimal number |
| 118 |
for each color component, with a single space between every two. (This means |
| 119 |
there are either 1 or 3 numbers). For the YCbCr color space (no \fB-rgb\fP), |
| 120 |
they are in the order Y, Cb, Cr. For the RGB color space (\fB-rgb\fP), they |
| 121 |
are in R, G, B order. For a monotone image, there is one number. |
| 122 |
.sp |
| 123 |
Where the component does not differ between the images, so the PSNR is |
| 124 |
infinite, the number is \fBinf\fP |
| 125 |
.sp |
| 126 |
But note that the number displayed is also modified by the effect of |
| 127 |
\fB-max\fP. In particular, with \fB-max\fP, you will never see \fBinf\fP. |
| 128 |
.sp |
| 129 |
This option has no effect when you also specify \fB-target\fP[\fIn\fP]. |
| 130 |
.sp |
| 131 |
This option was new in Netpbm 10.74 (March 2016). |
| 132 |
|
| 133 |
.TP |
| 134 |
\fB-max=\fP\fIn\fP |
| 135 |
This is meaningful only with \fB-machine\fP. |
| 136 |
.sp |
| 137 |
It specifies the maximum number \fBpnmpsnr\fP will print as a PSNR. |
| 138 |
If the PSNR is greater than \fIn\fP, \fBpnmpsnr\fP just prints \fIn\fP. |
| 139 |
\fIn\fP is a decimal floating point number. An infinite PSNR is considered |
| 140 |
greater than any number. |
| 141 |
.sp |
| 142 |
This is mainly useful to deal with infinite PSNRs. It is often much more |
| 143 |
convenient to have a program process only numbers than to make it deal with |
| 144 |
infinity, and often a very large number has the same effect on a program as |
| 145 |
infinity. |
| 146 |
.sp |
| 147 |
Note that the output is logarithmic, which means you will not see really |
| 148 |
large but finite numbers. If you specify \fB-max=1000\fP, the only way you |
| 149 |
will see 1000 in the output is if the PSNR is really infinite. Two images |
| 150 |
with as many pixels as there are electrons in the universe, differing in only |
| 151 |
one pixel, and only in the smallest amount representable in the Netpbm format, |
| 152 |
have a PSNR less than 1000. |
| 153 |
.sp |
| 154 |
This option was new in Netpbm 10.74 (March 2016). |
| 155 |
|
| 156 |
.TP |
| 157 |
\fB-target\fP=\fIn\fP |
| 158 |
|
| 159 |
This option causes \fBpnmpsnr\fP to run in comparison mode - rather than |
| 160 |
print the PSNRs, it just tells you whether the PSNRs exceed |
| 161 |
\fIn\fP (a floating point number), i.e. whether the compared images are the |
| 162 |
same within a given margin of error. If all the computed PSNRs (luminance for |
| 163 |
a PBM or PGM; luminance and chrominance or red, green, and blue for PPM) |
| 164 |
exceed \fIn\fP, the program prints 'match' to Standard Output. |
| 165 |
Otherwise, it prints 'nomatch'. |
| 166 |
.sp |
| 167 |
If you also specify any of \fB-target1\fP, \fB-target2\fP, or |
| 168 |
\fB-target3\fP, and the images are color, \fBpnmpsnr\fP ignores |
| 169 |
\fB-target\fP. |
| 170 |
.sp |
| 171 |
This is mainly useful for use in a program. If you're |
| 172 |
running \fBpnmpsnr\fP manually, you could just run \fBpnmpsnr\fP |
| 173 |
without \fB-target\fP and compare the PSNRs to your targets yourself. |
| 174 |
.sp |
| 175 |
This option was new in Netpbm 10.82 (March 2018). |
| 176 |
|
| 177 |
.TP |
| 178 |
\fB-target\fP{\fB1\fP,\fB2\fP,\fB3\fP}=\fIn\fP |
| 179 |
Like \fB-target\fP, these options cause \fBpnmpsnr\fP to run in comparison |
| 180 |
mode. But they provide separate targets for the individual color component |
| 181 |
PSNRs. \fB-target1\fP, \fBtarget-2\fP, and \fB-target3\fP are for either |
| 182 |
the Y, Cb, and Cr components, respectively, or the red, green, and blue |
| 183 |
components, respectively, depending upon whether you specified \fB-rgb\fP. |
| 184 |
.sp |
| 185 |
If you don't specify the corresponding \fB-target\fP\fIn\fP option for a |
| 186 |
component, \fBpnmpsnr\fP ignores the PSNR of that component in deciding |
| 187 |
whether the images match. |
| 188 |
.sp |
| 189 |
If the image is a PBM or PGM, these options have no effect, except that it |
| 190 |
stilll selects comparison mode, so if you don't \fIalso\fP |
| 191 |
specify \fB-target\fP, and the image is PBM or PGM, the program fails. |
| 192 |
.sp |
| 193 |
Note that the options are defined so that you could code a |
| 194 |
\fBpnmpsnr\fP command in a program that works on both color and monotone |
| 195 |
images, specifying individual PSNR targets for use on the color images and the |
| 196 |
single target for use on the monotone images. |
| 197 |
.sp |
| 198 |
These options were new in Netpbm 10.82 (March 2018). |
| 199 |
|
| 200 |
|
| 201 |
|
| 202 |
|
| 203 |
.UN seealso |
| 204 |
.SH SEE ALSO |
| 205 |
.BR "pnm" (1)\c |
| 206 |
\& |
| 207 |
.SH DOCUMENT SOURCE |
| 208 |
This manual page was generated by the Netpbm tool 'makeman' from HTML |
| 209 |
source. The master documentation is at |
| 210 |
.IP |
| 211 |
.B http://netpbm.sourceforge.net/doc/pnmpsnr.html |
| 212 |
.PP |