pamfunc - Man Page

Apply a simple monadic arithmetic function to a Netpbm image

Examples (TL;DR)

Apply the specified arithmetic function with n as the second argument to each sample in the specified PAM image: pamfunc -multiplier|divisor|adder|subtractor|min|max n path/to/input.pam > path/to/output.pam
Apply the specified bit string function with n as the second argument to each sample in the specified PAM image: pamfunc -andmask|ormask|xormask|shiftleft|shiftright n path/to/input.pam > path/to/output.pam

Synopsis

pamfunc { -multiplier=realnum | -divisor=realnum | -adder=integer | -subtractor=integer | -min=wholenum | -max=wholenum -andmask=hexmask -ormask=hexmask -xormask=hexmask -not -shiftleft=count -shiftright=count [-changemaxval] } [filespec]

All options can be abbreviated to their shortest unique prefix. You may use two hyphens instead of one. You may separate an option name and its value with white space instead of an equals sign.

pamfunc reads a Netpbm image as input and produces a Netpbm image as output, with the same format, and dimensions as the input. pamfunc applies a simple transfer function to each sample in the input to generate the corresponding sample in the output. The options determine what function.

The samples involved are PAM samples. If the input is PBM, PGM, or PPM, the output will be that same format, but pamfunc applies the functions to the PAM equivalent samples, yielding PAM equivalent samples. This can be nonintuitive in the PBM case.

pamarith is the same thing for binary functions -- it takes two images as input and applies a specified simple arithmetic function (e.g. addition) on pairs of samples from the two to produce the single output image.

Values

The functions fall into two categories: arithmetic (such as multiply by 5) and bit string (such as and with 01001000). For the arithmetic functions, the function arguments and results are the fraction that a sample is of the maxval, i.e. normal interpretation of PAM tuples. But for the bit string functions, the value is the bit string whose value as a binary cipher is the sample value, and the maxval indicates the width of the bit string.

Arithmetic functions

The arithmetic functions are those selected by the options -multiplier, -divisor, -adder, -subtractor, -min, and -max.

As an example, consider an image with maxval 100 and a sample value of 10 and a function of "multiply by 5." The argument to the function is 10/100 (0.1) and the result is 5 * 0.1 = 0.5. In the simplest case, the maxval of the output is also 100, so the output sample value is 0.5 * 100 = 50. As you can see, we could just talk about the sample values themselves instead of these fractions and get the same result (10 * 5 = 50), but we don't.

Where it makes a practical difference whether we consider the values to be the fraction of the maxval or the sample value alone is where pamfunc uses a different maxval in the output image than it finds in the input image. See -changemaxval.

So remember in reading the descriptions below that the values are 0.1 and 0.5 in this example, not 10 and 50. All arguments and results are in the range [0,1].

Bit string functions

The bit string functions are those selected by the options -andmask, -ormask, -xormask, -not, -shiftleft, and -shiftright.

With these functions, the maxval has a very different meaning than in normal Netpbm images: it tells how wide (how many bits) the bit string is. The maxval must be a full binary count (a power of two minus one, such as 0xff) and the number of ones in it is the width of the bit string.

As an example, consider an image with maxval 15 and a sample value of 5 and a function of "and with 0100". The argument to the function is 0101 and the result is 0100.

In this example, it doesn't make any practical difference what we consider the width of the string to be, as long as it is at least 3. If the maxval were 255, the result would be the same. But with a bit shift operation, it matters. Consider shifting left by 2 bits. In the example, where the input value is 0101, the result is 0100. But if the maxval were 255, the result would be 00010100.

For a masking function, the mask value you specify must not have more significant bits than the width indicated by the maxval.

For a shifting operation, the shift count you specify must not be greater than the width indicated by the maxval.

PBM Oddness

If you're familiar with the PBM format, you may find pamfunc's operation on PBM images to be nonintuitive. Because in PBM black is represented as 1 and white as 0 (1.0 and 0.0 normlized), you might be expecting adding 1 to white to yield black.

But the PBM format is irrelevant, because pamfunc operates on the numbers found in the PAM equivalent (see above). In a PAM black and white image, black is 0 and white is 1 (0.0 and 1.0 normalized). So white plus 1 (clipped to the maximum of 1.0) is white.

Options

In addition to the options common to all programs based on libnetpbm (most notably -quiet, see Common Options ), pamfunc recognizes the following command line options:

-multiplier=realnum

This option makes the transfer function that of multiplying by
realnum. realnum must be nonnegative. If the result
is greater than one, it is clipped to one.

Where the input is a PGM or PPM image, this has the effect of
dimming or brightening it. For a different kind of brightening,
see pambrighten(1) and ppmflash(1)

Also, see ppmdim(1), which does the same
thing as pamfunc -multiplier on a PPM image with a multiplier
between zero and one, except it uses integer arithmetic, so it may be
faster.

And ppmfade(1) can generate a whole
sequence of images of brightness declining to black or increasing to
white, if that's what you want.

-divisor=realnum

This option makes the transfer function that of dividing by
realnum. realnum must be nonnegative. If the result
is greater than one, it is clipped to one.

This is the same function as you would get with -multiplier,
specifying the multiplicative inverse of realnum.

-adder=integer

This option makes the transfer function that of adding
integer/maxval. If the result is greater than one, it is
clipped to one. If it is less than zero, it is clipped to zero.

Note that in mathematics, this entity is called an "addend,"
and an "adder" is a snake. We use "adder" because
it makes more sense.

-subtractor=integer

This option makes the transfer function that of subtracting
integer/maxval. If the result is greater than one, it is
clipped to one. If it is less than zero, it is clipped to zero.

Note that in mathematics, this entity is called a
"subtrahend" rather than a "subtractor." We use
"subtractor" because it makes more sense.

This is the same function as you would get with -adder,
specifying the negative of integer.

-min=wholenum

This option makes the transfer function that of taking the maximum of
the argument and wholenum/maxval. I.e the minimum value in
the output will be wholenum/maxval.

If wholenum/maxval is greater than one, though, every value
in the output will be one.

-max=wholenum

This option makes the transfer function that of taking the minimum of
the argument and wholenum/maxval. I.e the maximum value in
the output will be wholenum/maxval.

If wholenum/maxval is greater than one, the function is
idempotent -- the output is identical to the input.

-andmask=hexmask

This option makes the transfer function that of bitwise anding
with hexmask.

hexmask is in hexadecimal. Example: 0f