potrace(1) User Commands potrace(1)

NAME

potrace - transform bitmaps into vector graphics.

SYNOPSIS

potrace [options] [filename...]

DESCRIPTION

potrace is a tool for tracing a bitmap, which means, transforming a
bitmap into a smooth, scalable image. The input is a bitmap, which
means, a pixel-based image composed of the two colors black and white
only. The output is SVG, PDF, EPS, or one of a number of other vector
formats. A typical use is to create vector graphics from scanned
data, such as company or university logos, handwritten notes, etc.
The resulting image is not "jaggy" like a bitmap, but smooth. It can
then be rendered at any resolution.

potrace can read bitmaps in the following formats: PBM, PGM, PPM
(collectively known as PNM, see pnm(5)), as well as BMP (Windows and
OS/2 bitmap formats). The input image should only use the two colors
black and white. If other pixel values appear in the input, they will
be converted to black and white using a simple threshold method.

potrace can currently produce the following output formats: SVG, PDF,
EPS, PostScript, DXF, GeoJSON, PGM, Gimppath, and XFig. Additional
backends might be added in the future.

OPTIONS

The following options are supported. Dimensions (arguments of type
dim) can have optional units, e.g. 6.5in, 15cm, 100pt. The default
unit is inches (or centimeters, if this was configured at compile
time, see COMPILE TIME CONFIGURATION below). For pixel-based output
formats such as PGM, DXF, GeoJSON, and Gimppath, the default unit is
pixels.

General options:
-h, --help print help message and exit.

-v, --version print version info and exit. This also shows the
defaults that were compiled into this version of
potrace.

-l, --license print license info and exit.

Input/output options:
filename Each file can hold an input image, or multiple
concatenated input images. If filename arguments are
given, then potrace will by default create one output
file for each input filename given. The name of the
output file is obtained from the input filename by
changing its suffix according to the chosen backend.
If changing the suffix is impossible because the names
of the input and output files would be identical, then
the output filename is created by adding the "-out"
suffix to the name of the input file. If no filename
arguments are given, then potrace acts as a filter,
reading from standard input and writing to standard
output. A filename of "-" may be given to specify
reading from standard input.

-o filename, --output filename
write output to this file. All output is directed to
the specified file. If this option is used, then
multiple input filenames are only allowed for multi-
page backends (see BACKEND TYPES below). In this case,
each input file may contain one or more bitmaps, and
all the bitmaps from all the input files are processed
and the output concatenated into a single file. A
filename of "-" may be given to specify writing to
standard output.

-- End of options. Any remaining arguments are
interpreted as filenames. This also disables filter
mode, even if no filenames are given. This is useful
for shell scripts, because potrace -- $FILENAMES will
behave correctly even for an empty list of filenames.
However, -- with an empty list of filenames is not
permitted in conjunction with the -o option, because
this would generate a document of zero pages, which
none of the backends permit.

Backend selection:
For general information, see also BACKEND TYPES below.

-b name, --backend name
Select backend by name, where name is one of eps,
postscript, ps, pdf, pdfpage, svg, dxf, geojson, pgm,
gimppath, xfig. Backend names can be abbreviated by a
prefix as long as it is unambiguous. Backend names are
case insensitive.

-s, --svg, -b svg, --backend svg
SVG backend. The output is a Scalable Vector Graphics
(SVG) file. This is a single-page, variable-sized,
dimension-based backend. Note that unless the -r
option is given, the resolution of the input bitmap is
assumed to be 72dpi.

-b pdf, --backend pdf
PDF backend. The output is a file in the Portable
Document Format. If the input consists of multiple
bitmaps, they are each rendered on a separate page.
This is a multi-page, variable-sized, dimension-based
backend.

-b pdfpage, --backend pdfpage
The PDFPage backend is like the PDF backend, except
that it is fixed-size like the PostScript backend.

-e, --eps, -b eps, --backend eps
EPS backend (default). The output is an encapsulated
PostScript file. This is a single-page, variable-
sized, dimension-based backend.

-p, --postscript, -b ps, --backend ps
PostScript backend. The output is a PostScript file.
This is a multi-page, fixed-size, dimension-based
backend. If the input consists of multiple bitmaps,
they are each rendered on a separate page.

-b dxf, --backend dxf
DXF backend. The output is a file in the Drawing
Interchange Format (DXF). In this backend, all Bezier
curves are approximated by piecewise circular arcs;
this is suitable for processing in CAD software or for
machining applications using CNC tools. This is a
single-page, variable-sized, pixel-based backend. The
-u option has no effect for this backend.

-b geojson, --backend geojson
GeoJSON backend. The output is a file in the format
used by some applications processing geographical
data. In this backend, all Bezier curves are
approximated by 8 straight line segments. This is a
single-page, variable-sized, pixel-based backend. The
-u option has no effect for this backend.

-g, --pgm, -b pgm, --backend pgm
PGM backend. The output is a portable greymap (PGM)
file. It is a convenient backend for antialiasing a
bitmap image. This is a multi-page, variable-sized,
pixel-based backend. If the input consists of more
than one image, the images are concatenated in the
output.

-b gimppath, --backend gimppath
Gimppath backend. This backend produces output
suitable to be imported as a path by the GNU Image
Manipulation Program (Gimp) (in the Layers, Channels &
Paths dialog, select Paths, then right-click and
select Import Path). The output is actually an SVG
file. The differences to the SVG backend are: the
--opaque option has no effect, the --flat option is
always on, and the dimensions are pixel-based. This is
a single-page, variable-sized, pixel-based backend.

-b xfig, --backend xfig
XFig backend. This is a single-page, fixed-size,
dimension-based backend. The output is a file in the
XFig format. Note that XFig uses X-splines instead of
Bezier curves, thus it is not possible to translate
the output of potrace into the XFig format with
absolute accuracy. This backend does a reasonable
approximation using two control points for each Bezier
curve segment. The -u option has no effect for this
backend, because control points are always rounded to
the nearest 1/1200 of an inch in XFig. Curve
optimization is disabled. Implies --opaque.

Algorithm options:
For more detailed information on these options, see TECHNICAL
DOCUMENTATION below.

-z policy, --turnpolicy policy
specify how to resolve ambiguities in path
decomposition. Must be one of black, white, right,
left, minority, majority, or random. Default is
minority. Turn policies can be abbreviated by an
unambiguous prefix, e.g., one can specify min instead
of minority.

-t n, --turdsize n
suppress speckles of up to this many pixels.

-a n, --alphamax n
set the corner threshold parameter. The default value
is 1. The smaller this value, the more sharp corners
will be produced. If this parameter is 0, then no
smoothing will be performed and the output is a
polygon. If this parameter is greater than 4/3, then
all corners are suppressed and the output is
completely smooth.

-n, --longcurve
turn off curve optimization. Normally potrace tries to
join adjacent Bezier curve segments when this is
possible. This option disables this behavior,
resulting in a larger file size.

-O n, --opttolerance n
set the curve optimization tolerance. The default
value is 0.2. Larger values allow more consecutive
Bezier curve segments to be joined together in a
single segment, at the expense of accuracy.

-u n, --unit n set output quantization. Coordinates in the output are
rounded to 1/unit pixels. The default of 10 usually
gives good results. For some of the debug modes, a
value of 100 gives more accurate output. This option
has no effect for the XFig backend, which always
rasterizes to 1/1200 inch, or for the DXF backend. For
the GeoJSON backend, this option is only a hint; the
actual rounding may be more, but not less, accurate
than specified.

-d n, --debug n
produce debugging output of type n. This has different
effects for different backends. For the PostScript/EPS
backends, the values n=1,2,3 illustrate the
intermediate stages of the potrace algorithm.

Scaling and placement options:
-P format, --pagesize format
for fixed-size backends, set page size. The following
formats can be specified: A4, A3, A5, B5, Letter,
Legal, Tabloid, Statement, Executive, Folio, Quarto,
10x14. Format names are case insensitive. Also, an
argument of the form dimxdim is accepted to specify
arbitrary dimensions. The default page size is Letter
(or A4, if this was configured at compile time, see
COMPILE TIME CONFIGURATION below). Page format names
can be abbreviated by a prefix as long as it is
unambiguous. This option has no effect for variable-
sized backends.

-W dim, --width dim
set the width of output image (before any rotation and
margins). If only one of width and height is
specified, the other is adjusted accordingly so that
the aspect ratio is preserved.

-H dim, --height dim
set the height of output image. See -W for details.

-r n[xn], --resolution n[xn]
for dimension-based backends, set the resolution (in
dpi). One inch in the output image corresponds to this
many pixels in the input. Note that a larger value
results in a smaller output image. It is possible to
specify separate resolutions in the x and y directions
by giving an argument of the form nxn. For variable-
sized backends, the default resolution is 72dpi. For
fixed-size backends, there is no default resolution;
the image is by default scaled to fit on the page.
This option has no effect for pixel-based backends. If
-W or -H are specified, they take precedence.

-x n[xn], --scale n[xn]
for pixel-based backends, set the scaling factor. A
value greater than 1 enlarges the output, a value
between 0 and 1 makes the output smaller. The default
is 1. It is possible to specify separate scaling
factors for the x and y directions by giving an
argument of the form nxn. This option has no effect
for dimension-based backends. If -W or -H are
specified, they take precedence.

-S n, --stretch n
set the aspect ratio. A value greater than 1 means the
image will be stretched in the y direction. A value
between 0 and 1 means the image will be compressed in
the y direction.

-A angle, --rotate angle
set the rotation angle (in degrees). The output will
be rotated counterclockwise by this angle. This is
useful for compensating for images that were scanned
not quite upright.

-M dim, --margin dim
set all four margins. The effect and default value of
this option depend on the backend. For variable-sized
backends, the margins will simply be added around the
output image (or subtracted, in case of negative
margins). The default margin for these backends is 0.
For fixed-size backends, the margin settings can be
used to control the placement of the image on the
page. If only one of the left and right margin is
given, the image will be placed this distance from the
respective edge of the page, and similarly for top and
bottom. If margins are given on opposite sides, the
image is scaled to fit between these margins, unless
the scaling is already determined explicitly by one or
more of the -W, -H, -r, or -x options. By default,
fixed-size backends use a non-zero margin whose width
depends on the page size.

-L dim, --leftmargin dim
set the left margin. See -M for details.

-R dim, --rightmargin dim
set the right margin. See -M for details.

-T dim, --topmargin dim
set the top margin. See -M for details.

-B dim, --bottommargin dim
set the bottom margin. See -M for details.

--tight remove whitespace around the image before scaling and
margins are applied. If this option is given,
calculations of the width, height, and margins are
based on the actual vector outline, rather than on the
outer dimensions of the input pixmap, which is the
default. In particular, the --tight option can be used
to remove any existing margins from the input image.
See the file placement.pdf for a more detailed
illustration.

Color options:
These options are only supported by certain backends. The DXF and
GeoJSON backends do not support color.

-C #rrggbb, --color #rrggbb
set the foreground color of the output image. The
default is black.

--fillcolor #rrggbb
set the fill color of the output image, i.e., the
color of the "white" parts that are enclosed by
"black" parts. The default is to leave these parts
transparent. Implies --opaque. Please note that this
option sets the background color; to set the
foreground color, use --color instead.

--opaque fill in the white parts of the image opaquely, instead
of leaving them transparent. This only applies to
interior white parts, i.e., those that are enclosed
inside a black outline. Opaqueness is always in effect
for the XFig backend.

SVG options:
--group for SVG output, try to group related paths together.
Each path is grouped together with all paths that are
contained inside it, so that they can be moved around
as a unit with an SVG editor. This makes coloring
individual components slightly more cumbersome, and
thus it is not the default.

--flat for SVG output, put the entire image into a single
path. This makes it impossible to color the components
individually, and thus it is not the default. But the
resulting SVG file can be more easily imported by some
applications such as Gimp. In fact, the Gimppath
backend is a variation of the SVG backend with the
--flat option and pixel-based scaling. The --flat
option has no effect if --opaque has been selected.

PostScript/EPS/PDF options:
-c, --cleartext
do not compress the output. This option disables the
use of compression filters in the PostScript and PDF
output. In the PostScript backend, if -c and -q are
used together, the resulting output can be easily read
by other programs or even by humans.

-2, --level2 use PostScript level 2 compression (default). The
resulting file size is ca. 40% smaller than if the -c
option is used.

-3, --level3 use PostScript level 3 compression, if available. This
gives slightly smaller files than using -2, but the
resulting files may not print on older PostScript
level 2 printers. If support for PostScript level 3
compression has been disabled at compile time, a
warning message is printed and level 2 compression is
used instead.

-q, --longcoding
turn off optimized numerical coding in PostScript
output. Normally, potrace uses a very compact
numerical format to represent Bezier curves in
PostScript, taking advantage of existing redundancy in
the curve parameters. This option disables this
behavior, resulting in longer, but more readable
output (particularly if the -c option is also used).

PGM options:
-G n, --gamma n
set the gamma value for anti-aliasing (default is
2.2). Most computer displays do not render shades of
grey linearly, i.e., a grey value of 0.5 is not
displayed as being exactly half-way between black and
white. The gamma parameter corrects for this, and
therefore leads to nicer looking output. The default
value of 2.2 is appropriate for most normal CRT
displays.

Frontend options:
-k n, --blacklevel n
set the threshold level for converting input images to
bitmaps. The potrace algorithm expects a bitmap, thus
all pixels of the input images are converted to black
or white before processing begins. Pixels whose
brightness is less than n are converted to black, all
other pixels to white. Here n is a number between 0
and 1. One case is treated specially: if the input is
in an indexed color format with exactly 2 colors, then
the blacklevel is ignored and the darker of the two
colors is mapped to black.

Note: the method used by potrace for converting
greymaps to bitmaps is very crude; much better results
can be obtained if a separate program, such as
mkbitmap(1), is used for this purpose. In particular,
mkbitmap(1), which is distributed with potrace, has
the ability to scale and interpolate the image before
thresholding, which results in much better
preservation of detail.

-i, --invert invert the input bitmap before processing.

Progress bar options:
--progress display a progress bar for each bitmap that is
processed. This is useful for interactive use. The
default behavior is not to show any progress
information.

--tty mode set the terminal mode for progress bar rendering.
Possible values are "vt100", which requires a
vt100-compatible terminal, and "dumb", which uses only
ASCII characters. The default is system dependent.

BACKEND TYPES

Backends can be classified in several ways, which affects the
available command line options and their behavior:

Fixed-size or variable-sized:
For fixed-size backends, the size of the page is always the same
(for example Letter or A4, as specified at compile time or by
the -P option). By default, the image will be centered and
scaled to fit the page size. For variable-size backends, the
size of the page follows the size of the image. Currently the
PostScript (PS), PDFPage, and XFig backends are fixed-size, and
the remaining backends are variable-size.

Dimension-based or pixel-based:
In dimension-based backends, distances are measured in physical
units such as inches or centimeters. In pixel-based backends,
distances are measured in pixel units. The -r option only works
for dimension-based backends, and the -x option only works for
pixel-based backends. Currently, the DXF, PGM, Gimppath, and
GeoJSON backends are pixel-based, and the remaining backends are
dimension-based. Currently, all pixel-based backends are
variable-sized.

Single-page or multi-page:
Single-page backends can only accept a single image. Multi-page
backends can accept multiple images, typically one per page of
output. Currently, the PostScript (PS), PDF, PDFPage, and PGM
backends are multi-page, and the remaining backends are single-
page. Note that multiple input images can be read in two ways:
from multiple input files (with the -o option), or from a single
input file that holds several concatenated images.

COMPILE TIME CONFIGURATION

Certain aspects of the behavior of potrace can be configured at
compile time by passing the following options to the ./configure
script.

--disable-zlib
compile potrace without the zlib compression library. This means
PostScript level 3 compression will not be available.

--enable-metric
compile potrace with centimeters as the default unit instead of
inches.

--enable-a4
compile potrace with A4 as the default page size.

EXIT STATUS

The exit status is 0 on successful completion, 1 if the command line
was invalid, and 2 on any other error.

VERSION

1.16

AUTHOR

Peter Selinger <selinger at users.sourceforge.net>

Please see the file AUTHORS for a full list of other contributors.

TECHNICAL DOCUMENTATION

For a detailed technical description of the potrace algorithm, see
the file potrace.pdf, which is available from the potrace web site.
For information on the Potrace library API, see potracelib.pdf.

WEB SITE AND SUPPORT

The latest version of potrace is available from
http://potrace.sourceforge.net/. This site also contains a list of
frequently asked questions, as well as information on how to obtain
support.

SEE ALSO

mkbitmap(1)

COPYRIGHT

Copyright (C) 2001-2019 Peter Selinger

This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or (at
your option) any later version.

This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301, USA. See also http://www.gnu.org/.

Version 1.16 September 2019 potrace(1)