IMG2SIXEL(1) General Commands Manual IMG2SIXEL(1)

img2sixel - image converter to DEC SIXEL graphics

img2sixel [ -options ] imagefiles
img2sixel [ -options ] < imagefiles

img2sixel converts various images into high quality DEC SIXEL image format.

img2sixel has the following command-line options:

specify output file name (default:stdout).
-7, --7bit-mode
generate a sixel image for 7bit terminals or printers (default).
-8, --8bit-mode
generate a sixel image for 8bit terminals or printers.
limit arguments of DECGRI('!') to 255.
specify number of colors to reduce the image to (default=256).
transform image colors to match this set of colorsspecify map.
output monochrome sixel image. this option assumes the terminal background color is black.
allow to connect to SSL sites without certs (enabled only when configured with --with-libcurl)
assume the terminal background color is white. make sense only when -e option is given.
output 15bpp sixel image
use DECDMAC and DEVINVM sequences to optimize GIF animation rendering.
specify an number argument for DECDMAC and make terminal memorize SIXEL image. No image is shown if this option is specified
specify an number argument for the score of complexion correction. COMPLEXIONSCORE must be 1 or more.
render GIF animation without delay.
render animated GIF as a static image.
choose diffusion method which used with color reduction.
DIFFUSIONTYPE is one of them:
auto -> choose diffusion type automatically (default)
none -> do not diffuse
fs -> Floyd-Steinberg method
atkinson -> Bill Atkinson's method
jajuni -> Jarvis, Judice & Ninke
stucki -> Stucki's method
burkes -> Burkes' method
a_dither -> positionally stable arithmetic dither
x_dither -> positionally stable arithmetic xor based dither
choose method for finding the largest dimension of median cut boxes for splitting, make sense only when -p option (color reduction) is specified.
FINDTYPE is one of them:
auto -> choose finding method automatically (default)
norm -> simply comparing the range in RGB space
lum -> transforming into luminosities before the comparison
choose the method for selecting representative color from each median-cut box, make sense only when -p option (color reduction) is specified.
SELECTTYPE is one of them:
auto -> choose selecting method automatically (default)
center -> choose the center of the box
average -> calculate the color average into the box
histogram -> similar with average but considers color histogram
crop source image to fit the specified geometry.
REGION should be formatted as '%dx%d+%d+%d'.
resize image to specified width.
WIDTH is represented by the following syntax:
auto -> preserving aspect ratio (default)
<number>% -> scale width with given percentage
<number> -> scale width with pixel counts
<number>px -> scale width with pixel counts
resize image to specified height.
HEIGHT is represented by the following syntax
auto -> preserving aspect ratio (default)
<number>% -> scale height with given percentage
<number> -> scale height with pixel counts
<number>px -> scale height with pixel counts
choose resampling method used with -w or -h option (scaling).
RESAMPLINGTYPE is one of them:
nearest -> Nearest-Neighbor method
gaussian -> Gaussian filter
hanning -> Hanning filter
hamming -> Hamming filter
bilinear -> Bilinear filter (default)
welsh -> Welsh filter
bicubic -> Bicubic filter
lanczos2 -> Lanczos-2 filter
lanczos3 -> Lanczos-3 filter
lanczos4 -> Lanczos-4 filter
select quality of color quanlization.
QUALITYMODE is one of them:
auto -> decide quality mode automatically (default)
high -> high quality and low speed mode
low -> low quality and high speed mode
full -> quality and careful speed mode
select loop control mode for GIF animation.
auto -> honer the setting of GIF header (default)
force -> always enable loop
disable -> always disable loop
select palette color space type.
auto -> choose palette type automatically (default)
hls -> use HLS color space
rgb -> use RGB color space
select built-in palette type
xterm16 -> X default 16 color map
xterm256 -> X default 256 color map
vt340mono -> VT340 monochrome map
vt340color -> VT340 color map
select encoding policy
auto -> choose encoding policy automatically (default)
fast -> encode as fast as possible
size -> encode to as small sixel sequence as possible

specify background color
BGCOLOR is represented by the following syntax
#rgb
#rrggbb
#rrrgggbbb
#rrrrggggbbbb
rgb:r/g/b
rgb:rr/gg/bb
rgb:rrr/ggg/bbb
rgb:rrrr/gggg/bbbb
penetrate GNU Screen using DCS pass-through sequence.
[[deprecated]] read source images from stdin continuously.
show debugging info.
show version and license info.
print help.

img2sixel has the following command-line options:


specify background color.
overrided by -B(--bgcolor) option.
represented by the following syntax:
#rgb
#rrggbb
#rrrgggbbb
#rrrrggggbbbb
rgb:r/g/b
rgb:rr/gg/bb
rgb:rrr/ggg/bbb
rgb:rrrr/gggg/bbbb

specify number of colors to reduce the image to (default=256).
overrided by -p(--colors) option.

img2sixel includes two or more image decoder components.

libsixel includes stb_image, a public domain image loader.
img2sixel uses it as default built-in image decoder. It can decode almost all images. but a few images can not be decoded by its limitations.

Supported source formats:
JPEG baseline & progressive (12 bpc/arithmetic not supported, same as stock IJG lib)
PNG 1/2/4/8/16-bit-per-channel
TGA (not sure what subset, if a subset)
BMP non-1bpp, non-RLE
PSD (composited view only, no extra channels)
PIC (Softimage PIC)
PNM (PPM and PGM binary only)

Limitations:
no 12-bit-per-channel JPEG
no JPEGs with arithmetic coding / JPEG 2000
no 1-bit BMP


If libpng library is linked at compile time, img2sixel uses it for decoding PNG image.


If libjpeg library is linked at compile time, img2sixel uses it for decoding JPEG image.


If gdk-pixbuf2 library is linked at compile time, img2sixel uses it automatically in some cases.


If GD library is linked at compile time, img2sixel uses it automatically in some cases.


img2sixel can load SIXEL as source image format, because it uses libsixel as a SIXEL image decoder.

Former SIXEL encoders(such as ppmtosixel) are mainly designed for dot-matrix printers. They minimize the amount of printer-head movement distance. But nowadays this method did not represent the best performance for displaying sixel data on terminal emulators. Encoded SIXEL data for VT-2xx/VT-3xx terminals were found in 80's Usenet, But the technology of how to create them seems to be lost.

kmiya's sixel(kmiya,2014) introduces an efficient encoding method which is re-designed for terminal emulators to optimize the overhead of transporting SIXEL with keeping compatibility with former SIXEL terminal. Now libsixel and ImageMagick's sixel coder follow it.

Araki Ken, known as the maintainer of mlterm, proposed the method for more compressed SIXEL encoding. Now libsixel adopted that method. Araki Ken describes about the way to generate high quality SIXEL.

See http://mlterm.sourceforge.net/libsixel.pdf(in Japanese).

sixel(5) sixel2png(1)

img2sixel is maintained by Hayaki Saito. We imported whole code of stb_image v2.12, written by Sean Barrett and its contributers, for loading various images, and imported some code from pnmquant.c (netpbm library) for image quantization.

Copyright (c) 2014-2016 Hayaki Saito

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

This software derives from the following implementations.

This software includes stb_image-v2.12 (stb_image.h), a public domain JPEG/PNG reader.
https://github.com/nothings/stb
The implementation of median cut algorithm for color quantization in quant.c is imported from pnmcolormap included in netpbm library.
http://netpbm.sourceforge.net/
pnmcolormap was derived from ppmquant, originally by Jef Poskanzer.

Copyright (C) 1989, 1991 by Jef Poskanzer.
Copyright (C) 2001 by Bryan Henderson.
Permission to use, copy, modify, and distribute this software and its documentation for any purpose and without fee is hereby granted, provided that the above copyright notice appear in all copies and that both that copyright notice and this permission notice appear in supporting documentation. This software is provided "as is" without express or implied warranty.

some part of converters/loader.c are derived from kmiya's "sixel" original version (2014-3-2).
http://nanno.dip.jp/softlib/man/rlogin/sixel.tar.gz
It is written by kmiya@culti.
He distributes it under very permissive license which permits useing, copying, modification, redistribution, and all other public activities without any restrictions.
He declares this is compatible with MIT/BSD/GPL.

We added some resampling filters in reference to the line-up of filters of MagickCore's resize.c.

http://www.imagemagick.org/api/MagickCore/resize_8c_source.html

Araki Ken (@arakiken)
Markus Elfring (@elfring)
Akinori Hattori (@hattya)
isaki (@isaki68k)
NOKUBI Takatsugu (@knok)
Yasuhiro MATSUMOTO (@mattn)
Masami HIRATA(@msmhrt)
OBATA Akio (@obache)
Izumi Tsutsui (@tsutsui)
Iwamoto Kouichi (@ttdoda)
haru (@uobikiemukot)
Vertis Sidus (@vrtsds)
Bruce Mitchener (@waywardmonkeys)
Kazuhiro YOSHIKAWA (@yoshikaw)
Turenar <sora@turenar.xyz>
Yusuke Endoh <mame@ruby-lang.org>
mattn <mattn.jp@gmail.com>
Akinori Hattori <hattya@gentoo.org>
Øyvind Kolås <pippin@gimp.org>
Henri Salo (@fgeek)
hongxu (@HongxuChen)
pwd (@YourButterfly)
Nicholas Luedtke (@nluedtke)
cool-tomato (@cool-tomato)

Send bug-reports, fixes, enhancements to saitoha@me.com
Aug 2016 4th Berkeley Distribution