SYNOPSIS

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

DESCRIPTION

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

COMMAND-LINE OPTIONS

img2sixel has the following command-line options:

-o, --outfile

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.

-p COLORS, --colors=COLORS

specify number of colors to reduce the image to (default=256).

-m FILE, --mapfile=FILE

transform image colors to match this set of colorsspecify map.

-e, --monochrome

output monochrome sixel image. this option assumes the terminal background color is black.

-k, --insecure

allow to connect to SSL sites without certs (enabled only when configured with --with-libcurl)

-i, --invert

assume the terminal background color is white. make sense only when -e option is given.

-I, --high-color

output 15bpp sixel image

-u, --use-macro

use DECDMAC and DEVINVM sequences to optimize GIF animation rendering.

-n MACRONO, --macro-number=MACRONO

specify an number argument for DECDMAC and make terminal memorize SIXEL image. No image is shown if this option is specified

-C COMPLEXIONSCORE, --complexion-score=COMPLEXIONSCORE

specify an number argument for the score of complexion correction. COMPLEXIONSCORE must be 1 or more.

-g, --ignore-delay

render GIF animation without delay.

-S, --static

render animated GIF as a static image.

-d DIFFUSIONTYPE, --diffusion=DIFFUSIONTYPE

choose diffusion method which used with color reduction.
DIFFUSIONTYPE is one of them:
auto -> choose diffusion type automatically (default)
none -> do not diffusion
fs -> Floyd-Steinberg method
atkinson -> Bill Atkinson's method
jajuni -> Jarvis, Judice & Ninke
stucki -> Stucki's method
burkes -> Burkes' method

-f FINDTYPE, --find-largest=FINDTYPE

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

-s SELECTTYPE, --select-color=SELECTTYPE

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

-c REGION, --crop=REGION

crop source image to fit the specified geometry.
REGION should be formatted as '%dx%d+%d+%d'.

-w WIDTH, --width=WIDTH

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

-h HEIGHT, --height=HEIGHT

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

-r RESAMPLINGTYPE, --resampling=RESAMPLINGTYPE

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

-q QUALITYMODE, --quality=QUALITYMODE

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

-l LOOPMODE, --loop-control=LOOPMODE

select loop control mode for GIF animation.
auto -> honer the setting of GIF header (default)
force -> always enable loop
disable -> always disable loop

-t PALETTETYPE, --palette-type=PALETTETYPE

select palette color space type.
auto -> choose palette type automatically (default)
hls -> use HLS color space
rgb -> use RGB color space

-b BUILTINPALETTE, --builtin-palette=BUILTINPALETTE

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

-E ENCODEPOLICY, --encode-policy=ENCODEPOLICY

select encoding policy
auto -> choose encoding policy automatically (default)
fast -> encode as fast as possible
size -> encode to as small sixel sequence as possible

-B BGCOLOR, --bgcolor=BGCOLOR

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

-P, --penetrate

penetrate GNU Screen using DCS pass-through sequence.

-D, --pipe-mode

read source images from stdin continuously.

-v, --verbose

show debugging info.

-V, --version

show version and license info.

-H, --help

print help.

Image loaders

img2sixel includes two or more image decoder components.

stb_image

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-bit-per-channel
   TGA (not sure what subset, if a subset)
   BMP non-1bpp, non-RLE
   PSD (composited view only, no extra channels)
   GIF/Animated GIF
   PIC (Softimage PIC)
   PNM (PPM and PGM binary only)

Limitations:
   no 16-bit-per-channel PNG
   no 12-bit-per-channel JPEG
   no JPEGs with arithmetic coding / JPEG 2000
   no 1-bit BMP

libpng

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

libjpeg

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

gdk-pixbuf2

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

GD

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

libsixel

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

HISTORY

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).

AUTHORS

img2sixel is maintained by Hayaki Saito. We imported whole code of stb_image v2.02, written by Sean Barrett and its contributers, for loading various images, and imported some code from pnmquant.c (netpbm library) for image quantization.

COPYRIGHT

Copyright (c) 2014-2015 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.

THANKS

This software derives from the following implementations.

stb_image-v2.02

This software includes stb_image-v2.02 (stb_image.h), a public domain JPEG/PNG reader.
https://github.com/nothings/stb

pnmquant.c (netpbm library)

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.

sixel 2014-3-2

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.

REFERENCES

resize.c (ImageMagick)

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

CONTRIBUTORS

Araki Ken (@arakiken)
Markus Elfring (@elfring)
isaki (@isaki68k)
NOKUBI Takatsugu (@knok)
Yasuhiro MATSUMOTO (@mattn)
Izumi Tsutsui (@tsutsui)
Iwamoto Kouichi (@ttdoda)
haru (@uobikiemukot)
Vertis Sidus (@vrtsds)
Bruce Mitchener (@waywardmonkeys)
Kazuhiro YOSHIKAWA (@yoshikaw)