APDL Eductation Resources

home *** CD-ROM | disk | FTP | other *** search

/ APDL Eductation Resources / APDL Eductation Resources.iso / programs / graphics / gif2rpc / !Help next >

Wrap

Text File | 1996-01-21 | 22.2 KB | 875 lines

db od8o.o8bo 88[ ]8[]P8` 88[ YP 88[ Y8 ]88 88[ P` ]88 o88bodboo8 .o88bo d8[ .od,o8[.odbo88o .d88o ]8[ 88,""88 88[ .8P "888"88'Y88`'88, .8P ]8[ Y8[ 88` 88 88[ d8 88P ' ]8[ Y8b 88[ " '8bo8P 88 88[ .8` 88[ ]8[ ]88 88[ .d""` 88 88[ .8` 88[ ]8[ 88[ Y8[ ., ]8oooo,.o88b.o88bo.888888 o88bo ]88bd88 '88oo8 '888888 """" """"` """""P '"""` ]8['"" """ . """8 ]8[ 8[ .8 ]8b '88o8P` ]8888 A freeware GIF™ -> RISC O S sprite converter By Cy Booker <cy@cheepnis.demon.co.uk> Version 1.02 21-Jan-1996 ================================================================ gif2rpc ==== [1] Synopsis . . . . . . . . . . . . . . . . . . . . . 1 [1.1] License . . . . . . . . . . . . . . . . . . . . . . 1 [1.2] Trademarks . . . . . . . . . . . . . . . . . . . . 2 [2] How to use . . . . . . . . . . . . . . . . . . . . 2 [2.1] How to use with ArcWeb . . . . . . . . . . . . . . 2 [2.2] How to use with Webster . . . . . . . . . . . . . . 3 [2.3] Command line parameters . . . . . . . . . . . . . . 3 [2.4] Default options: gif2rpc$options . . . . . . . . . 8 [3] Output image format . . . . . . . . . . . . . . . . 9 [4] Why use gif2rpc . . . . . . . . . . . . . . . . . . 10 [4.1] Speed comparisons . . . . . . . . . . . . . . . . . 10 [5] Error codes . . . . . . . . . . . . . . . . . . . . 11 [6] History . . . . . . . . . . . . . . . . . . . . . . 12 ================================================================ gif2rpc ==== [1] Synopsis gif2rpc is a program for RISC O S (>= 3.10) that converts GIF image files to native sprite files. A GIF file is a 2…256 colour image, with an associated 24-bit palette and an optional “transparency” mask. This program was written mainly to allow the generation of 32K colour images that look as good as they can. This is because, despite what you intuitively feel, a 256 colour image displayed on a 32K colour screen does *not* look that good. [I mainly browse the World Wide Web in a 32K colour screen, and am generally disappointed with the way GIFs are displayed, and so wrote this program that provides a very good conversion of GIFs to 32K colour modes.] Since then it has extended to be a general image processing program. Note that gif2rpc is extremely lenient towards malformed GIF's (ie it won't crash if fed garbage). [1.1] License This software comes with NO WARANTEE. It is distributed in the hope that it's useful. The author will not be liable for any incurred damages, directly or indirectly due to the use or inability to use this software. It is FreeWare, and remains the copyright of the author. Portions of this code are copyright Stephen A. Bennet: * DECODE.C - An LZW decoder for GIF * Copyright (C) 1987, by Steven A. Bennett * * Permission is given by the author to freely redistribute * and include this code in any program as long as this * credit is given where due. The code to match a colour to the default wimp palette is based on a UseNet article: Message-ID: <12463@acorn.co.uk> Date: 30 Jan 92 12:20:23 GMT From: RWilson@acorn.co.uk Subject: How to find the closest colour out of the RISC OS 256 Page 1 ================================================================ gif2rpc ==== [1.2] Trademarks GIF is a trademark of Compuserve Inc. [2] How to use gif2rpc is a command line utility. It takes quite a few parameters, but basic usage is: gif2rpc <in> <out> If you would like a logfile of any errors that are encountered then append “ >>& logfile” to the command line, or use the -log parameter. Note that if you have he DDEutils module then gif2rpc will take advantage of the new facility for long command lines. [2.1] How to use with ArcWeb This applies to ArcWeb 1.14. First open the Browser configuration window, and enable the “GIF87 handled by gif2spr” option. Select OK. Don't forget to choose the menu item “Configure->Save options”! Next make a back up copy of the file “!ArcWeb.gif2spr”. Finally copy these files: gif2rpc -> !ArcWeb.gif2rpc WWW.gif2spr -> !ArcWeb.gif2spr WWW.gif2rpcFE -> !ArcWeb.gif2rpcFE And that is that. You may wish to modify the (new) “!ArcWeb.gif2spr” file to change the initialisation of the system variable “gif2rpc$options”. Page 2 ================================================================ gif2rpc ==== [2.2] How to use with Webster This applies to Webster v0.11. First open the Choices dialogue box, and select the “Inlne Viewers” tab. Ensure that the “GIF Decoder” value is left at the default of “<Webster$Dir>.gif2spr -q %0 %1”. Click the Save button. Next make a back up copy of the file “!Webster.gif2spr”. Finally copy the *contents* of the WWW directory into the !Webster directory, i.e. creating the three files: gif2rpc -> !Webster.gif2rpc WWW.gif2spr -> !Webster.gif2spr WWW.gif2rpcFE -> !Webster.gif2rpcFE And that is that. You may wish to modify the (new) “!Webster.gif2spr” file to change the initialisation of the system variable “gif2rpc$options”. [2.3] Command line parameters Here are all the parameters: [-gif] <file-name> this is the name of the GIF file. It does not need to be type’d as a TIFF file, as this program will verify that the file is in GIF format [-sprite] <file-name> this will be the file that is produced, containing the single sprite representing the original GIF image [-help] produce some terse help on these parameters [-verbose <n>] by default <n> is 0, which disables all text output while the GIF file is being read and processed. By setting <n> to 1, then some basic information is displayed. This can be redirected to a file by appending “> file” to the command line. By setting <n> to 2 then rather more information is displayed Page 3 ================================================================ gif2rpc ==== [-quiet] by default, if gif2rpc detects an error in the input GIF, then it will first display a textual message saying where and what failed, then attempt to continue (but see -fussy). This message can be redirected to a a file by appending “2> file” to the command line, or using -log (below). For a list of the error messages, see below [-hourglass] because some GIFs are big, they take a long time to process (even on an ARM 700). Therefore you may tell gif2rpc to use the “hourglass” to give some visual feedback over how far it has got in processing the GIF. The percentage indicator is used to show how much of the image has already been decoded After decoding, while the image is processed the percentage counts DOWN from 100% to 0%. This is an arbitrary aesthetic decision [-name <text>] by default the image saved in the -sprite file-name will be “gif2rpc”. If this offends you then you can rename it. Note that names of sprites must be lower case. gif2rpc enforces this by using the current locale’s idea of lower case. Also note that sprite names are limited to twelve characters. See also -autoname Page 4 ================================================================ gif2rpc ==== [-bpp <n>] bpp stands for bits per pixel. This parameter determines what sort of sprite image gif2rpc will try to produce. It takes the following values 0 use GIF’s palette to produce 1/2/4/8 bpp image. No image processing is done 1 produce a monochrome black/white image 2 produce a four colour image, using default wimp palette (a grey scale) 4 produce a 16 colour image, using the default wimp palette 8 produce a 256 colour image, using the default wimp palette 16 produce a 32K colour image, with a fixed palette 32 produce a 16M colour image -1 (this is the good one) it uses the current screen display to determine whether it will produce a 2, 4, 16, 256, 32K, or 16M colour image also, the current screen palette is used the default value is -1 [-accurate] when generating a 256 colour image, this will force the use of a more accurate colour matching algorithm that runs about 3-4 times slower [-force_66] when generating a 32K colour image, this will force the use of routines that use an internal resolution of 22 bits per red/green/blue. Normally only 16 bits are used. These 66 bit routines are slower [-square] a GIF file can specify an aspect ratio for an image. If a 32K or 16M colour image is produced then, by default, this aspect ratio information is embedded in the sprite. For <= 256 colour images it is not possible for RISC O S to be told of the aspect ratio. Using this switch will disable embedding this aspect ratio information Page 5 ================================================================ gif2rpc ==== [-autoname] the sprite in the spritefile can be give a name derived from the input GIF file with this option [stats] this will output a single text line to show how well (in terms of size) the conversion done, relative to keeping the 256 colour image The fields are "gif2rpc " "%s" input file name "(" "%4d" horizontal pixel size "x" "%4d" vertical pixel size ":" "%d" input bits per pixel "p" [if palette in screen] "s" [if palette is sorted] "." "i" [if interlaced] "p" [if palette in image] "s" [if palette is sorted] "t" [if transparent] ") -> " "%s" output file name "(" "%s" output colours ") " "%d" output size - gif2rpc output size " " "%f" total duration, in seconds " " "%f" percentage time spent filtering " " "%s" filter used [-log <file>] if present, this overrides any command line re- direction of stdout and stderr [-fussy] by default gif2rpc will allow several mistakes in a GIF file, and silently (if -quiet) ignore them. With -fussy, these errors will cause conversion to fail, and no sprite to be produced Page 6 ================================================================ gif2rpc ==== [-filter <name>] If -bpp is not 0 then gif2rpc will process the image to produce a sprite. These are the algorithms it will use (these are listed fastest to slowest) nearest: you do *not* want this. It is crap dither2x2: strongly recommended. It is only slightly slower than nearest, and produces good images (if >= 256 colours) sierra2_4a: very good, very fast. Personally, I think it looks better than floyd-steinberg floyd_steinberg: as used by ChangeFSI. An old favourite burkes: a 2-line filter as per floyd_steinberg, where the coefficients are quick to calculate sierra2: very similar to burkes, but poor quality sierra3: a 3-line filter stucki: another 3-line filter, but slow to calculate jarvis_judice_ninke: a 3-line filter which is very slow (22 times slower than dither2x2). It is questionable whether it produces better output Note that you can *not* use shortcuts for the switches (ie -v is not acceptable, use -verbose). Page 7 ================================================================ gif2rpc ==== [2.4] Default options: gif2rpc$options If the system variable “gif2rpc$Options” is defined, then it will be appended to any parameters given on the command line. This is particularly useful when using the logging facilities of the program. For example, using *Set gif2rpc$Options -stats -log $.log will cause all calls to gif2rpc to append statistics information to the file “$.log”. 2.5 Typical usage For no-hassle usage use: gif2rpc -quiet gif sprite For some reasonable information then use something like: gif2rpc -stats -log $.log gif sprite and the log file will be appended with any errors, and a single line giving some statistics of the conversion For “problem” gif's use: gif2rpc -verbose 2 -fussy gif sprite Page 8 ================================================================ gif2rpc ==== [3] Output image format You may think that -bpp 0 is the best option, as this will generate a sprite that contains all the information in the original GIF. But when you come to display that GIF, the OS has quite a bit of work to do before it can plonk the data on the screen. The miriad of options provided by gif2rpc are there to do a lot of this work once, and not every time the GIF is displayed. The main processing that gif2rpc does is to remove any explicit palette information, and alter the image so that a default (RISC O S) palette is used. This usually saves 2K per image: Size none 32K image 16M image 8x8 2168 184 312 16x16 2360 568 1080 32x32 3128 2104 4152 64x64 6200 8248 16440 1024x4 6200 8248 16440 And with a transparency mask Size none 32K image 16M image 8x8 2232 216 344 16x16 2612 632 1144 32x32 4152 2232 4280 64x64 10296 8760 16952 1024x4 10296 8760 16952 Note that the -stat file may be used to give statistics about the file size (and savings) encountered. Page 9 ================================================================ gif2rpc ==== [4] Why use gif2rpc There are two other programs out there that perform near identical functions: !ChangeFSI This program, unfortunately, does not seem to cope with “transparent” GIFs. Since these are the norm, rather than the exception on the WWW, it is bad news gif2spr This is a nice program, but has no support for generating 32K colour images You’d use gif2rpc if you usually view GIFs in 32K modes, or would like to convert GIFs to 16M colour sprites. It’s logging facility, and fuller diagnostics of errors may be useful to you. Also, its' ability to use other filter algorithms (faster ones like dither2x2, or slower ones) can be very useful. [4.1] Speed comparisons As a test bed, a 128 colour 640x480 GIF was used. It was converted to a sprite 10 times, and a script recorded the total time, accurate to +/- one second. It was ran in a 32K screen mode (1024x768), and a 32K image was generated: ChangeFSI dithered 36 (S16,90,90) gif2rpc floyd-steinberg 30 (-quiet, -bpp 16) ChangeFSI simple 24 (S16,90,90, -nodither) gif2spr 11 (-q) gif2rpc 2x2 dither 11 (-quiet, -bpp 16) For generating a 256 colour image in a 256 colour screen mode (1600x1200): ChangeFSI dither slow 67 (28r) ChangeFSI dither fast 59 (28) ChangeFSI simple slow 52 (28r, -nodither) ChangeFSI simple fast 46 (28, -nodither) gif2spr 11 (-q) gif2rpc 8 (-quiet, -bpp 8) Page 10 ================================================================ gif2rpc ==== [5] Error codes These are the error numbers that will be reported (unless you use the -quiet option) if the GIF is found to be corrupted: 10: the start of the file was not a valid GIF header 11: invalid lzw table size (not in range [2, 9]) 12: short file 13: unexpected record 14: corrupt graphic extension record size 15: there was no image in the file 16: image size invalid (pixel size less than one) 17: corrupt graphic extension record terminator 18: (internal error) processing mask 19: (internal error) processing image 20: no palette defined in screen or image 21: last row was incomplete 22: extra byte(s) after image found 23: bad end of image marker Page 11 ================================================================ gif2rpc ==== [6] History Version 1.02 21-Jan-1996 Better colour matching for 256 colour sprites Version 1.01 14-Jan-1996 Added -autoname feature Tidied up this documentation Wrote the gif2rpcFE kludge program Version 1.00 21-Sep-1995 This was never released, but was left languishing on the hard disc 7 Author Program and documentation by (but see copyright notice above): Cy Booker 86 Church View Main Road Crockenhill Swanley Kent BR8 8JW U.K. Internet: cy@cheepnis.demon.co.uk Arcade BBS (Fidonet#2:254/27.0): Cy Booker Musical entertainment provided by Frank Zappa