Graphics Plus

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

/ Graphics Plus / Graphics Plus.iso / msdos / animutil / pcmovie / pcmovied.exe < prev

Wrap

Text File | 1994-01-10 | 51.2 KB | 648 lines

Producing Animations on IBM PC's and PS 2's by Viktor K. Decyk, Physics Department and Joan Slottow, OAC A. Introduction The ability to animate graphics on personal computers is an important visualization tool which can lead to new scientific insights and discoveries. Much of science involves dynamics in time and the human eye is a marvelous tool for recognizing patterns in time. When animating data, one sees details and patterns one never notices in the still pictures. Animation also allows a researcher to examine more data more quickly. To use animation as a discovery tool, it need not be fancy, just simple and easy to do. It is very handy to carry these movies around on diskettes and pop them into a PC to show someone. If needed, a videotape can be used to show the animation to a larger audience, such as a conference session. Such a videotape can easily be made of a computer animation using the equipment in UCLA's Visualization Demonstration Laboratory. Viktor Decyk, of the UCLA plasma physics simulation group, has created a library of Fortran subroutines that produce a file of compressed raster images, which may then be sent to a personal computer for rapid playback. Joan Slottow of OAC has added routines that allow existing raster images to be compressed and has added compressed raster image for the IBM PC to the output formats that can be produced by the CGM to Raster program, thus allowing graphics created by DISSPLA and SAS/GRAPH to be displayed as both stills and movies on IBM PCs. The software for playing these movies is written by the National Energy Supercomputer Center (NERSC) at Lawrence Livermore National Laboratory and currently runs on IBM PCs (Ref. 1,2). Versions for Suns and Macintoshes will be available soon. The IBM PC version supports three formats: 4-color CGA (320x200 pixels), monochrome EGA (640x350 pixels), and 256-color VGA (320x200 pixels). We have given this program to the MIC for distribution along with the documentation from NERSC. There are several ways to create the compressed movie files: - One way is to use the portable Fortran graphics library described in a previous Perspective article (Ref. 3). This library contains nine general-purpose graphics subroutines, six of which are line drawing routines and three of which are contouring routines. - Alternatively, you could run DISSPLA or SAS/GRAPH, both of which are large graphics packages, and create your graphics in a CGM metafile. You could then run the program CGM to Raster to create the compressed movie file. CGM to Raster was described in (Ref. 4) and an update on it appears in this issue of Perspective. - If you happen to have raw raster images from another source, these can also be compressed into compressed movie files. - Finally, you can create a compressed movie file by rasterizing a file of Tektronix commands. Again, this allows you to use another software package to create the images, since many graphics packages, such as DISSPLA, can create Tektronix files as output (Ref. 5). The compressed movie file is sent from whatever computer it was created on (in this case the IBM 3090) to your PC for playback, using either Y-Term (which can be slow) or, if your PC is on the network, FTP. In either case you must use binary transfer mode. There is no limit to the length of the movie files you can create, because the software from NERSC that reads the movie files and displays the movies reads the frames from disk as it displays them. However, you might want to limit yourself to movies which will fit on a diskette, since that makes them easily transportable. When reading a compressed movie file from a hard disk, speeds of 18, 12, and 6 frames per second are typical for EGA, CGA, and VGA respectively, with slower speeds when reading directly from diskette. B. Calling the Routines in Viktor Decyk's Fortran Graphics Library This library of subroutines was initially developed for use in an interactive plasma simulation program. It includes six X-Y plotting subroutines: GRAF2, GRAF1, GRAF3, DISP, DISP1, and DISP2, and three contour-plotting subroutines: CONTUR, DISPCN, and DSPCN2. The subroutine library, as written by Viktor, supports numerous graphics output devices. The subroutine library as installed at OAC includes only IBM PC compressed movie output. Viktor distributes the entire library to interested parties. See Reference 3. The subroutines and their arguments are: CALL GRAF2( y, labely, x, labelx, n, chr, ncr, irc ) Produces a line plot of Y versus X, which fills the maximum area of the display region. CALL GRAF1( y, labely, ymax, ymin, x, labelx, xmax, mxim, n, chr, ncr, irc ) Produces a scattergram (point plot) of Y versus X with the maximum and minimum values of the display region given by YMAX, YMIN, XMAX, and XMIN. CALL GRAF3( f, labely, x, labelx, n, m, nv, chr, ncr, irc ) Produces M line plots of subarrays of F versus X. Each plot contains N points. The total graphic fills the maximum area of the display region. CALL DISP( f, label, xmin, xmax, n, m, nv, isc, chr, ncr, irc ) Produces M line plots of subarrays of F versus a linear function in x, where XMIN < x < XMAX. A common scale given by YMAX = 2**ISC, YMIN = -2**ISC is used for the Y-axis. If ABS(ISC) > 116, the program finds the minimum value of ISC which will contain the plots. CALL DISP1( f, label, xmin, xmax, n, m, nv, isc, ist, chr, ncr, irc ) DISP1 is similar to DISP except that it includes an additional parameter IST. If IST > 0, then YMIN = 0 is used for the common scale, while if IST < 0, then YMAX = 0 is used for the common scale. If IST = 0, then DISP1 functions exactly as DISP. CALL DISP2( f, label, xv, xmin, xmax, n, m, nv, isc, chr, ncr, irc ) DISP2 is similar to DISP, except that only two subarrays are plotted, the first as a line plot and the second as a set of small closed circles. In addition, a vertical line at location x = XV is drawn. CALL CONTUR( f, link, label, n, m, nc, nv, chr, ncr, irc ) Produces a contour plot of the function F, for values of the first index in the range 1 < i < N, and of the second index 1 < j < M. NC contour intervals are chosen, spaced equally between the maximum and minimum values of F. CALL DISPCN( f, link, label, n, m, nc, nv, isc, chr, ncr, irc ) DISPCN is similar to CONTUR, except that the NC contour intervals are chosen equally spaced between FMAX = 2**ISC and FMIN = -2**ISC. If ABS(ISC) > 116, then the program finds the minimum value of ISC which will contain the function values. CALL DSPCN2( f, link, label, xv, yv, n, m, nc, nv, isc, chr, ncr, irc ) DSPCN2 is similar to DISPCN, except that a pair of vertical and horizontal lines are drawn at x = XV and y = YV, where the units are 0 < x < N and 0 < y < M. Where: Y, X, and F are real arrays of dimension: y( n ), x( n ), f( nv, m ) LINK is a logical*1 array of dimension: link( 2, n, m ) LABELY, LABELX, and LABEL are character*20 labels. CHR is a character*128 label, of which NCR characters are to be printed on the plot. When NCR = 0, none of the characters in CHR are printed. IRC is always returned to you as 0. Before the first plot, your program must initiate graphics with: CALL GOPEN In addition, a common block: COMMON /PLOTCM/ IPLOT, NPLOT, IDTYPE needs to be declared at the beginning of the main program, and you should set the variable NPLOT to the desired number of plots per frame of the movie. IPLOT is used to control the position of each individual plot in the frame, and is not usually set by the user. After the last plot, your program must terminate the graphics with: CALL GCLOSE Also, you must open a file on Fortran Logical Unit Number 19 to contain the compressed movie. This must be allocated as a card-image dataset. You can use a DD statement such as the following to define this dataset: //GO.ddname DD DSN=aaaaiii.name,UNIT=DATA,DISP=(NEW,CATLG), // DCB=CARD,SPACE=(TRK,(p,s),RLSE) By default, the output is generated in the IBM PC VGA (256-color) format. To produce output in either the 4-color CGA or for monochrome EGA format, you have to provide an input parameter on SYSIN as follows: Format Input Parameter CGA 1 EGA 2 VGA 3 For example, to get CGA mode code: //GO.SYSIN DD * 1 Each execution of a call to one of the graphics subroutines produces a complete plot. It takes NPLOT calls to the subroutines to fill one frame of the movie. Example 1: This sample program makes a simple test movie of a traveling sine wave. The 250-frame movie created by this test program occupies 553K bytes in 4-color CGA mode, 667K bytes in monochrome EGA mode, and 757K bytes in 256 color VGA mode. PROGRAM MOVIET PARAMETER(NX=128) COMMON /PLOTCM/ IPLOT, NPLOT, IDTYPE CHARACTER*128 CHR CHARACTER*20 CHP DIMENSION F(NX) 91 FORMAT (8H TIME = ,F7.2) DATA NLOOP,NCURVE,ISC,NCR /250,1,0,15/ DATA FREQ,DT /1.,.2/ DATA CHP /' POTENTIAL'/ C OPEN OUTPUT DEVICE FOR RASTER FILE OPEN(UNIT=19,FILE='MGRAPH',FORM='FORMATTED',STATUS='UNKNOWN') C CALCULATE CONTANTS DNX = 6.2831853/FLOAT(NX) XMIN = 0. XMAX = FLOAT(NX - 1) C NPLOT = NUMBER OF GRAPHS PER PAGE NPLOT = 1 C OPEN GRAPHICS CALL GOPEN C MAIN TIME LOOP DO 20 I = 1, NLOOP TIME = DT*FLOAT(I - 1) WT = FREQ*TIME C CALCULATE NEW FUNCTION VALUES DO 10 J = 1, NX F(J) = COS(2.*DNX*FLOAT(J - 1) - WT) 10 CONTINUE C DISPLAY FUNCTION WRITE (CHR,91) TIME CALL DISP (F,CHP,XMIN,XMAX,NX,NCURVE,NX,ISC,CHR,NCR,IRC) IF (IRC.EQ.1) GO TO 30 20 CONTINUE C CLOSE GRAPHICS 30 CALL GCLOSE STOP END C. Creating a Movie from images in a CGM Clear Text Metafile The program CGM to Raster has been updated to allow you to specify CGA, EGA, or VGA as one of the output file types. See the article on CGM to Raster also appearing in this issue. D. Creating a Movie from Raw Raster Images You can call Subroutines in the Fortran Graphics library that create a compressed movie file from raw raster images. Such a movie can contain any number of frames. You can create a compressed movie file out of a single image for display on an IBM PC and then tell the Movie software from NERSC to single-step through the movie. You will see the image displayed until you press Return. The NERSC software allows you to control the speed at which movies are displayed so that you can play them slow or fast. Your raw-raster image should be placed in a CHARACTER*1 matrix: raster( x,y ) Where x is the number of columns in the image and y is the number of rows. The raw-raster image must be of one of these three types: Type Description 1-bit Each pixel is represented in the low-order 1 bit of each character (byte) the 7 high-order bits are unused. A low-order 0 bit represents the background color. A low-order 1 bit represents the drawing color. 2-bits Each pixel is represented in the low-order 2 bits of each character (byte). The other 6 bits are unused. 00 represents the background color. 01, 10, and 11 are the drawing colors. 8-bits All the bytes of each pixel are used, giving 256 colors from 0 through 255. An 8-bit raster can also be accompanied by a color lookup table (CLUT). If the PC display mode with which the movie is to be displayed has at least as many bits as the raw-raster images, the images are simply compressed into the movie file. If, however, the PC display mode with which the movie is to be displayed has fewer bits than the raw-raster images, the images are filtered to reduce the number of colors. IBM PC Format Number of bits CGA 2 EGA 1 VGA 8 For example, an 8-bit raster image has to be filtered to be displayed on a CGA display; a 2-bit raster image does not. Four filters are: Filter Description 0 For EGA, the low-order 1 bit of each byte is used. For CGA, the low-order 2 bits of each byte are used. 1 All 0 pixels are mapped to the background color. All non-0 pixels are mapped to the drawing color. 2 Thresholding is used. For EGA, all pixel values less than (max-min)/2 are mapped to the background color. All other pixel values are mapped to the drawing color. Where max is the largest pixel value in the raster image and min is the smallest. For CGA, all pixel values falling within the lowest fourth are mapped to the background color, pixel values in the next fourth are mapped to magenta, in the next fourth to cyan, and in the highest fourth to white. 3 Dithering is used. Dithering is a process by which colors are replaced by dots. For a monochrome image, areas of darker color are replaced by densely clustered dots and areas of light color are replaced by sparse dots. For EGA, white dots on a black background result from the dithering process. For CGA, the result of dithering depends upon whether a CLUT is supplied by the user or if the default CLUT is used. Each pixel value is looked up on the CLUT to determine its red, green, and blue component. These are dithered and combined into a resultant picture with the following mapping: red is mapped to magenta green is mapped to cyan blue is mapped to cyan yellow is mapped to white magenta is mapped to magenta cyan is mapped to cyan white is mapped to white black is mapped to the background The default CLUT maps pixels values as follows: 0 black 1 blue 2 green 3 cyan 4 red 5 yellow 6 magenta 7 white 8-255 black Filters 0 and 1 work well when the raw-raster image contains a line drawing. Filters 2 and 3 work well for colored in images. The following CALLs are required to create a compressed movie file: CALL GMOPEN( iout, clut, ncolors ) Initiates the graphics. CALL COMPRG( raster, nx, ny, ixdim, nbits, ifilter, iflip ) Compresses a single raw-raster image as a frame of the movie. CALL GCLOSE Terminates the graphics. Where: IOUT should be set to 1 for CGA, 2 for EGA, or 3 for VGA output. CLUT is a CHARACTER*1 color lookup table of 0 to 256 entries. If you are not supplying a color lookup table, you must still provide some sort of dummy argument here. If you are supplying a color lookup table, dimension it as 3 times the number of entries (768 max), initialize it as follows: CLUT(1) = red value for the first entry, CLUT(2) = green value for the first entry, CLUT(3) = blue value for the first entry, CLUT(4) = red value for the second entry, etc. Each value should range form 0 (no color) to 255 (maximum color). NCOLORS should be set to the number of entries in CLUT. Set NCOLORS = 0 if you are not supplying a color lookup table (and are supplying a dummy argument instead). RASTER is the CHARACTER*1 matrix of dimension: raster( ixdim, ny ) containing the raw-raster image. IXDIM, NY are the dimensions of RASTER NX is the number of columns to select from the raster image. See the discussion of the raster image size below. NBITS is the number of significant bits in the raw-raster image: 1, 2, or 8. IFILTER is the filter you want to be used. If the output format, EGA, CGA, or VGA has at least as many bits as the raw-raster image (NBITS), IFILTER is ignored. IFLIP set IFLIP to 0 if the image is right side up in RASTER and to 1 if it is upside down. In addition, you must open a file on Fortran Logical Unit Number 19 to contain the compressed movie. This must be allocated as a card-image dataset. The maximum number of pixels that can be displayed on the IBM PC screen depends upon the display type: Type X Y EGA 640 350 CGA 320 200 VGA 320 200 Images smaller than the display size are padded with the background color. Images larger than the display size are truncated on the right and the bottom. For images larger that the display size, you can use the RASTER, NX, NY, and IXDIM parameters to select a subsection of the image to compress into the movie file. For example when a specific 500x387 image is displayed on VGA, only the upper right portion is visable. However, since both the left and the top of the image contain a considerable amount of the background color, it is desired that a different portion of the image be displayed. CALL COMPRG( raster, 320, 387, 500, nbits, ifilter, iflip ) This selects the leftmost 320 columns out of the 500 and is equivalent to what you would get by setting NX to 500 and allowing the image to be truncated. CALL COMPRG( raster(181,1), 320, 387, 500, nbits, ifilter, iflip ) This selects the rightmost 320 columns starting with column 181. CALL COMPRG(raster(181,188), 320, 200, 500, nbits, ifilter, iflip) This selects the rightmost 320 columns and the bottommost 200 rows. Example 2: This example reads in 30 raw raster images and a CLUT and generates a compressed movie file. program rmovie parameter( nx=259, ny=218 ) character*1 raster( nx, ny ) character*1 clut( 768 ) * *********************************** * OPEN OUTPUT DEVICE FOR RASTER FILE * *********************************** OPEN(UNIT=19,FILE='MGRAPH',FORM='FORMATTED',STATUS='UNKNOWN') * *********************************** * read the clut * *********************************** read( 1, 1 ) clut 1 format( 200a1, 56a1 ) CALL GMOPEN( 1, clut, 256 ) * *********************************** * read the raster image * *********************************** do 5 i = 1, 30 do 3 iy = 1, ny read( 2, 2 ) ( raster( ix,iy ), ix = 1, nx ) 2 format( 200a1, 200a1, 100a1 ) 3 continue * *********************************** * put out the image * *********************************** CALL COMPRG( raster, nx, ny, nx, 8, 1, 0 ) * *********************************** * bump the dataset sequence number * *********************************** read( 2, 2,end=5 ) ( raster( ix,iy ), ix = 1, nx ) 5 continue CALL GCLOSE STOP END E. JCL required to use the Fortran Graphics Library //aaaaiiic JOB job-parameters // EXEC FORTCG,LIBB='APP1.PCMOVIE.LOAD' (Fortran Program) //GO.ddname19 DD DISP=(NEW,CATLG),DSN=aaaaiii.cmf, // DCB=CARD,UNIT=DATA,SPACE=(TRK,(p,s),RLSE) //GO.ddname DD dd-parameters //GO.SYSIN DD * n Where: aaaaiii is your Userid. c is an optional character to make the jobname unique. ddname19 is your ddname for Fortran Logical Unit 19. aaaaiii.cmf is the dataset to contain the compressed movie file. p and s are the primary and secondary space requests for this dataset. //GO.ddname DD dd-parameters are any additional DD statements required by your program including those necessary to read in a color lookup table and raster images, if any. //GO.SYSIN DD * and n is required only when you are calling the graphics routines in the Fortran Graphics Library and want something other than the default VGA output. F. Creating a Movie from Textronix Files If you have already saved Textronix Files from some other source, Viktor will be happy to give you access to his software. G. Test Movie Player for the IBM 3179s You can preview the frames in your compressed movie file on the IBM 3179 terminals before transfering it to the PC. Logon to TSO and enter: PCMOVIE dsname H. References 1. Jean Shuler, "Kinetic Output for Viewing Your Graphics Files Dynamically on Your PC," NMFECC Buffer, Vol. 12, No. 6, p. 6 (1988). 2. MOVIE Document National Energy Supercomputer Center Lawrence Livermore National Laboratory 3. V. K. Decyk, "A Portable Fortran Graphics Library," UCLA-OAC Perspective, vol. 14, No. 2, p. 16 (1990). 4. OAC Writeup MVS: Visualizing Supercomputer Scientific Data on Macintoshes (SG22) 5. V. K. Decyk and L. Xu, "Transmission of Graphics by Electronic Mail," UCLA-OAC Perspective, Vol. 11, No. 4, p. 30 (1987).