Documentation for the "render" Program

Overview

Command line

render inp out outsize [dimg] [pimg]

where inp is the input file list, out is out output file list, outsize is the name of your output file(s), and dimg and pimg are the names of your output distance and percent images respectively. These are each described below.

inp

This parameter is the list of your input files, separated by commas. The first file must be the elevation map (topography). The following file(s) are the corresponding image data set(s). Either one image file (for grey scale) or three image files (for color) may be given. The number of output files given must match the number of image files given here. All file are VICAR images, and may be either BYTE or HALF format. See examples.

out

This parameter specifies the output images to be computed. The same number of files must be specified as input image files (the elevation file does not count here). The output files will be VICAR images, unless the NOLABELS option (described below) is given.

outsize

This parameter gives the size of the output images to be computed. It is given as number of lines then number of smaples, separated by a comma. (Do not use parentheses). See examples.

dimg

This optional parameter is the name of the distance image to be output if desired. The distance image is a VICAR image in REAL (32 bit floating point) format, where each pixel represents the distance from the eye to the ground for that location.

pimg

This optional parameter specifies the name of the percent image to be output if desired. The percent image is a VICAR image in BYTE format, where each pixel represents the percentage of the pixel which is terrain (as opposed to sky, or background). For example, if 4 rays were cast for that pixel, and only three hit the terrain, the pixel value in the percent image would be 75.

render la.ele,la.bw frame.img 480,512 - renders a 480 line by 512 sample black and white frame.

render la.ele,la.red,la.grn,la.blu f.red,f.grn,f.blu 200,200 dist.img per.img - render a 200 square color frame, and also generates the distance and percent images.

Options

	setenv FOOBAR ON

To make the descriptions simpler, each variable is described as integer, float, string, or keyword. Integer variables may take only integers, float variables may take any real number, string variables may take strings (valid values will also be given) and keyword variables take nothing. For example:

		Command			Type
		setenv FOOBAR 3		integer
		setenv FOOBAR 4.55	float
		setenv FOOBAR ON	string
		setenv FOOBAR		keyword

Required Options

PARMFILE

(string) This option specifies the name of the parameter file for render to use. The parameter file will be either the file output from the VICAR program PATH2TEXT, which will convert your path file from 3DI, or output from the program surv2rend, which will convert your Surveyor flight path to a parmfile that render can read.

Each line in the parmfile contains the viewing parameters for a single frame, which is selected when rendering with the FRAME parameter. The parameters which appear on the line are (in order of appearance): AZLOOK, AZMUTH, DIST, ELEV, FOV, PITCH, XYRES, ZRES, SL, SS, NL, NS, ROLL. The SL, SS, NL, and NS parameters are integers and describe the SIZE field. All the other parameters are described below. SL, SS, NL, NS, and ROLL are optional.

FRAME

(integer) This option specifies which frame in the parmfile to render. The first frame is 1.

Recommended Options

FIELD

(string, valid values: EVEN, ODD, BOTH. Default: BOTH) Specifies which lines should be rendered. Using EVEN or ODD will result in a half-height image, with only even or odd lines rendered. This option is used when recording on video tape, where the odd and even lines are interlaced. When generating the parameter file with PATH2TEXT, generate twice as many frames (i.e., 60 frames per second instead of 30). Then render ODD fields for the odd frames, and EVEN fields for the even frames, and use the program weave to combine the two into a single frame for recording.

MINELEV

(integer, default 0) Specifies the minimum elevation to be found in the elevation data set, in pixel value (DN) units. Specifying MINELEV and MAXELEV can greatly speed up the rendering process.

MAXELEV

(integer, default 255 for byte data, 32767 for halfword data) Specifies the maximum elevation to be found in the elevation data set, in pixel value (DN) units. Specifying MINELEV and MAXELEV can greatly speed up the rendering process.

PYRAMID

(integer, default 1) Specifies the number of levels that are present in the image pyramid from which the terrain is to be rendered. Highly recommended for all but the smallest of data sets, because processing time is greatly reduced and rendering quality is greatly enhanced. A pyramid can be made with the makepyr command on the DIAL Unix machines.

RAYS

(integer) Specifies the number of rays to be cast per pixel. When generating a high quality still frame or animation, using multiple rays per pixel is highly recommended; it does, however, take longer to compute. When possible, 9 or 16 rays per pixel are recommended. For tests, use only one ray per pixel.

SIZE

(keyword, or set of 4 integers). When used as a keyword, SIZE specifies that the SIZE field in the parameter file should be ignored, and that the inputs should be mapped to virtual memory. When working with large data sets, this is recommended, and often necessary. When used as a set of four integers, SIZE specifies the area in the input from which data is to be fetched. This option is not currently recommended and will not be discussed further here.

Viewing Parameters

AZLOOK

(float) Specifies the clockwise angle in degrees from the center of the input image to which the camera view is rotated. If AZLOOK is 0.0 or 360.0, then the camera will look directly at the center of the image.

AZMUTH

(float) Specifies the clockwise angle in degrees around the center of the input image at which the camera position is located. Due north is 0.0, east is 90.0.

DIST

(float) Specifies the distance in pixels from the center of the input image of the camera position. If your input image was 100 lines, an AZMUTH of 0 and a DIST of 50 would place your camera in the center of the top line.

ELEV

(float) Specifies the elevation of the camera above 0 in the elevation image.

FOV

(float) Specifies the field of view of the camera. The wider of the image dimensions of the output frames will subtend this angle.

PITCH

(float) Specifies vertical angle of the camera view direction. A pitch of 0.0 will look horizontally at the horizon; a pitch of -90.0 will look straight down.

ROLL

(float) Specifies camera roll. The camera will be rolled ROLL degrees counter-clockwise about the view direction (the image will appear to rotate clockwise). A roll of 0.0 is upright, 180.0 is upside down. NOTE: because of the algorithm used to speed up the ray-casting, large values for roll (> 20 degrees) can sometimes develop artifacts, so check your frames carefully when using roll, especially if you have very tall (spiky) mountains and/or a very wide (> 120 degree) field of view.

XYRES

(float) Specifies a scaling to be applied to the input image in the plane of the image. This parameter is usually set to the resolution in meters (or some other unit) of the data.

ZRES

(float) Specifies a scaling to be applied to the elevation. For a vertical scale of 1 (no exaggeration), ZRES is normally set to the units of the pixels in the elevation image in meters (or some other unit). To double the vertical exaggeration, you can double ZRES.

Other Options

ASPECT

(float, default is calculated) Specifies the aspect ratio of the device you will use to display the final image, width:height. For NTSC video, ASPECT should be set to 1.33333, assuming you are filling your frame buffer. For example, for an Abekas video disk, you would use an output size of 486 lines by 720 samples, with ASPECT set to 1.3333. The image will then be warped so that after reshaping your image to the ASPECT ratio, you will get square pixels. So for the Abekas case just described, if you display the image on your workstation screen instead of an Abekas the image will look short and fat.

BACKGRND

(set of 3 integers, default 0,0,0 [black]) Specifies a color (R,G,B) for the background in the rendered scene. Any pixel which has no data (sky) will be set to this color, unless HAZE is also set.

BLUR

(keyword) If blur is turned on, stochastic sampling is used to apply motion blur to the animation. Each ray cast randomly assigned to one of the cells within its pixel on the view plane (each ray gets its own cell; that is, they are shuffled), and is jittered in time between the current frame and the next frame in the path file. Blur works best with many rays per pixel (RAYS) to obtain good results.

CAMERA

(string, valid values: ABOVE, BELOW. No set default.) This option specifies whether you want to render the top or bottom of the terrain data. If the camera is ABOVE, the top is rendered, and if the camera is BELOW, the bottom is rendered. By default, this parameter is determined by the actual position of the camera relative to the terrain.

EMINLEVEL

(integer, default 0) Specifies the minimum level in the pyramid of elevation images that is present. This option allows you to delete the highest resolution level(s) of your pyramid when they are not needed. For example, if your image data is 28.5 meter resolution, and your original elevation data was 100 meter resolution, part of your preprocessing would be to resample the elevation data to be the same resolution as the image data, making the file 3.5 times the resolution of the actual data. Once the image pyramid is created, you could remove the original image and set EMINLEVEL to 1 with no degradation in rendered image quality. You could also remove the next level of the pyramid and set EMINLEVEL to 2 with only slight degradation (probably not noticeable) in rendered image quality.

GAMMA

(float, default 1.0) Specifies that gamma correction be applied to the rendered image as it is rendered. Gamma correction is like a stretch which compensates for non-linearity in video monitor response. A typical monitor has a gamma of about 2.2. A gamma of 1.0 applies no correction. For most image data, gamma correction is not needed, because it is processed on monitors and the correction is generally built into the image.

HAZE

(set of 3 integers, default same color as BACKGRND) Specifies a color (R,G,B) for haze. If HAZEDIST is zero, the effect of HAZE is to provide a horizon (zero pitch) color, which then blends smoothly going up until at vertical (+90 deg. pitch) it is the BACKGRND color.

HAZEDIST

(float, default 0.0) Specifies a distance from the eye at which a horizontal ray (pitch=0) would become purely the HAZE color. The haze as a function of distance is modulated the the function of height; so as a ray looks up more (positive pitch), the effect of haze is less. Thus, the sky looks the same whether HAZEDIST is used or not. A value of zero turns off haze over the terrain.

JITTER

(float, default 0.5) JITTER randomly varies the position of a ray within its viewplane cell. A value of 0.0 would mean the ray is fixed in the center of the cell, and a value of 1.0 would mean the ray will be randomly cast somewhere within the cell. JITTER is used to replace some of the aliasing which occurs when rendering with some noise, to break up artifacts and make the aliasing less visible.

NOLABELS

(keyword) Specifies that the output image file(s) should not have any VICAR labels on them. The resultant images will be raw raster image files with no header.

NOSPREAD

(keyword) Specifies that when a ray intersects the surface, the pyramid level currently being used in the elevation be used for the image data directly. The default is to recalculate the level of detail for the image pyramid based on the angle of incidence with the surface, to allow some smoothing for anti-aliasing.

SMEAR

(float, default 1.0) Specifies a scaling to be applied to pyramid transition distances. A scaling of 2.0 causes lower resolution levels of the pyramid to be accessed at half the normal distance; 3.0 at a third, etc. For most intermediate distances, a value of SMEAR greater than 1.0 will cause the resultant image to look a little blurrier; but when close to the data the full resolution is still there (unsmeared).