To install VistaPro 3, GeoMorph and MakePath to your hard drive, just drag the entire drawer to where you'd like to install it. The drawer is only 5.5MB in size. Next month we'll place lots of DEMs in the graphics drawer. I've placed both the FPU and non FPU versions in this drawer with a tiny IconX script that detects which version you are running. You should have no problems with this so everyone should be able to click on the programs and have the appropriate version run. If it doesn't work for any reason, run the iconless required file manually, they are called; VistaPro3.0.fpu VistaPro3.0.nofpu MakePath.fpu MakePath.nofpu GeoMorph.fpu GeoMorph.nofpu Lyap2DEM.fpu Lyap2DEM.nofpu Covering my bases here. :) Mat. The VistaPro 3 changes file follows; <--- Cut ---> There have been several minor changes to VistaPro 3.0 since the manual went to press. These are detailed here for your convenience. 1> On the Stereo Control Panel, the single Displacement field has been replaced by a Camera Offset and an Image Offset field. The Camera Offset can be thought of in the same way as your eyes. The difference in eye location gives us stereo vision. Of course, the camera offset is exagerated as we are trying to make two 2-d images appear as 3-d. The Image offset spreads the images apart, forcing our eyes to focus properly for the stereo effect. 2> The description for the markers used in the Place Control Panel to mark the location of buildings, roads, trees and grass is incomplete. See the image "Place.iff" in your IFF drawer for a complete description of these markers. 3> On the Graphics Control Panel, an additional button has been added. This is the "No Flicker" button. This button will only be visible if you are running version 39 or later of the Graphics.library. This button regulates the Flicker Free mode in the DBLNTSC and DBLPAL monitor types. 4> Ham8 tends to be subject to artifacting if you use Pixel Dithering. Ham8 can display 256,000 colors out of a palette of over 16 million, so pixel dithering is not needed. The additional noise it adds to the image confuses the Ham coloring routine. This is controlled by the PixDth value in the Main lower control panel. This will be particularly important if you are creating animations with Ham8. 5> The Viewer program will now handle Ham8 and 256 color images as well as the earlier Amiga modes. 6> A new script command allows rendering of images up to 4000 by 4000 (memory permitting). VideoNoDisplay turns off the rendering of the image to chip RAM. Since all of the IQ scripts and the s:vistapro.prefs files are VistaPro scripts, this command could be added to them. ------------------------------------------------------------------------- Changes since Version 3.00 -------------------------- 3.01a If VistaPro.prefs was found in S:, a flag was set telling the script handler to ignore looking for a .list file, and not reset after the prefs were read. This resulted in aborted anims not being able to be continued. In 68000 mode, the increase in width of the control panel caused the topo map to move so that it was an odd number of bytes from the left edge of the screen. This resulted in an address exception when the topo was being initialized. Pine, Oak, Palm trees and grass were seeding the random number generator with the camera position, causing the trees to change how they were rendered as the camera position changed. This resulted in moving tree branches during animations. The tree_hidden routine improperly calculated the location of the space that the tree occupied so that at times trees that were partially behind hills would not render and would pop into full view long after they should have been visible. In HAM8 rendering, it was possible that the error sum could overflow the error table which allowed for 16 bit error sums. This was boosted to 32 bits to protect from this. In HAM8 rendering, the routine which chose the colors to use was using a full 6 bits of red, green and blue to choose the color register or adjustment to use, while the error table was built with 5 bits of red green and blue. This could cause severe artifacting in bright colors. The Interlace button was locked in the 'on' position after it was pressed. The "VideoNoDisplay" script could still open up the screen and window even though no image was being displayed. Load Region failed to update the panel gadgets. Actually, it did update the gadgets, but disable_gadgets() had been called, and the refresh that was done was not effective, because there were no enabled gadgets. The refresh was moved to past the enable_gadgets() call. When loading an extended scape, the lock_grass and lock_trees flag status was not being used to update the button position. The flag was set, but the button did not reflect the proper flag setting. Reworked the code dealing with Multiscan and Euro72 monitors once more. These monitor modes are quite convoluted, with flags changing bit positions when compared with the other modes. 3.02a Found bug in Range code. calc_dist() didn't return a value, yet the code for range checked the return to determine whether or not it would render the polygon. Forced a register choice at the beginning of each line in the HAM8 code. This eliminated artifacting at the left edge of the image. Still some problems with screen size vs image size that will trash image at times. Found a bug reported by Mark Manes whose eyesight is better than mine. When calculating the HAM8 palette, the color values were shifted one too few places to the left, resulting in a palette that was twice as dark as it should have been. I may argue, and disagree, but I still look. Got new integer determinant routine from John which fixes a possible overflow problem. Could cause things like topless trees. Got new branch shading code from John. Makes trees look more realistic. Please try this out in all monitor modes that you can, and at as many image sizes/display sizes that you can. Run "Vista >test.doc" and send me the doc file if anything strange occurs. 3.02b Modified texture code as requested by John. Lowered the FrDim to 100 from 150. This adds a shadow effect to the texture of the ground making the trees/under-trees area more realistic looking. Continued to fiddle with the graphics modes. So far this looks good. I have tried in several monitor modes and graphic modes and the images look alot better. No more garbaged images (at least the tests that previously messed up now work). Please test as many different modes as you can. There are over 9200 possible combinations which would take me months to render in order to minimally test everything, so try as many as you can so I can shake the last of the bugs out. Added the monitor panel back in. I didn't change the code attached to this panel, but now it seems to be working. I wish I knew what other changes I did to make it work, but... well there is this saying about a gift horse... 3.02c Discovered that rasters _must_ be mod(64) bits with AGA modes even though autodocs for WritePixelLine8() call for mod(16). This fixes problems with odd size screens, particularly in Super72 mode. Try running with Super72 mode and let me know how the control panels are scaled. I have had one report that the font is not correct, but am unable to duplicate the symptoms on my system. You must have Topaz 11 in your fonts: drawer. I need to find out if anyone else can duplicate this problem. 3.02d Trying 'generic' HAM8 algorithm which uses the widest spread of 256K colors out of 16 million rather than the best 32K colors. A little less banding in the sky, I think. The routine which chooses the best 256K out of 16 million colors is too expensive, requiring an additional 1024K in tables and about 45 seconds at the end of the render to calculate the color palette. Comparing images, you don't get much extra for the expense. Fixed HiRes and LoRes buttons for DCTV and HAME. HiRes is always forced on in these modes, and the numbered buttons should now reflect this as well. Removed the block on Interlace for DCTV and HAME. Please try these. I do not even know whether or not they are supposed to work, as I don't have any spec on either device. Try it, if it works, I'll leave it in, otherwise, I'll re-lock the Interlace button in these modes. 3.02e Fixed 'Show Render' palette selection. Modified default Cloud Density to 40 from 50. Fixed Load USGS DEM so that it re-initializes after each load. More tweaks to GRMode panel. Set all default extensions to "#?" and "*". Users can specify "#?.dem" etc., in their VistaPro.prefs file. 3.02f Fixed '1', '2', '4' and '8' keyboard keys so that Poly buttons don't get drawn if the Main lower panel is not up. 3.02g Increased size of region behind the ASL requester to include the entire screen so that the screen would not get trashed if the requester was moved. Forced default image size for DCTV to be same as screen size at the current interlace setting. Added specific OpenFont and SetFont to panel window instead of just naming the font in the NewScreen structure. Updated load and save extended for some stereo variables and sky and horizon intensity 3.02h More changes to extended to keep tracking with PC version. DCTV default screen size set to 768x482. ImageWidth and ImageHeight and aspect ratio initialized for detected monitor type. 3.02i Removed DCTV Filter flag. Some reports of lower quality DCTV images have been reported, this was the only change that I could see since Feb 8, 1992. Added a 'clear_status_line()' to the script processing, so that frame numbers don't leave a little garbage behind. Added some new fields to the Extended Header to support additions to the PC version of VP3. Added test for 'no_display' to routine which reallocates memory for screen bitmaps. Shouldn't ever get called if VideoNoDisplay is active. Found that HAM8 bit wasn't getting masked out of ModeID when saving HAM8 images to IFF. Same for 256 color images. 3.02j Fixed bug in save_iff that prevented the Hi-Res bit from being set in the CAMG flag for all non-AGA iff modes, including HAM-E and DCTV. Fixed problem in fade_color24 where value was not returned, but was expected and used. Added new Heading and Pitch code from John which fixes a problem caused by the new load/extended code he sent earlier. Fixed a problem where it was possible to have the DisplayWidth not match the actual width of the buffers in the bitmap. Fixed problem where the initial values for ImageHeight didn't match the value in the string gadget at startup. Improved the HAM8 register selection process. Removed previous fix for Exposure problem, as the resulting image quality was less for all other images. 3.02k Fixed bug where if Menu item for DEM Size Large or Huge was selected and there was not enough memory, it would switch to a lower size, but the menu item would remain checked. Fixed problem with Viewer where DCTV or HAME VANIMs wouldn't display correctly. Made the default width for Multiscan and Euro72 to be 640. The Lo-Res screen without Productivity set opens this wide. Added "Q" keyboard control which causes the image just rendered _not_ to be displayed, rather just writes to the bitmap wherever it may be. 3.02 final Fixed bug in Point. Mixup between CamZoom from PC and ICamZoom from Amiga. Fixed another Viewer bug. Same one as last version, but in a different place. Fixed bug in colormap of HAM8 Iff load/save. Caused by changing algorithm used to calculate HAM8 palette. 3.03a Fixed bug in hzline that caused totally black lines not to be rendered at all, using whatever color was leftover in the buffer instead. DBEQ loop was used where a DBRA was needed. Fixed problem where the use of DCTV or HAME cleared the Monitor ModeID without popping the button. When DCTV or HAME was then exited, the default NTSC was used regardless of what button was down. Fixed several button problems with Monitor modes. Made overscan buttons reflect the requirement of the width being an even multiple of 64 pixels in AGA modes. Added busy_pointer() call when exiting Graphics Panel until everything is set back up again. 3.03b Oops... Added a delay in maybe_front_to_back() which slowed the entire rendering process down drastically. Removed that delay and one in CheckAbort which could have similar effects. Fixed bug in Preview that caused it to ignore the arrow keys. Fixed bug where Display Width and Height gadgets weren't being refreshed after the script command VideoDisplaySize. Fixed bug in Viewer that caused it to skip every other frame. 3.03c Fixed bug where IFF anim mode did not save AGA modes correctly. Fixed bug where preview_script wrote outside of boundary of topo map. Added new code from John for "P" button handling, which allows for proper 3-d representation from very steep camera angles. 3.03 final