miner@ulowell.UUCP (10/16/86)
I have been working with image processing routines on images loaded into my own window on a CUSTOM screen. For what I was doing in the past I could use the commands in the graphics library (and my own routines) for accessing this image memory. We have built a piece of image processing hardware that sits as a plug in card (PIC) in a Zorro expansion box, and performs DMA reads and writes to chip memory. I need to have a several open windows on a screen that have their own bitmaps so that the PIC can DMA-read my window as one contiguous bitmap. +-------------+ Normally the windows use the screen's memory and do not | +-----+ | have their own bitmaps. I have considered using super- | |win#1| | bitmapped windows. The benefit is that I get my own | |_____+---+ | bitmap made up of individual bitplanes whose size I can | |win#2| | control. The problem is that doing image processing | |_____+ | requires deep bitmaps - four to five bitplanes. When +-------------+ I open a super bitmap window 256 x 256 x n (0<n<5) on a hires interlaced screen it eats at least 128K (640 x 400 x 4) for Intuition to do house cleaning for my little 256 x 256 images. I cannot afford to waste this much chip RAM for each image buffer. It seems like a waste to use super-bit mapped windows anyway - they are really designed to handle bitmaps larger then the window - not what I need. I talked with the people at Westchester, they gave me new ideas but nothing I am completely happy with. Two other alternatives are to either create my own off-screen bitmap and then do a bit blit to rastport or to use the layers library. The creation of my own bitmap off screen would work out great, but I like the idea of seeing the PIC, or my routines, process the image while it is on the screen. I would have to process it off screen and then copy the result to the screen. The problem with the layers library is that I have to say goodbye to Intuition, something we all try to avoid. What I need is the ability to just point a window, which can be fixed in both size and location, to its own bitmap. Any ideas? Thanks, Rich Miner UUCP: ...wanginst!ulowell!miner ARPA: miner@ulowell.CSNET University of Lowell, Comp Sci Dept TALK: (617) 452-5000 x2693 Lowell MA 01854 USA HAL hears the 9000 series is not selling. "Please explain Dave. Why aren't HAL's selling?" Bowman hesitates. "You aren't Amiga compatible."