Path: chuka.playstation.co.uk!news From: "Martin Swaine" Newsgroups: scee.yaroze.programming.2d_graphics Subject: Re: Animating sprites Date: 20 May 1998 10:09:33 GMT Organization: Lightwork Design Lines: 105 Message-ID: <01bd83df$772808e0$80c809c0@antonio.lightwork> References: <01bd7ac4$e4eda480$7397989e@antonio.lightwork> <01bd803a$2c971800$80c809c0@antonio.lightwork> <355E0F4F.1CFAECA7@ndirect.co.uk> NNTP-Posting-Host: lightwk.demon.co.uk X-Newsreader: Microsoft Internet News 4.70.1162 Thanks Alex for your explanation. I am however still unsure if I can do anything practical to avoid this problem. It seems a shame to abandon this work because of a disapointing display. If you are interested, I have uploaded the source etc. to http://www.netyaroze-europe.com/~mtn446/ftp/spline.zip Can anyone suggest code changes that might avoid the ghosting. Alex Herbert wrote in article <355E0F4F.1CFAECA7@ndirect.co.uk>... > > > Martin Swaine wrote: > > > Thanks people for your comments. I now have better information. > > > > This effect can be replicated with the Sony's Check program (the one where > > all the balls bounce around the screen) that comes with the Yaroze kit. > > > > If you change the source so the deltas for x and y are quite large (say 10 > > or 20) you see the effect I mean when you keep the number of balls at one. > > The illusion of the "trail" of balls is clearly visable. I guess this means > > I am using the drawing primitives and double buffering correctly. > > > > In response to the other questions, my sprite has semi-transparency off and > > is of a single colour (red). However, since this problem is reproducable > > with the Sony demo code then I believe there can be no unique problem with > > my code. > > > > I too was drawn to Robert's suggestion that it maybe a deception of the > > eye, but am a little suprised as I have never seen this manifest with TV or > > video images before. > > Oh right, I see. Sorry, I did know about this effect, but from your arriginal > description I thought it was something more complcated. Robert's suggestion is > true to a degree. But if you are running at the full fame rate (i.e. 50 or > 60fps) then this effect it not mormally noticable. Drop your frame rate to > 25fps and the "movement ghosting" becomes very obvious. Try a fast full screen > scroll at 50fps, then at 25fps. At 50fps it look silky smooth, but at 25fps it > ugly as hell. It is also more noticable if there is a high contrast between > the object which is moving, and the surface it is moving over. > > However, this effect is less noticable in most 3D environments. This is > because 'realistc' 3D worlds tend to be built up with mid-tones and colours > with less saturation. And, the relative screen movements tend to be smaller > and less consistent. In this way, the ghosts get lost (to some degree) in the > overall picture, and you can get away with a slower frame rate. Arcade machine > manufacturers such as Namco and Sega, usually design their software to run at > the full 60fps to get that perfect smoothness. Compare Ridge Racer on the PSX > and in the arcade, and weep at the difference! > > With regard to TV and video... > > Game video has an effective camera exposure time of 0, and you are creating > movement from static images. In the real world, video and film cameras have a > longer exposure time and catch movement within one frame. (Pause your VCR at a > point where there is fast movement on the screen to see what I mean.) This > "motion blur" hides the ghosting effect (by not having any hard edges > perpendicular to the direction of movement). Conversly, if you are watching a > cartoon (which is made up of static images) you can see the ghosting. It is > especially noticable when the camera is panning over a background. > > Interestingly, motion blurring can be created in software realtime, but it > takes a hell of a lot of processing power to get it right. This is something I > expect to see done in hardware in the next generation of games machine. > > Anyway, I gone on far too long already... > > > > Herbs > > >