The ARTS: Episode 5 -- Featuring Legalize/Pilgrimage!
category: general [glöplog]
The ARTS Episode 5 is a very special show, in that it's our drinking episode.
Hear first hand as the gradual inebriating effects set in on SPINSANE. Topics discussed include: Grapevine Issue 2, Mindcandy vol2, Assembly airfare, new ASCII packs, alternative underground talk shows, and a whole lot more.
Special guest on this ep is LEGALIZE, beloved organizer of the Pilgrimage demoparty. He speaks at length regarding the history of 3D computer graphics and their origins, breaking down a lil' "Demo Coding 101".
Running at a cool 00:50:33, we promise this is our WORST show ever!
Decide for yourself:
http://www.acid.org/radio/index.html#ARTS-EP05
Hear first hand as the gradual inebriating effects set in on SPINSANE. Topics discussed include: Grapevine Issue 2, Mindcandy vol2, Assembly airfare, new ASCII packs, alternative underground talk shows, and a whole lot more.
Special guest on this ep is LEGALIZE, beloved organizer of the Pilgrimage demoparty. He speaks at length regarding the history of 3D computer graphics and their origins, breaking down a lil' "Demo Coding 101".
Running at a cool 00:50:33, we promise this is our WORST show ever!
Decide for yourself:
http://www.acid.org/radio/index.html#ARTS-EP05
By the end of the episode you'll realize that Spinsane isn't "Faking It". :)
rich sure talks alot. kinda reminded me of the volumetric light effect not beeing real, discussion at csipd...
We now have a button (88x31 GIF). If you'd like to link back to us, please feel free to use it:
where can i find a list of all legazlize's demos?
Legalize is a member of Polygony which doesn't have any demos and as far as I can tell has two members. Still juice by him for being so damn supportive by *cough* starting a demo party! Sounds like he's pretty darn knowlegeable of 3D "stuff".
On another note: *takes credit for the 88x31 button*
On another note: *takes credit for the 88x31 button*
where can i find a list of all the demoparties Legalize attended? :D
next up on channel DK-DEMO-RADIO - YOUR NUMBER ONE CHANNEL OF NONSTOP DEMOSCENE! - I'll be explaining to you how to track music and do pixel graphics!
[just because i know *everything* about tracking and pixelling]
[just because i know *everything* about tracking and pixelling]
Rasmus first you should do an in-depth documentary of how to be a jackass on Pouet, as you appear to have a scholarly knowledge of that subject... :)
chris, nah, having first hand knowledge usually disqualifies one from making public announcements :D
Rasmus: I have done some demos, but they were all for the Evans & Sutherland ESV Workstation and the Evand & Sutherland Digistar II shown at SIGGRAPH. IIRC, it was SIGGRAPH 1991 and 1992. I have some photographs of the screen for one of the demos, perhaps I'll get it scanned someday. Alas, I have neither the source code (it was owned by E&S being a "work for hire") nor any digital files of screen captures.
The ESV was a workstation manufactured by E&S in 1989. It featured 100K Z buffered, lit, Gouraud shaded, depth cueued 10 pixel polygons/second and 1 M color interpolated, 10 pixel, antialiased, depth cueued vectors/second. The CPU was a MIPS R3000 RISC chip running at 25 MHz. The graphic subsystem consisted of up to 24 AT&T DSP32C DSP chips operating in parallel feeding an 88-bit deep frame buffer at 1280x1024, 32-bit Z buffer, double-buffered 24-bit color buffers and 8 bits of window ID used for fast window clipping and frame buffer clearing. The DSP rasterized primitives into spans and spans were rasterized into pixels by a custom ASIC. The ASIC also handled all frame buffer pixel operations (copy, readback, etc.). Glenn Eckart and I wrote a functional simulation of the ASIC that was used to verify the clock level simulation of the ASIC design as well as serve as a backend prototype for the X server team before first tapeout of the ASIC. The chip was usable for production on the first tapeout of the ASIC (a significant accomplishment for the time).
The ESV demo was a terrain display using color interpolated antialiased vectors -- the ESV was capable of extremely high quality rendering of AA points and lines and although current PC hardware vastly exceeds the ESV in terms of polygon performance and texturing effects, they still cannot match the quality of the AA points and lines the ESV could do at a sustained rate of 1 Mvecs/second (not peak rates as most vendors claim in their benchmarks). The model was constructed from a terrain heightfield model with a co-registered aerial photo texture map, obtained from the Simulation division at E&S. The actual terrain is a small canyon outside Salt Lake City called Pine Canyon.
Instead of drawing texture mapped polygons (the ESV had no texture mapping), I took each scanline of the height field and turned it into a color interpolated polyline with the vertex color chosen from the texture map. As the scaling was shrunk down, the high AA quality of the ESV gave the illusion of textured polygons. An illusion so convincing that a herd of SGI engineers and marketing people (our big competitor at the time) gathered outside our booth to figure out what we had done. Fooling SGI engineers with a rendering effect on our hardware is something that I'm still proud of to this day. The display ran in real-time and could be translated, rotated and scaled through the use of a knob box. The rendering API was PEX (PHIGS Extensions to X), which E&S was the first to implement with real-time performance for large, complex scenes. The knob box was controlled through the X Input Extension and the UI was implemented with the Motif widget set.
The Digistar II was a projection system made by E&S that was an adaptation of their simulation technology for planetariums. It projected a vector graphics display through a spherical distortion lens onto an overhead dome. The system at SIGGRAPH used a 30 ft' diameter overhead dome. The Digistar II was only capable of grayscale graphics with point and line primitives, although the lines could color interpolate along the primitive. The display could be animated and updated in real-time, although very large models would cause the vector-based display processor to generate some flicker. (This is an inherent limitation of all vector-based refresh displays and one reason why frame buffers became so popular as they decouple display rates from rendering rates.)
The demos I wrote for the Digistar were dynamically loadable modules into a dataflow scientific visualization system called AVS at the time, although it is now marketed under the name AVS/Express. The modules I wrote generated 3D models that were then transformed by the visualization system into the model format used by the Digistar. The Digistar then displayed the 3D models on the overhead projection dome. The whole system was interactive and the models could by dynamically updated by manipulating UI controls in AVS.
The models generated were a vector-based visualization of the Mandelbrot set: instead of coloring pixels by the length of the orbit, I instead drew the orbit itself, leading to a grid of vector spikes, one spike per orbit. Periodic orbits became cylindrical spiral columns and diverging orbits became outward flowing curves and converging orbits became vertical spikes. To my knowledge this was the first publicly displayed vector visualization of the Mandelbrot set in this form. Even now, I have not seen any visualizations of the Mandelbrot set that have reproduced the visualization I showed at SIGGRAPH.
The second AVS module I wrote was an adaptation of the vector-based terrain display I had written for the ESV. The same decomposition of height field + registered aerial photo was used to generate a grayscale polyline set. The height field was of the Mosel river valley in Germany with a registered grayscale aerial photograph. This model was also the first time that a terrain had been visualized with the Digistar system. Both of these Digistar datasets were contributed to the Digistar user's library and were displayed as demos in planetarium systems around the world.
As for demo parties, as I reported last year, I have only attended Pilgrimage, the party that I organized.
The ESV was a workstation manufactured by E&S in 1989. It featured 100K Z buffered, lit, Gouraud shaded, depth cueued 10 pixel polygons/second and 1 M color interpolated, 10 pixel, antialiased, depth cueued vectors/second. The CPU was a MIPS R3000 RISC chip running at 25 MHz. The graphic subsystem consisted of up to 24 AT&T DSP32C DSP chips operating in parallel feeding an 88-bit deep frame buffer at 1280x1024, 32-bit Z buffer, double-buffered 24-bit color buffers and 8 bits of window ID used for fast window clipping and frame buffer clearing. The DSP rasterized primitives into spans and spans were rasterized into pixels by a custom ASIC. The ASIC also handled all frame buffer pixel operations (copy, readback, etc.). Glenn Eckart and I wrote a functional simulation of the ASIC that was used to verify the clock level simulation of the ASIC design as well as serve as a backend prototype for the X server team before first tapeout of the ASIC. The chip was usable for production on the first tapeout of the ASIC (a significant accomplishment for the time).
The ESV demo was a terrain display using color interpolated antialiased vectors -- the ESV was capable of extremely high quality rendering of AA points and lines and although current PC hardware vastly exceeds the ESV in terms of polygon performance and texturing effects, they still cannot match the quality of the AA points and lines the ESV could do at a sustained rate of 1 Mvecs/second (not peak rates as most vendors claim in their benchmarks). The model was constructed from a terrain heightfield model with a co-registered aerial photo texture map, obtained from the Simulation division at E&S. The actual terrain is a small canyon outside Salt Lake City called Pine Canyon.
Instead of drawing texture mapped polygons (the ESV had no texture mapping), I took each scanline of the height field and turned it into a color interpolated polyline with the vertex color chosen from the texture map. As the scaling was shrunk down, the high AA quality of the ESV gave the illusion of textured polygons. An illusion so convincing that a herd of SGI engineers and marketing people (our big competitor at the time) gathered outside our booth to figure out what we had done. Fooling SGI engineers with a rendering effect on our hardware is something that I'm still proud of to this day. The display ran in real-time and could be translated, rotated and scaled through the use of a knob box. The rendering API was PEX (PHIGS Extensions to X), which E&S was the first to implement with real-time performance for large, complex scenes. The knob box was controlled through the X Input Extension and the UI was implemented with the Motif widget set.
The Digistar II was a projection system made by E&S that was an adaptation of their simulation technology for planetariums. It projected a vector graphics display through a spherical distortion lens onto an overhead dome. The system at SIGGRAPH used a 30 ft' diameter overhead dome. The Digistar II was only capable of grayscale graphics with point and line primitives, although the lines could color interpolate along the primitive. The display could be animated and updated in real-time, although very large models would cause the vector-based display processor to generate some flicker. (This is an inherent limitation of all vector-based refresh displays and one reason why frame buffers became so popular as they decouple display rates from rendering rates.)
The demos I wrote for the Digistar were dynamically loadable modules into a dataflow scientific visualization system called AVS at the time, although it is now marketed under the name AVS/Express. The modules I wrote generated 3D models that were then transformed by the visualization system into the model format used by the Digistar. The Digistar then displayed the 3D models on the overhead projection dome. The whole system was interactive and the models could by dynamically updated by manipulating UI controls in AVS.
The models generated were a vector-based visualization of the Mandelbrot set: instead of coloring pixels by the length of the orbit, I instead drew the orbit itself, leading to a grid of vector spikes, one spike per orbit. Periodic orbits became cylindrical spiral columns and diverging orbits became outward flowing curves and converging orbits became vertical spikes. To my knowledge this was the first publicly displayed vector visualization of the Mandelbrot set in this form. Even now, I have not seen any visualizations of the Mandelbrot set that have reproduced the visualization I showed at SIGGRAPH.
The second AVS module I wrote was an adaptation of the vector-based terrain display I had written for the ESV. The same decomposition of height field + registered aerial photo was used to generate a grayscale polyline set. The height field was of the Mosel river valley in Germany with a registered grayscale aerial photograph. This model was also the first time that a terrain had been visualized with the Digistar system. Both of these Digistar datasets were contributed to the Digistar user's library and were displayed as demos in planetarium systems around the world.
As for demo parties, as I reported last year, I have only attended Pilgrimage, the party that I organized.
ZZZZZZZZZZzzzzzzzzZZZZZzzzz<snort> did the train wreck happen yet? No? Ok, wake me up when something's on fire... ZZZZZZzzzzzzzzzzzzZZzzzz
muhaha truck!
oh.. shut up guys!
I predicted this would happen -- if only someone read my prophetic quatrains!
my god, he's like an informative optimus!
leg: you need to win sceneevent with a 64k intro saying "i am rasmus" to get some recognition for your demomaking talent around these parts man ;)
tech talk and scene dedication doesnt impress us, "woooo" effects and artistic talent do :)
tech talk and scene dedication doesnt impress us, "woooo" effects and artistic talent do :)
the talk on the show was nice for starting 3d coders though. props for that.
^^^ nice in a non sarcastic way =)