From: dhaynie@jersey.net (Dave Haynie) Newsgroups: comp.sys.amiga.misc Subject: Re: The Boing is BACK! Date: Wed, 03 Sep 1997 18:35:10 GMT Organization: PIOS Computer Lines: 156 Message-ID: <340da574.2746650094@news.jersey.net> References: <5ujbfb$tdh$1@coranto.ucs.mun.ca> NNTP-Posting-Host: chip175.jersey.net X-Newsreader: Forte Free Agent 1.1/32.230 Path: 195.95.96.10!news.unisource.be!planetinternet.be!newsgate.unisource.nl!newsgate.cistron.nl!het.net!Cabal.CESspool!bofh.vszbr.cz!nntprelay.mathworks.com!news.mathworks.com!uunet!in3.uu.net!209.66.0.12!news.jersey.net!not-for-mail On 3 Sep 1997 09:43:07 GMT, gnoel@morgan.ucs.mun.ca (George Noel) wrote: >>g> I bet a lot of those companies would love to get their hands on some of >>g> the Amiga chipset patents to make their chipset stand out more.. I doubt it. If these guys even know about the Amiga patents, they're either worried about, or working around, the chipset patent. Which is basically about the blitter (we always called it "the blitter patent", but it's actually, in the way of patents, covering the points of the original Amiga chipset that were unique at the time). >Perhaps they don't see the value when Windows will not make any use of >them and probably because they don't feel, due to brainwashing, that the >Amiga custom chipset has anything they'd be interested in. I don't know about that. One way or another, graphics chip companies have moved architecturally in the Amiga direction on 2D stuff. Most of the chips support three operand blits, line draw, area fill, etc. In fact, this was all "standardized" in IBM's XGA chipset. No one exactly cloned XGA, but many of the SVGA chipsets have grown in that direction over the years. The only primary features that Windows would not use are copper and HAM. When you see you can update a typical PCI display in 24-bit faster than any Amiga can update a HAM8 screen, at equal resolutions, it does put a certain spin on the whole HAM thing. AAA had some more useful built-in compression modes, but in this day of cheap memory and fast buses, it's questionable if even that would be all that useful today. >>g> >g> Expensive custom chips? Wasn't AGA's cost to manufacture something >>g> like $55? >>g> > >>g> >Something like that. Kinda hard to compare it to stuff now that is half >>g> the cost for 10x performance huh? ;) >> >>g> Half the cost FOR GRAPHICS ALONE.. get that point into your head. >>g> OCS/ECS/AGA/AAA was a lot more then just graphics. >>Yeah, 8 bit voice. Oh boy. The internal cost (something closer to the real cost than what you would pay at replacement prices) for AA was about $15.00 for the three chips. But you wind up with another $20 or so in support logic: fast PALs, buffers, part of Gary/Gayle/Akiko, the DAC, etc. That's price comparable to off-the-shelf parts today, it's just that you get so much more in the off-the-shelf parts. A SuperI/O chip, which supplies keyboard, mouse, two RS-232s, parallel, floppy, and maybe some general purpose I/O runs around $10-$15, depending on specifics (add $2 for two 8520s into the Amiga side of things here). A cheap 64-bit graphics chip with PCI interface and on-board clock synthesizer and DAC, maybe only 5x-10x faster than the Amiga chips, runs $15-$20, depending on who's you buy. A cheap 16-bit DAC costs about $1.00, but it's not something you can directly hook to a personal computer (these tend to run I2S or some-such serial bus). A 16-bit CODEC runs around $5.00, for stereo audio in and out. If you pay more, you can get MIDI, synthesis, and game ports on the same chip. >Which still sounds better in some ways then some 16 bit PC cound cards I >have heard stuff being played from. One big problem with sound cards is that, regardless of whether or not they'll handle 16-bit digital audio, there's still plenty of stuff out there using the old SoundBlaster FM synthesizer. Or 8-bit sampling. Meanwhile, everything on the Amiga is at least 8-bit sample based, which does sound better than the FM (any decent MOD file will convince you of this). And cheap sound cards aren't made very quietly. But put that in prespective -- you're not likely to find anything worse than maybe 70-75dB S/N (good stuff in 16-bit runs up around 90dB). That's better quality that you can get from FM radio, or most folks cassette tape machines (yes, a professional or audiophile tape deck with metal tape can exceed this). The point being, the systems are capable of sounding as good as most of the music most people listen to. If it doesn't, blame the source material, not the hardware. Or blame yourself, if you an audiophile expecting to get awesome sound from cards made by computer people. Get a real sound card if you need something better, start at around $250... >It is more then likely that they don't see the obvious benefits and they >are already too involved in other projects. I wager, if someone came knocking at their door, asking to license the Amiga chipset, for the purpose of making a clone (or improvement), Gateway wouldn't let them out the door without signing on the dotted lines. Chipset development is expensive -- they'd just at the chance of getting someone else to pay for this. >>g> >Got a few billion dollars for R&D and marketing you could chip in to AI? >>g> Billion?!?!? $250 million for R&D of a new chipset and then add however >>g> much else onto it that they wanted to spend on advertising. $250 million? Get real. Intel may spend $100 million on a new CPU, but they're pushing the limits of a poor and ancient architecture. And they have the money. You're talking under $25 million for the whole chip project, at least at an existing design center. If Gateway did it, they'd have a fairly substantial investment to make in Suns, software, designers, etc. You could easily spend $50 million on advertising, and another $50 million paying for applications to be ported. >Sony did a lot of things from scratch when developing their chipset and >they take a loss on the hardware to get the volumes out there and make it >back with the software. On paper, they can't take a loss on the hardware, that's dumping, it's illegal, at least in the US. Of course, they don't have to make a profit, either. But don't think any reasonable new Amiga chipset wouldn't also be 100% from scratch (well, unless you hooked up with Hepler and got his post-Hombre ideas). You're not going salvage anything from AA. Not a transistor. >>g> But it is the new people that we need to attract so an Amiga priced at PC >>g> clone prices wouldn't stand out as much anymore if the price performance >>g> ratio is there again. If the Amiga can BEAT the PC clone price and have >>g> better price/performance ratio then we are back to the way the Amiga was >>g> back in '85. >>So your going to lose the old supporters trying to get new customer base? Not >By this time you won't lose the old supporters as long as you provide a >perceived improved price/performance ratio also. And given that 90% of the supporters have A500s or A1200s, that's not the hard target. Building something that's very exciting at the $500 price point means making something that would still be interesting at $1000, but still managing to sell it for $500, at a profit. >>Is it much more then the $50 of the base gfx card to add a 2D chip? > >I guess not. The base gfx card contains a 2D chip. No one makes them without blitters anymore. >>So I'll pay a little bit more then $50 for a 64 bit card for video output, not >>a biggie. > >$50 for a card? Perhaps in the used market and older ones at that. Most of the basic PCI cards, 64-bit DRAM based things, are around the $50 price point, especially if you buy "generic" cards, not Diamond or Number Nine or such. Which isn't much of a risk -- the whole thing is just graphics chip and memory. There's a Trident chip with video output, including 3-line flicker filtering and 2D acceleration, for around $50 new. S3's Virge is just over $50 (I think I paid about $60 last Spring for the card in my garage), while their 2D-only stuff is below $50. Dave Haynie Vice President, Technology PIOS Computer haynie@pios.de "...no RISC, no fun"