What did you think of 3dfx?

What did you think of 3dfx?

  • Pretty good chips, but too slow to react to a rapidly changing market (voodoo3 16-bit anyone?)

    Votes: 0 0.0%
  • Last decent product was Voodoo2 and Savage3D was faster. Nice antialiasing though.

    Votes: 0 0.0%
  • Lucky enough to get first decent 3D board out, never knew where they were going

    Votes: 0 0.0%
  • Who?

    Votes: 0 0.0%

  • Total voters
    182
Radea said:
Thats how I feel anyway. The Voodoo3 lacked 32bit color, so overall the TNT2 was the better card (Among other reasons). The Voodoo2 did have speed on itside, so it was a decent 3Dfx product.

Fair enough.

However, I disagree with the AA statement in this choice. Maybe it was nice at the time, but now their AA is mediocre at best. It hardly deserves all the praise it gets. :?

Uh... excuse me? 3dfx's AA used nearly the same sample pattern as ATi's 4x (rotated the other direction), and is Supersampling. The only thing it's missing is gamma correction :) Mediocre indeed...
 
I dont find any of the vote options reflects my opinion of 3dfx. Such an option might be:

"Pretty good chips, but they didn't spend enough on marketing to balance with the onslaught of the competition's marketing efforts."
 
Reverend said:
I don't believe John "worked" (i.e. employed, as in salary-based) for 3dfx... IIRC he was a freelancer (game reviews) for 3dfx's web spin-off, 3dfxgamers. Dave Barron and Kristof, however, were salaried 3dfx employees.

Correct, though that position got me some juicy contacts. Me knew Rampage specs and product lineup long before most others (even most 3dfx employees).
 
Uh... excuse me? 3dfx's AA used nearly the same sample pattern as ATi's 4x (rotated the other direction), and is Supersampling. The only thing it's missing is gamma correction :) Mediocre indeed...
:oops: Well from what I've personally seen, it didnt do as good of a job on removing jaggies. And myself, I dont like SS since it blurs small-print text to the point where its unreadable (Ex. RPG games) :p
 
Tagrineth said:
Uh... excuse me? 3dfx's AA used nearly the same sample pattern as ATi's 4x (rotated the other direction), and is Supersampling. The only thing it's missing is gamma correction :) Mediocre indeed...

I can only agree with Tagrineth wholeheartedly. It's taken ATI to finally make FSAA usable again, but it's still not as good (can we say texture aliasing?) as V5.... of course, thats all the V5 had....... but, at the time, it was enough.
 
Radea said:
Well from what I've personally seen, it didnt do as good of a job on removing jaggies.

Well then you pretty much stand alone. No one who has seen V5 4xRGSS AA disagrees it wasnt excellent at edge aliaising, as well as good texture aliaisng.

I agree with text issues in a lot if games though.
 
martrox said:
Tagrineth said:
Uh... excuse me? 3dfx's AA used nearly the same sample pattern as ATi's 4x (rotated the other direction), and is Supersampling. The only thing it's missing is gamma correction :) Mediocre indeed...

I can only agree with Tagrineth wholeheartedly. It's taken ATI to finally make FSAA usable again, but it's still not as good (can we say texture aliasing?) as V5.... of course, thats all the V5 had....... but, at the time, it was enough.

Our AA is significantly better at edge antialiasing than V5s thanks to gamma correction. Your point about texture antialiasing thanks to supersampling is reasonable, so it's probably close to a wash overall.
 
Randell said:
Well then you pretty much stand alone. No one who has seen V5 4xRGSS AA disagrees it wasnt excellent at edge aliaising, as well as good texture aliaisng.

RIP The English Language 10 April 2003. :LOL:

Back on track.

The other thing about 3dfx's AA is that it wasn't just a rotated grid - the sample points were jittered too. :)
 
andypski said:
Our AA is significantly better at edge antialiasing than V5s thanks to gamma correction. Your point about texture antialiasing thanks to supersampling is reasonable, so it's probably close to a wash overall.

Who is "Our"? And I don't think there's much difference between ATI's currrent FSAA edge sampling (on R300) and V5's - at least at 4X.... of course, I'd like to add that I'd glady trade the texture antialiasing for the "other" things the R300 brings to the table..... but we can wish for it all, can't we?
 
martrox said:
Who is "Our"? And I don't think there's much difference between ATI's currrent FSAA edge sampling (on R300) and V5's - at least at 4X.... of course, I'd like to add that I'd glady trade the texture antialiasing for the "other" things the R300 brings to the table..... but we can wish for it all, can't we?

If you look at my profile you can see who "our", is although I think it should be reasonably clear from the context anyway.

There are many situations in which gamma corrected sub-samples are far superior to uncorrected samples (which to the best of my knowledge includes V5's). In terms of sample positions in 2X and 4X mode V5 and R300 are pretty similar - it's the quality of each sample that is higher in R300.

On low contrast edges you would probably only notice a marginal improvement in quality. On high contrast edges (White->Black or even better something like Red->Green) V5's uncorrected antialiasing should look significantly worse.

It would be interesting to do a direct AA quality comparison, but I don't have access to a V5 here. If anybody does and would like to do this test then it would be good to see.

- Andy.
 
You wouldn't believe it, but I actually have a Radeon 9700 Pro and my Voodoo 5 6000 AGP in the same machine, at the same time. Yes... 2 AGP cards in the one system at the same time.

Now reasaon why is I was given one of those PCI->AGP riser cards. Since the V5 6000 ran in PCI mode anyway, it works fine in the riser cards.

I'll try to get some pics of my system as it looks quite strange. The V5 6000 now takes up a massively huge amount of space in the system. :)

Anyway, my opinion is that the Radeon's edge aa quality is far superior to the Voodoo5 at the same quality level. The Voodoo5's supersampling of course does have it's benefits.
 
Althornin said:
if you only knew about the never released Ravenous chipset...
I know it gets 348 fps in quake3 at 1600x1200 with 128bit FP color, 128x AA, and 384tap AF.
Too bad it was never released. BUt it was almost finished, and i heard that some of the engineers from S3 who had worked on this technological marvel went to antoehr company where they are scheming to release their tech into the marketplace. I cannot say anything more about the mystery engineers or the mystery company tho, for fear of reprisals against said people.

ROTFL... :D
 
Tagrineth said:
Randell said:
Well then you pretty much stand alone. No one who has seen V5 4xRGSS AA disagrees it wasnt excellent at edge aliaising, as well as good texture aliaisng.

RIP The English Language 10 April 2003. :LOL:

coming from an American, that's rich :!:

yes, messages typed in a lunch time rush don't read well, do they?
 
Don't moan. Beyond3D has about the best English of any board I've read on the net. While wandering round the net I've seen nothing that disproves my theory that the quality of language and the quality of posts are usually related.

Which is a big credit to the average intelligence of the crowd here, and an even bigger credit given the number here for whom English is a foreign language... yes, I do include Americans in that....
 
Dio said:
Don't moan. Beyond3D has about the best English of any board I've read on the net. While wandering round the net I've seen nothing that disproves my theory that the quality of language and the quality of posts are usually related.

Which is a big credit to the average intelligence of the crowd here, and an even bigger credit given the number here for whom English is a foreign language... yes, I do include Americans in that....

i doneno wat u meen? :p
 
Tagrineth said:
Randell said:
Tagrineth said:
RIP The English Language 10 April 2003. :LOL:

coming from an American, that's rich :!:

Oi, I'll have you know that I'm a Brit! And if you search through my posts over time you'll notice I still use proper spellings! ^_^
are you, sorry, fair cop then guv!

How did you end up in Florida?
 
Randell said:
are you, sorry, fair cop then guv!

How did you end up in Florida?

Abridged version:

Born in England. Parents divorced when I was seven. Moved to Germany with mother and sister. Mother remarried, a Floridian. Moved to Florida.
 
Tagrineth said:
The other thing about 3dfx's AA is that it wasn't just a rotated grid - the sample points were jittered too. :)
And now you're going to tell us what the difference is between a rotated grid and subsample jittering, meaning that the samples were shifted from the center of a pixel by certain values in order to form a rotated square grid...


Yes, you could change the sample positions (like Colourless' AA adjuster does), but I doubt you'd find a better pattern than 4xRG.
 
Colourless said:
You wouldn't believe it, but I actually have a Radeon 9700 Pro and my Voodoo 5 6000 AGP in the same machine, at the same time. Yes... 2 AGP cards in the one system at the same time.

Now reasaon why is I was given one of those PCI->AGP riser cards. Since the V5 6000 ran in PCI mode anyway, it works fine in the riser cards.

I'll try to get some pics of my system as it looks quite strange. The V5 6000 now takes up a massively huge amount of space in the system. :)

Anyway, my opinion is that the Radeon's edge aa quality is far superior to the Voodoo5 at the same quality level. The Voodoo5's supersampling of course does have it's benefits.
Quite a while ago, I had used Basic's program and it showed my V5500 was utilising Multisampling.

I no longer have the V5500 currently, though I'd like to know what your programs showed.... Multisampling or Supersampling?

Thanks.
 
Back
Top