In response to recent events
- Thread starter Hellbinder
- Start date
- Status
legion88 said:
Actually, if I recall correctly, the V3 generally outperformed the TNT2 and had considerably better performance-to-quality because of the 2x2 post filter.
Where the additional features useful? Not really. 512x512 textures were supported in Q3 on it, but was it really an option to use things such as that, 32-bit color and high resolutions in Q3? No. V3 offered substantially better 16-bit quality that was very fast.
Oh but I do recall commenting quite freely about the lack of 32mb of memory and high res textures. I did talk about the lack of 32-bit, but at the time it wasn't nearly as much of an issue.
By then the V3 was old news.
I don't recall saying that the hardware AA on the V5 made it faster, but that it made it more compatible. Yes I did mention about the V5 not being able to do trilinear, FYI.
As a matter if fact I didn't. The whole job thing didn't come up until the summer, after I had graduated. I was still in school at the time of the preview/review.
As legion said, TNT2 had, depending on model, comparable fillrates to V3.Dave said:
That goes for both sustained and peak rates I'd say.
I have a TNT2U in my old machine, and I'm glad it has large texture support, 32MB, and 32-bit colour. They were useful features.
The way you are downplaying the advantages of TNT2 and at the same time saying V3 had "considerably better performance-to-quality" and "substantially better 16-bit quality that was very fast" might actually be your honest opinion, but does make you look biased in my eyes.
I'm saying that you can run at 16-bit and get quality closer to 32-bit. So you are getting better performance. 16-bit on the TNT2 is rather poor compared to the V3.
Biased? About a dead company. I really don't think so.
As for them being useful features, I am curious where you found them useful?
Biased? About a dead company. I really don't think so.
As for them being useful features, I am curious where you found them useful?
I personally think that the post filter was one of the most underrated features. Granted first implementations sufferend from the fact that they were line based but the latest incarnations of the filter were actually quite good. And it did manage to somewhat reverse the ugly dither patterns. Obviously it would never be equal to true 24/32 bits but it was better than plain 16 bit.
Not that I read more than the 5 lasts posts of this thread, so don't know what is actually being discussed
K~
Not that I read more than the 5 lasts posts of this thread, so don't know what is actually being discussed
K~
Kristof said:
it was "underrated" for a reason - it didn't manage to reverse the ugly dither patters exactly were it was needed most - at locales of hight contrast transitions, for the simple fact that it was trying to achieve two mutually-exclusive things:
* to somewhat 'revert' dithering, which was naturally most noticible at high-contrast transitions, i.e. at a black-white transition you'd get a 'chess-board' effect
yet
* in order to preserve details it was forced 'disabled' when coming across such high-contrast locales.
see the issue?
darkblu said:Kristof said:
Well thats not exactly true. Dithering dithers between 2 colors that should not be very far apart. I doubt dithering ever dithers between black and white, it should dither between 2 grey scale values that are pretty close together.
High contrast was to avoid that true texture details and polygon edges would get blurred which is not the idea of the post filter although they could have claimed edge AA with it
I do have to say that the post filtering only worked really well on gradients, whcih are quite comon in 3D Graphics.
Now I did read through the last couple of pages of this article and am not going to go into quotes since its too easy to rip things out of their historical context and article context.
For what its worth I am trying to get back to technical article writing and will very hard try not to say that company X, Y or Z **cks
K~
BenSkywalker
Regular
Dave(UK)-
I didn't say that you hadn't decided on a direction yet, I'm just stating what we can see reading things every day. Actually, with each change in leadership there has been an obvious difference in at least the tone of things if not necessarily the overall direction. When Rev was at the healm the site actually started paying attention to some gaming related issues and how the different boards stacked up in different areas there. I thought that that was a welcome addition. Since your taking over we have already seen your broader scope in terms of paying attention to the players who aren't one of the big two at the time in 3D. It is nice to see 3DLabs/Integraph and Matrox at least paid some attention to here, along with actual reviews of ATi based products. I'm not saying that you haven't done anything, just the site has a somewhat generic feel to it right now. In no way do I mean that as a slam, I write for what is likely the most generic and directionless gaming site I've ever seen.
When the T&L articles were published the forum members gave them many different examples of actual games that rebutted their assertions along with many different developer comments on their statements in particular involving lighting(the GeForce being much 'slower' then a CPU). We listed off many different benches that showed the exact opposite, in fact most of them did. This was brushed aside. Why? Because it didn't come from someone at nVidia? This doesn't just include nVidia products, although I never saw the same slams against the Radeon in terms of an article(which was actually considerably slower then the GeForce). Rev, although IIRC it was over at the Pulpit, gave explicit examples of T&L helping to make a game run faster and look better. If that had been done here, even if it was simply a case of 'it can happen but it is rare', it would have sent a message that a true attempt at impartiality was attempted. That is just one example.
How can you figure things out without companies? When the hardware is readily available along with extensive benchmarks to demonstrate it, why do you need a company's imput? If something works all the time then the company isn't needed. If something works part of the time, and you are pointed directly to when it is and isn't, you don't need a company. If something never works, then I could perhaps seeing questioning the company to find out why that is the case, or perhaps if you are trying to find something out about an unreleased product. Other then that, I would prefer that company contacts were kept to a minimum at any of the sites I need. Do I want nVidia to walk you through how to review their board? No, and I would wager heavily you don't either.
It was integrated into DirectX as the article mentions. Has anyone ever thought of asking Microsoft about the details? Question 2 seems like the only one they wouldn't have been able to answer for you(as it is nV specific). I would think that anyone looking for the most impartial answer as far as 3D hardware is concerned would rather talk to MS. Who would you trust more on the benefits of PS 1.4 over PS 1.3, the vendors who have a stake in them or a comparitively neutral party? Whenever it is possible, it seems like it would make a lot more sense to get the views of someone who is familiar with the technology but has no interest in platform specific implementations.
If you seperated it from the Wildcat it would
Wanting to run a site and truly caring about what it is you cover are two different things. How much of what you do is because you should, and how much of it is because you can't wait to get to do it? That's what the missing element seems to be now as opposed to previously. I honestly think that you enjoy what you do very much, I appreciate the lengths you go through to keep up on all the players(which has been missing prior to your taking the lead), but this doesn't show in terms of how the overall site is viewed.
Bullshit Dave. I agree that you understand the importance, and I think getting K to pump out some new content is a very good thing, but if you honestly think you can't write the same types of articles you are selling yourself far too short. You understand the importance of bringing in a wider audience, and you realize that it is a desireable thing. With all due respect to K(as he does do a hell of a job), how many people do you think can understand his writings that don't already read this site that may be interested and would learn anything? There is a staggering gap between the, relatively speaking, laymen sites and B3D and you can fill in most of that gap yourself. I understand you don't have the time to cover it all, but think about a few things for a minute-
How many times have you had to answer what is to you, a very simple question on 3D on other boards?
How many times have you been discussing 3D on other tech sites and found out that the people who write for that site have a hard time simply following the discussion(I know of at least once first hand)?
Do you think that any of the authors for the 'major' tech web sites know more then you do about 3D technology(not talking about access to NDAs)?
How many times do you think the members of these boards have run in to the same types of things?
What you are more then capable of doing is bridging the gap between places like Anand's or Tom's and a site like this. Right now there isn't another site that really attempts it(Digit-Life is the closest and they leave some to be desired). If you were to do what you are capable of in this sense you would also attract that kind of people B3D likes to have around, people who are genuinely interested in learning about 3D along with the minds to teach them. I'd rather see the boards full of questions about how different types of anisotropic filtering are implemented or the varrying levels of complexity of pixel or vertex shaders and how a given platform would handle them then another BS "NV30 is xxxxx" thread.
Kristof finishes up an article right now on Vertex Shaders and the long time readers and forum goers will certainly greatly appreciate it(I know I will). The average reader of Tom's or Anand's, even if they really want to be able to understand it, won't be able to. You can teach a lot of things to 99% of the geek community(at least), and a good amount of people are interested in learning about what you already know too.
I'm not trying to criticize you, you have already brought new angles to the site. What I'm talking about, the change in direction on this site, happened years ago and simply hasn't reverted back. I think that all of the people that write for the site should realize we don't expect everyone to know as much as Kristof when writing an article, but that in no way means that we don't want to read them authored by you, Jon or Marco(or previously Rev).
Dave(US)-
Was that aimed at me? I was trying to point out how the perception was easy to grasp from the content of the site. We've discussed these things at great lengths in the past and I know that we have a clear rift, at least we did, in the general goals that the different companies were trying to reach. The fact is that yours mirrored 3dfx's(and mine nV's at the time), and that did show through in your articles. Do I agree with the rest of your post? Almost entirely not, but you already knew that I have always believed that you were telling the truth in your writings Dave, and that your POV was yours and not because of an infatuation with 3dfx(well, that took some time to get to know you). If I were sitting in a position of authority at nVidia and all I had to go by was reading your articles, I would have had you labled as a '3dfx zombie'.
JB-
Giants.
I didn't say that you hadn't decided on a direction yet, I'm just stating what we can see reading things every day. Actually, with each change in leadership there has been an obvious difference in at least the tone of things if not necessarily the overall direction. When Rev was at the healm the site actually started paying attention to some gaming related issues and how the different boards stacked up in different areas there. I thought that that was a welcome addition. Since your taking over we have already seen your broader scope in terms of paying attention to the players who aren't one of the big two at the time in 3D. It is nice to see 3DLabs/Integraph and Matrox at least paid some attention to here, along with actual reviews of ATi based products. I'm not saying that you haven't done anything, just the site has a somewhat generic feel to it right now. In no way do I mean that as a slam, I write for what is likely the most generic and directionless gaming site I've ever seen.
When the T&L articles were published the forum members gave them many different examples of actual games that rebutted their assertions along with many different developer comments on their statements in particular involving lighting(the GeForce being much 'slower' then a CPU). We listed off many different benches that showed the exact opposite, in fact most of them did. This was brushed aside. Why? Because it didn't come from someone at nVidia? This doesn't just include nVidia products, although I never saw the same slams against the Radeon in terms of an article(which was actually considerably slower then the GeForce). Rev, although IIRC it was over at the Pulpit, gave explicit examples of T&L helping to make a game run faster and look better. If that had been done here, even if it was simply a case of 'it can happen but it is rare', it would have sent a message that a true attempt at impartiality was attempted. That is just one example.
How can you figure things out without companies? When the hardware is readily available along with extensive benchmarks to demonstrate it, why do you need a company's imput? If something works all the time then the company isn't needed. If something works part of the time, and you are pointed directly to when it is and isn't, you don't need a company. If something never works, then I could perhaps seeing questioning the company to find out why that is the case, or perhaps if you are trying to find something out about an unreleased product. Other then that, I would prefer that company contacts were kept to a minimum at any of the sites I need. Do I want nVidia to walk you through how to review their board? No, and I would wager heavily you don't either.
It was integrated into DirectX as the article mentions. Has anyone ever thought of asking Microsoft about the details? Question 2 seems like the only one they wouldn't have been able to answer for you(as it is nV specific). I would think that anyone looking for the most impartial answer as far as 3D hardware is concerned would rather talk to MS. Who would you trust more on the benefits of PS 1.4 over PS 1.3, the vendors who have a stake in them or a comparitively neutral party? Whenever it is possible, it seems like it would make a lot more sense to get the views of someone who is familiar with the technology but has no interest in platform specific implementations.
If you seperated it from the Wildcat it would
Wanting to run a site and truly caring about what it is you cover are two different things. How much of what you do is because you should, and how much of it is because you can't wait to get to do it? That's what the missing element seems to be now as opposed to previously. I honestly think that you enjoy what you do very much, I appreciate the lengths you go through to keep up on all the players(which has been missing prior to your taking the lead), but this doesn't show in terms of how the overall site is viewed.
Bullshit Dave. I agree that you understand the importance, and I think getting K to pump out some new content is a very good thing, but if you honestly think you can't write the same types of articles you are selling yourself far too short. You understand the importance of bringing in a wider audience, and you realize that it is a desireable thing. With all due respect to K(as he does do a hell of a job), how many people do you think can understand his writings that don't already read this site that may be interested and would learn anything? There is a staggering gap between the, relatively speaking, laymen sites and B3D and you can fill in most of that gap yourself. I understand you don't have the time to cover it all, but think about a few things for a minute-
How many times have you had to answer what is to you, a very simple question on 3D on other boards?
How many times have you been discussing 3D on other tech sites and found out that the people who write for that site have a hard time simply following the discussion(I know of at least once first hand)?
Do you think that any of the authors for the 'major' tech web sites know more then you do about 3D technology(not talking about access to NDAs)?
How many times do you think the members of these boards have run in to the same types of things?
What you are more then capable of doing is bridging the gap between places like Anand's or Tom's and a site like this. Right now there isn't another site that really attempts it(Digit-Life is the closest and they leave some to be desired). If you were to do what you are capable of in this sense you would also attract that kind of people B3D likes to have around, people who are genuinely interested in learning about 3D along with the minds to teach them. I'd rather see the boards full of questions about how different types of anisotropic filtering are implemented or the varrying levels of complexity of pixel or vertex shaders and how a given platform would handle them then another BS "NV30 is xxxxx" thread.
Kristof finishes up an article right now on Vertex Shaders and the long time readers and forum goers will certainly greatly appreciate it(I know I will). The average reader of Tom's or Anand's, even if they really want to be able to understand it, won't be able to. You can teach a lot of things to 99% of the geek community(at least), and a good amount of people are interested in learning about what you already know too.
I'm not trying to criticize you, you have already brought new angles to the site. What I'm talking about, the change in direction on this site, happened years ago and simply hasn't reverted back. I think that all of the people that write for the site should realize we don't expect everyone to know as much as Kristof when writing an article, but that in no way means that we don't want to read them authored by you, Jon or Marco(or previously Rev).
Dave(US)-
Was that aimed at me? I was trying to point out how the perception was easy to grasp from the content of the site. We've discussed these things at great lengths in the past and I know that we have a clear rift, at least we did, in the general goals that the different companies were trying to reach. The fact is that yours mirrored 3dfx's(and mine nV's at the time), and that did show through in your articles. Do I agree with the rest of your post? Almost entirely not, but you already knew that
I have always believed that you were telling the truth in your writings Dave, and that your POV was yours and not because of an infatuation with 3dfx(well, that took some time to get to know you). If I were sitting in a position of authority at nVidia and all I had to go by was reading your articles, I would have had you labled as a '3dfx zombie'.JB-
Giants.
BenSkywalker said:
I wrote the articles and can't even remember what they were about. Did we discuss speed in general or did w the article link it to specific typical game situations like flexible animation and dynamics ? Maybe I should read my own articles again one of these days
OOohhh... don't get me started on that game
Hope its not representative for TnL... K-
BenSkywalker said:
I fully agree. Dave has yet to realize how good he is.
Now if I can just get him to stop being the king of run-on sentences.
That's a little harsh, Ben, but you're entitled to your opinion. However, consider this: a site can either parrot all IHV's marketing or they can question things, probe, examine, etc. The key is in fairly doing so with all companies involved. You obviously feel this wasn't the case in a lot of B3D articles, an opinion that's shared by others. Fair enough. However, there will always be IHV's who are more aggressive with their PR and marketing, less truthful with their claims for product performance. And I say this without suggestion of unethical/immoral conduct. It is, afterall, the job of a good PR employee to put subtle, yet insinuative, spins on performance claims. So an online journalist can either copy 'n paste company-issued marketing documents and claims--which, let's face it, explains 95% of all PC hardware reviews of the past five years--or he/she can try to get at the veracity of such claims. The latter certainly isn't easy, nor is it necessarily healthy for the growth of your site. And IHV's can be very resistant to this, some obviously more than others. They want their PR/marketing regurgitated to the masses and favor any site willing to do so.
I'm guess I'm trying to say that I think Dave Barron was being honest when he wrote that he and K were in agreement with 3dfx's philosophy/approach to what was needed more in '99 than with Nvidia. Their writings therefore reflected this. I don't know about most others, but personally I would rather have a reviewer take a stand on issues and offer subjective opinion than read another barrage of endless 3DMark/Quake benchmark scores buried within regurgitated marketing documentation. Whether or not the proferred opinion is correct, time will prove. The key, as I wrote above, is in applying established testing standards fairly to all products.
Kristof said:
dithering can produce patterns of arbitrary-high contrast if the original color which is being approximated happens to have closest neighbours of such contrast (WRT each other) along the approximation curve in the reduced color space.
of course R5G6B5 is not monochrome, but it doesn't have that great a resolution along the R/G/B axes either. anyhow, nobody would have mentioned the 'nasty artefacts of dithering' on the voodoos if the contrast across the dithered pattern was below the threshold of human perception in the first place, no?
so at the end of the day we get the fact of visibly high-contrast dithered patters, which are to be countered by smoothing filters, which, on their turn, try to avoid smoothing high contrasts. back to square 1.
High contrast was to avoid that true texture details and polygon edges would get blurred which is not the idea of the post filter although they could have claimed edge AA with it
apparently.
umm, just as common as high-contrast details.
For what its worth I am trying to get back to technical article writing and will very hard try not to say that company X, Y or Z **cks
i'll be expecting those with great anticipation
The point you don't want to understand Dave is that 'looked better' is a subjective thing. What is better to you may be not to me. Many people thought you were biased toward 3Dfx, and I believe after your last comment they have consolidated their beliefsDave said:
In fact, in my humble opinion, you and Kristof were and are plaing wrong on all the T&L usefulness debate. That doesn't mean this view of mine is the only truth and that all the other people have to perceive things as I do or otherwise they are all wrong.
ciao,
Marco
- Status
