Reaction to Radeon

Positive Overall

I got a lot of reaction on that article.

About 60% of the responses elaborated on the theme “Thank God somebody said it.”
About 25% of the responses elaborated on some technical theme they didn’t think I got quite right.
About 15% of the responses elaborated on the theme “You’re an idiot.”

I must admit the last have been the most entertaining.

My favorite line so far was the one that said that folks like me “are a threat to freedom of speech and
a threat to democracy.” Move over, Osama. 🙂

Second place goes to defining of “journalistic integrity” as “covering up for other journalists.”

Let me address a few of the themes addressed in the emails:

“You’re a nVidiot”

Some were certain I had some connection or reason to favor nVidia.

This is rather funny given that I’ve been using a Radeon 8500 the last few months, and have found it OK. The last
nVidia card I used regularly was the Riva TNT (in between have been Matrox cards).

By the way, I bought and paid for that Radeon 8500. Overclockers.com is on no reviewer list for video cards. I don’t
believe we have even tried to get on any such video card list.

The Radeon 9700 had been a card I had been planning on buying. The marketing assault turned me off the card (as it
did many of you, more on that later), which I think is quite the opposite of what ATI wanted.

After some thought, I decided I’m going to have to buy the card just to make sure certain items get verified.

Negative Waves?

I know, in the heat of my wrath, I talked about being the “anti-ho” and just talk about the negatives. Even when mad, that
was more a rhetorical statement than anything else, but it illustrates a big problem we have living in the Land of Hype.

Imagine your friends want you to date somebody. She looks and dresses like Linda Cardellini as Velma in Scooby Doo. Your
“friends” tell you she looks like Jessica Simpson, only better.

I’m one of your friends, too. I think you ought to know the
truth, after all, you will the second you see her. Maybe you’ll like Velma, maybe you won’t, but if you’re expecting Jessica
Simpson Plus, you’re going to be in for a letdown.

When I try to tell you the truth, don’t you think I’m going to emphasize the negative just to counteract what you’ve been fed already?
That doesn’t mean I think Velma is ugly or undesirable, just that she isn’t Jessica Simpson Plus.

That’s why I sound so negative at times. When there’s a tidal wave of marketing hype coming right at you, you react strongly the other way to get some sort of balance in the overall environment.

Resolution

The vast majority of benchmarks and benchmark descriptions we’ve seen have been occurring at 1600X1200.

Since, in earlier surveys, relatively few of you indicated you used 1600X1200 in gaming, I pointed this out, and also noted from the very few sub-1600X1200 benchmarks that the card didn’t really kick in until higher resolutions and/or additional video processing were used.

The probable reason for this is the doubled bandwidth of the Radeon 9700. At the point where nVidia cards start getting handicapped due to lack of bandwidth, the Radeon has some more headroom.

But not as much as some benchmarks would have you think.

The problem with the ATI-dictated benchmarks is that they’re mostly taken at a point where the nVidia card is crippled, while the Radeon is merely handicapped.

For instance, let’s look at the frame rate for the Ti4600 on one benchmark reported by Anandtech:

Unreal Tournament 2003 w/ Anisotropic Filtering
dm-antalus @ 1024×768
Ti4600: 45.1

Unreal Tournament 2003 w/ Anti-Aliasing
dm-antalus @ 1024×768
Ti4600: 42.8

Unreal Tournament 2003 w/ AA+Aniso
dm-antalus @ 1024×768
Ti4600: 31.1

These are fairly low, and this is at only 1024X768 resolution.

I know a lot of you don’t believe in analysis and extrapolation and estimates, so I’m not going to go into what I did to come up with the numbers below, except to say it’s called math, and only math.

Just take this as a prediction, and see how close or far off I am later on.

When someone runs these tests and provides framerates, I predict the frame rate for the first two items mentioned above at 1600X1200 will be something like 18-20 fps.

For the last, I predict it will probably be a bit less than 15 fps.

So I’m predicting the numbers in a comparison between the Ti4600 and Radeon 9700 will look roughly like this:

Unreal Tournament 2003 w/ Anisotropic Filtering
dm-antalus @ 1600×1200
Ti4600: 19
Radeon 9700: 47

Unreal Tournament 2003 w/ Anti-Aliasing
dm-antalus @ 1600×1200
Ti4600: 18
Radeon 9700: 45

Unreal Tournament 2003 w/ AA+Aniso
dm-antalus @ 1600×1200
Ti4600: 14
Radeon 9700: 35

Looks a lot less impressive than “250% of Ti4600, doesn’t it?”

That’s hardly an advertisement for the Ti4600. Obviously the Radeon does much better, but is 35-45 fps enough for you to maybe buy a new monitor, or at the least spend $350+ for this card and go with 1600X1200 and the trimmings?

Anandtech didn’t think so:

“At 1600×1200 with 4X AA enabled, the Radeon 9700 is 2.51x faster than the GeForce4 Ti 4600. While the frame rates are high enough to be playable, even the almighty Radeon 9700 would need to back down to 1280×1024 with 4X AA enabled to be smooth as silk.”

Please note something very important here. In this particular comparison, the Ti4600 is on the other and wrong side of the 24 fps barrier. That’s the point where you get the illusion of continuous motion from frames.

Imagine you’re a reviewer. You’re either brought in to see, or “allowed” to test at 1600X1200 with additional fixings: Ti4600 vs. Radeon.

Ti4600 looks like Keystone Cops at sub-20 or even 15 fps, while the Radeon looks fine running the demo at 35-50 fps. That’s a dramatic difference. That might “blow away” somebody.

But the Radeon’s performance, while certainly a whole lot better than the Ti4600, is likely not quite good enough for serious gaming at 1600X1200, where all the benchmarks were taken.

This brings up the next big question. Why were all these AA and anisotropic tests run at just 1600X1200? I see one very likely and a second possible reason.

The very likely reason is that running at a lower resolution side-by-side just is nowhere near as dramatic.

If you’re running a side-by-side comparison at 1024X768, 45fps vs. 90 or 100 or 110 fps is unlikely to “blow people away.” However, those numbers would probably be a hell of a lot more relevant and useful to the average serious gamer. He would figure, correctly, that he could go to 1280X1024 with all the fixings and still be in good shape.)

The less-than-probable, more-than-possible reason is that the differential ratios drop a lot with resolution. They’ll almost certainly drop somewhat due to the lower bandwidth required. The big question (which is the one most gamers need to know a lot more than how it does at 1600X1200) is: By how much?

Some of you basically told me that they saw the Radeon 9700 as the card powerful enough to bring anti-aliasing and anisotropic filtering to the masses. That’s a very understandable sentiment, and a powerful argument in favor of the card.

So why didn’t ATI pursue it by allowing a couple benchmarks with all the goodies to be run at 1024X768 or 1280X1024? That would have been a killer argument for many gamers.

Either ATI was just very, very stupid, or those numbers drop a whole lot at lower resolutions. It’s certainly something to be looked into, and one we will look into once we buy the card.

Designated Drivers, Not Cheerleaders

There is this notion floating out there these days that reviewers should be all excited and enthusiastic about products. My God, a marketeer must have made that one up. Wrong.

Reviewers should be the designated drivers of any product party, not the first on the joy juice line.

Reviewers shouldn’t get excited, either. Reviewers are supposed to judge a product. Judges don’t get excited; cheerleaders do. When you get excited, you don’t think as well. Ask any seduced cheerleader. She got excited, and look what happened to her. 🙂

Do you know what would excite me? A working holodeck. That’s about what it would take, forget anything less in this field.

But I don’t have to get excited to like, or approve, or be impressed by a product. I just don’t need to go nuts to do so.

I mean, really. If you need serious advice about something, do you look for the most crazed person you can find to get it? So why do it here?

There are times when you’d like to see some enthusiasm, but you want that enthusiasm directly towards you, not some product.

To me, buying a video card is like buying an air conditioner, which is something I did recently. You look and think and compare, and buy what is best for your particular needs.

In my air-conditioner research, I saw little excitement even from buyer comments, and none from professional reviewers. This was perfectly fine by me; I didn’t want excitement, I wanted information and thoughtful recommendation.

So I came, I saw, I bought it. Did it get me excited? No. I didn’t want to get hot and bothered over it, actually, quite the opposite. 🙂

I didn’t want to get excited about it, I wanted to be pleased by it. Was I? Yes. That’s all I wanted, that’s all I expected, and I was quite satisfied with what I got.

I remember going from a Matrox G450 to a Radeon 8500. The difference between the two in gaming was dramatic, even more dramatic than any Ti/R300 pony show.

I didn’t foam at the mouth, or start uncontrollably spasming and jerking around like a Holy Roller, either physically or verbally. I just thought to myself, “This is pretty good.” But then, I expected it, just like I would have expected the Radeon 9700 to do very well under conditions that up to now have been bandwidth-constricted.

Had I written a review at the time, I would have strongly recommended the Radeon 8500 over the G450—for gaming. I wouldn’t have recommended it for everybody and everything, because the Matrox is as good or better than the Radeon for everything else but gaming. But then, I was just very impressed, not virtually insane. I wasn’t an excited cheerleader.

Save the excitement for after the purchase, when you’re having fun with the product.

At The Preview

Let’s pretend for argument’s sake I had been invited to this “preview.”

I would have emphasized the limitation, controls, and lack of freedom of the whole exercise.

I probably would have asked if a controlled environment was exactly the smartest move in the world after the Quack 3 debacle.

I would have told you the questions/complaints I had, and the answers to them.

I doubt I would have just regurgitated the limited benchmarks given me; I would have wanted a more representative range of resolutions. If I were refused, I would have pointed out the refusal and most likely would have given some general impressions, but emphasized the issues I’ve been raising.

If I ran or saw some benchmarks where the Radeon 9700 didn’t do all too well, I wouldn’t have blamed the benchmark for doing something wrong, I would have pointed out that if you played that particular game a lot, maybe this card isn’t meant for you.

If I had seen a crawling Ti4600 vs. a nice Radeon 9700, I would have known that ATI was exploiting the 24 fps barrier, and pointed that out.

In short, I would have been extremely skeptical, because that’s my job. A reviewer should be skeptical so you don’t have to be.

My job is not to make any manufacturer happy; it’s to make you happy. If I wanted to make ATI happy, I’d be honest about it and get hired by their marketing department.

Some of you are probably thinking, “That would be the last preview Ed would ever go to,” and you’re probably right. 🙂

But all that indicates is how sad the current environment is.

Let’s face it. This is about a video card, not the Manhattan Project. We’re not racing to beat the Nazis to the Bomb. It’s not even really typical news with a short-shelf life. It’s not like they’re only going to sell it for three days.

If you want something badly, that’s just what you’ll get.

I had a few folks say something to the effect of “you shouldn’t hurt [the other reviewers’] feelings.”

None of these folks seem to be concerned about the buyers’ feelings. When a buyer gets all wound up and excited from the hype and doesn’t read some of the fine print, buys the product, and finds out it doesn’t do what he had been led to believe it would do? Shouldn’t somebody be watching out for him?

(I got a few notes essentially telling me, “Isn’t capitalism wonderful? The suckers deserve it.”)

If I had to bet right now on the final outcome (presuming there’s no skeletons in the closet), I would bet I will end up saying something like the following:

“Primarily due to its doubled bandwidth, the Radeon 9700 can sometimes be a stellar performer. It doesn’t overclock terribly well, mainly because the 0.15 micron GPU is being pushed to the max. If you do A, B and C, and are willing to pay $350+ now, this is a good idea. If you do D, E and F, this is probably not a good idea, especially at the initial price.

“If you’re thinking, “well I’ll buy a Ti4600 or Radeon 9700″ when they get cheap, the Radeon 9700 will be a better choice, primarily due to the increased bandwidth. Also keep in mind that ATI prices tend to go klunk suddenly, much more so than nVidia cards. Whether it will be better than a cheaper NV30 remains to be seen. A 0.13 micron version of the GPU would be a nice upgrade, also. This could be a good Christmas present to those of you looking for a bargain video card.”

This is hardly a scorched-earth, stop-ATI-at-all-costs attitude, but God, it seems to qualify as one these days.

The reality is this is a very diverse group of people, and no one answer fits all. Even good products aren’t appropriate for everyone.

The Most Important Thing I Learned From This

The biggest surprise to me was not the attacks, not the “you idiot” notes, not the “I’ll never come here again” comments. I expected that, and actually expected more than I got.

No, the biggest surprise was the proportion of people who said how fed up they were with this sort of “journalism.” It wasn’t overwhelming, wasn’t even a majority, but it was significantly higher than I’ve seen in the past.

Some folks had better watch it and realize they’re on a sliding slope.

Ed