I want your opinions and feedback

Seems a lot of people want apples to apples.

Fact is: Go somewhere else to find that. Just about every other tech site does it. We don't need another carbon copy. [H] does this on purpose and I like it personally. It really shows what kind of gaming experience you're going to get.
 
I think everyone else has hit the nail on the head - you guys are doing a great job.

One point of consideration: I definitely agree with playable testing - that is when you show the frame rates of one card with what it can do (as this is the most relevant data to get).

One thing I would still like to see however is a direct power comparison - when I buy a car I definitely want to at what maximum speed it is most comfortable driving at, however horsepower is relevant to many people.

With this analogy still having a comparison between two competing cards at the same resolution is really good - being able to see that one card is 20% faster doing the same task than another is relevant data (much like in cpu tests where they don't measure in "maximum multitasking comfortably doable").

Don't get me wrong, I agree with the direction of letting people know how the card will play a game, and at what level it will do so comfortably, however having even-tasked benchmarking as well would only satisfy all crowds.

Keep up the great work!
 
Intel_Hydralisk said:
It really shows what kind of gaming experience you're going to get.
Click to expand...

What it shows is the reviewer's opinion of what kind of gaming experience we're going to get.

While I still come to [H] first for reviews (I appreciate the usually no-nonsense approach to manufacturers and willingness to tell it as it is) I would just prefer to see my reviews showing maybe 5 or 6 appropriate resolution/aa settings for each game and each card compared at these levels so I can see how much 'raw' power one has over another. The subjective reviews idea isn't something I disagree with in principle; it just doesn't fit with what I want to see in a review.
 
I would agree with the above, out of all website reviews I would say this site has the best but in terms for improvment I agree, what would be good is to add how much power each card uses.
 
SPARTAN VI said:
Good idea. Lots of 2005fpw/2405fpw gamers, including myself. :D
Click to expand...


Yes , wide screen resolutions are certainly over due
 
I vote for more widescreen info too. Not just the widescreen LCD's out there but HDTV's. How the cards do at 720p, 1080i and 1080p resolutions would be great. More and more of us are gaming on out home theater setups, and once it's tried theres really no turning back.

peace.
unloaded
 
Heya.

Thanks for asking for feedback - it's always appreciated to feel like our opinions matter in ways other than pure page hits ;)

Like someone mentioned further up the postings, i'd really love to see a set of baseline "mainstream" systems used that don't change, just so that we can get an idea of what, why and when it's worth us upgrading other bit's of the system instead of the GPU.

I guess it's one of those things that I personally take the GPU stats as full system test for that gfx card and scale it down based on the cpu - and so far, i've got it badly wrong twice.
So it might be a bit more hassle, but how about baselineing a AMD 64, 3000+ , 512MB ram and using it as a baseline for all tests in the up and coming year and a bit? Means that we get a real "If i upgrade to this card, will it actually help my framerates if i'm stuck to my LCD's 1280x1024, or can i up the quality to max without the cpu bottlenecking it"?

For an example - Doom3.
It runs fine at 1280x1024 on my machine 3000+, 2GB Ram, 6600 at low quality.
Will getting a 7800GT be a total waste cos the CPU will still bottleneck the GPU's abiltiy to kick out quality - and should i just stump up for a 3800 X2 instead?

Yeah, i guess i want everything handed to me on a silver platter ;)

Cheers,
Wolfsbane2k
 
Despite what all the die hards say, Apples to Apples is what really counts.

The reason people go read reviews is because they cant purchase 10 cards at a time. My monitor runs 1280x1024 and I want to know what card performs the best at that resolution. Show me each card at that res with varying degrees of AA/AF and let me choose. DONT simultaneously graph one card at a lower resolution and lower AA/AF with one of higher image quality. You guys merely take your apples to apples data and graph varying cards at varying settings on the same chart. Thats not helpful. Its not helpful for me to see your results, then have to decipher them to determine things like.....oh wait, that card is running at 10x7 with no AA whilst the other is at 12x10 with 4xAA/4sAF. I cant think of a more unscientific way of representing data.

The reason all the other sites use apples to apples is because it works! Different isn't always better, especially just because it is different. Put your apples to oranges comparo opinions in your final conclusion paragraph, but leave the charts to show useful data.


PS. Every review should have a 9800pro and X800 as baselines for all us "mere mortals"
 
Brent,
for what you are doing here to the audiance I think your reviews are very nicely done. I would have to echo the widescreen numbers.

Also more IQ comparisons. I know I know some may say but you will not see them in a game. BS. Its the same thing as a scratch on your car, once you notice them, then you can still see them far away. Same thing. Once you notice things its then its easy to spot them even in action. Plus for a team game like CS/DOD you do spend some time dead in flyby where IQ issues really jump out a person. And finaly no two peoples eyes are the same..what one may not see, could be a huge issue to another..

But again you guys do a very good job with your reviews and I use this along with other sites to get a feeling of what a card can and will do...
 
I would say, that the biggest change that needs to be done is that instead of having it on a system where its a fresh install of windows and the only thing running is the game, put it on a system that 50+ processes running in the background, and have a majority of the system RAM taken up by these processes.

That is just my 2c, but it would really seperate the big H from everyone else
 
Ya guys i decided that im gonna run multiple Benchmarks on widescreen resoultions on my SLI 7800GTX 512 cards to see what kind of performance they get. I myself am a little curious, id love to go to a widescreen LCD for gaming but its hard to break away from my CRT NEC 22in moniter. Other then that i think that Hard OCP is one of the mose trusted websites for Hardware reviews. I know i dont have to worry about them catering to a company when it comes to there opinions, they are unbiased and exact to the letter. I generally wait on all hardware releases e.g. Motherboards, CPUS, Video Cards, before i make a buying decision.
 
I want to know what each card is capable of doing at my max resolution, which is 1280x1024 for a 17" LCD. I am sure a large number of us are using 17 or 19 inch LCDs at this point and seeing what a card will do at 1600x1200 just doesn't provide much usable information for me/us.
 
Elstaf said:
I want to know what each card is capable of doing at my max resolution, which is 1280x1024 for a 17" LCD. I am sure a large number of us are using 17 or 19 inch LCDs at this point and seeing what a card will do at 1600x1200 just doesn't provide much usable information for me/us.
Click to expand...

Yes it does .. if it can run 1600 x 1200 4XAA and UQ then you can run 1280 x 1024 MAX everything ...
 
Im sorry if this has been mentioned before as I have not read the entire thread. I really like the current review structure, but I would like to see more tests done at different resolutions.

At every opportunity the reviews are pushed up to 1600x1200. But this is pushed well out of range of many monitors, and is usually not a native resolution of gaming LCDs. Even my CRT will only go to 1280x1024 at a decent refresh rate.

Although 1600x1200 is a better test for added realism and performance, the maximum playable settings for cards at 1280x1024 would be interesting to know as well. For example, can you sacrifice that resolution to enable better Ansitropic or AA modes?
 
On that note, I sometimes use 1280x960 for a bit of extra 'oomph' and a 'square' picture since 1280x1024 is generally a bit skewed.
 
Add a benchmark at a widescreen resolution like 1920x1200 to really max out the high end cards, especially since some games are getting completely maxed out at 1600x1200 w/ Ultra high details, 4xAA, 8xAF, etc...

Add power consumption measured at wall for same configs, just different videocards.

Add case temps for different set ups to see how efficient the HSF are at exhausting the heat out of case, like in the dual slot coolers with exhaust vents, vs. standard single slots, etc...

I still like an apples to apples page, which I didn't notice one on the last review.
 
The "Playability" ratings were cool when the 3dMark controversy was going on...but give it a rest now and lets see some FPS.

OK im out of this thread too.
 
I'm sure I'm in the minority, but I would like to see more bar charts, and also more comparisons of the video card being reviewed with previously reviewed cards. I usually end up going to other websites for their reviewes because I can see a braoder spectrum of how that newer hardware fares. I realize that the rest of the review hardware will most likely not be the same, but it is still something that would be of great interest to me. It would help me determine which card will provide the performance I need at a price point I'm actually willing to pay.
 
I'd have to chime in with more widescreen benchmarks. I'd like to see how the new cards stack up at 19x12.

Also, I'd be interested in seeing more apples to apples benches.

And something I wouldn't mind see as an avid EQ2 player, is the benches run with shadows off. It's one of the biggest performance hits in the game and one of the first things players turn off to get better frame rates.

I'm not sure if they're CPU or video card intensive, but I'd love to see what level of detail an SLI rig could handle with thm off.
 
I agree with using some older cards. A point of reference would be nice, ie; 7800 is 10x faster than a 4600, 9600, 9800, whatever. Something to put it in perspective.

Just my thing, apples to apples comparison. 1600x1200x8xAAx16xAF, just pick one for a direct comparison. The graphs are not my favorite part and the lack of direct comparison (image quality is subjective) and running one card at 1600x1200 w/AA/AF vs 1280x1024 w/AF only, yea we know which is faster, but I want a true side by side comparison.

As it is lately, I've been reading other reviews more often. I don't have the time to try to remember which is running what res in what game on what graph. At least one comparison in the graph needs to be a direct comparison.
 
Intel_Hydralisk said:
Seems a lot of people want apples to apples.

Fact is: Go somewhere else to find that. Just about every other tech site does it. We don't need another carbon copy. [H] does this on purpose and I like it personally. It really shows what kind of gaming experience you're going to get.
Click to expand...
It isn't without its uses, but if I don't read the review because I want apples to apples, their advertisers won't pay for my click.

The old style reviews were preferred and H was my first review to read. Now, I tend to go to Anandtech to get the quick answer and don't read the H reviews as much anymore.
 
Powerhouse2k said:
6. Updates to the review (or perhaps a whole new Section/Review area) that include new Drivers with their performance (gain/loss). An example would be like this...

CPU: FX-60 with 2Gigs of Memory, WindowsXP 32bit version
Video Cards...................................Min. FPS...Max. FPS...Avg.FPS...Playable Settings
Drivers used: Forceware 81.94 for Nvidia Based Cards, Catalyst 5.11 for ATI Cards....
Click to expand...
I agree to keep some kind of graph section that holds data from all the graphs that you guys create for reviews. I also agree with alot of the thoughts expressed though except with one change. Most of these ideas would be great as their own section/review/article and would be too much if added to an existing article. The driver evaluations is a prime example of this, having a separate article for that would be better than adding it to an otherwise lengthy review/preview.

One thing i disagree about is everyone asking for more resolutions, save for WS resolutions. All these graphs and data and stuff that kyle/brent/whoever puts out for these cards are great. It seems like you guys dont want to do any more work besides lookin at th enumbers, If you know A card did A numbers at X resoution and Z resolution, im sure you can use a few brain cells and figure out how it performs under Y resolution. Same thing goes with the want for more cards in a review.

I have been reading H for 5+ years ever since i made fun of my freshmen year roomate for reading such a geeky site. The reviews/previews/articles presented always gave me more than enough information and if you didnt explicitly mention certain aspects of a hardware's performance it was easy to figure it out from all the other data provided.

And yes bring back 3dmark! But dont add it to your video card reviews, it would be better served as a separate article. Maybe "Real World VS Synthetics" would be nice.
 
0ldman said:
The old style reviews were preferred and H was my first review to read. Now, I tend to go to Anandtech to get the quick answer and don't read the H reviews as much anymore.
Click to expand...
thats funny, the same thing happened to me, vice versa.....
 
Well, I liked the extra section in at the end where you used to have "apples to apples"...it doesn't tell you anything except how one card stacks up to the rest, but when I am in a hurry, I would skip the rest of the benchmarks, and go right to that section.
 
I still think apples to apples is a valid comparison. I think it might help to give a better overlook on how the card fare over on another,and might give an insight to future games. I also question for instance, that one is shown say at 2xAA 8XAF, and another is at 4xAA 8XAF maybe at lower framerates, although it may be playable, is it really adding that much to the gameplay? Is the higher AA really making a visual difference? If you lower it to the same settings, is the one that could enable higher AA faster?
I like the way [H] reveiws the video boards, but to say shimmering does not really impact gameplay so its a non issue, does one more level of AA really impact the experience? I think that by you showing framerates, it may be throwing off some people into thinking that one board is slower even though it has higher settings, so apples to apples may help them. I understand that in your conclusions you sum all that up, but I personally know many people that just want to see the numbers.
 
You guys do an excellent job with your vid-card evals. That said, there's one thing I'd love for you to do that I think no-one else does on a regular basis: evaluate video playback for quality/CPU usage.

I know this is a gaming site, but we don't game all the time. Some of us play back MPEG-2/4 or HD WMV video in our spare time, and it has been proven that quality varies from card to card. Some cards don't interlace properly. Some are buggy, resulting in abnormally high CPU usage (i.e., early PureVideo. I'd really like to know when I buy a vid-card that I'm going to get quality video playback in addition to good gaming performance.

Thanks for taking the time to listen to reader input.
 
A justification for little apples to apples comaprisons is (well at least this generation) the remarkably different solutions that ati and nvidia have took for the 7x00 and X1x00 series.
 
I think some1 already mentioned this...

I play on a 19 inch LCD @ 1280x1024 would be nice to to know what the max AA/AF setting is at that res. I had to go to another site to see a comparison hench I'm gonna order 2x7800 GTX OC's instead of blowin over £1000 on the 512's which I thought about doin....

maybe include 10x7 and 12x10 for us plebs without the 16x12's ;)
 
It seems that people want reviews to cover the following:

Packaging/Bundle
Construction Quality
Noise
OC Speeds
Temperatures
AA Performance
AF Performance
DX8 (Older Generation Performance) Analysis with Many Games
DX9 (Current Generation Performance) Analysis with Many Games
CPU Scaling (Int/Amd)
Ram Scaling
Common LCD Resolutions
Common Wide Screen Resolutions
Ultra HiRes for those with nice monitors
Comparison to older generation video cards
Value Analysis
Tech Support

And a free lollypop.


I know I suggested this before, but the best way (IMHO) to do this is to go really in-depth with the reference cards and for cards from the resellers, just report the deltas as they pertain to the card. That way, you can cover most of what was outlined above ^^^^^


M
 
Sims2, though widely popular, is not something that people are buying a 7800gtx for. If I buy a 7800gtx, I am buying it under the assumption that Sims2 will run maxed out at roughly 60fps. Again, though widely popular, I don't see any high end video cards flying off the shelves to play Sims2.

I also think heat and noise measurements are badly needed. I know my current x800xt heats up my case pretty well when it's oc'd, and I'd like to buy a video card knowing how my case temps will be impacted.
 
DocFaustus said:
Brent,

The thing that stands out most to me is Tables below each chart. Right now they are Setup to be 5 columns stating: Card, MinFPS, MaxFPS, AvgFPS, Playable Settings. The Charts are supported by the tables below, the tables are in turn supported by the paragraph to follow. So, it is easier to read them all from bottom to top.

I believe the Tables could stand on their own much better if you added a 6th column that simply stated the difference in settings. Examples: EverQuest2 the 6th column would say “High Quality” or “Balanced” for game setting. FEAR would have “+soft shadows” if they fixed the AA bug in the future and only 1 of the cards in the test could run it fine that way.

Settings that are the same for all cards in the comparison wouldn’t need notes in the table as they will be stated well enough in the paragraph that follows “..all cards played very well with textures at extreme”

I believe this would make the tables much more meaningful to your readers and hardly change the way things are being done by you.
Click to expand...

I second that....

Also, NVidida and ATI seem to be going down different paths as to how they improve image quality. Seeing as how your evaluations have become more and more subjective (not that I mind - but playable settings for you might be different than playable settings for me), How about an "I enjoyed this game most on this graphics card at this resolution" (or there's really no difference bewteen the two cards in this game).

You already do that a little bit, but I would like to see it said more forcefully. Especially because maybe the highest playable setting is not where you would play the game or maybe in the future a lower resolution on card A is actually nicer than a higher resolution on card B because of AA and AF.

Me.
 
I also would like to see more 1280x1024 benchies, because that's a popular resolution with a lot of the average-sized LCD screens.

I also agree whole-heartedly with everyone else who wants to see more apples to apples comparisons. I honestly do appreciate how much [H] does to be different and does things for gamers, but I just dont get the information I'm looking for when I read video card reviews here. For each game, I want to see a level playing field (as in same res, same settings, same everything) and see how the cards stack up from there. Seeing two cards at different settings showing different FPS scores doesn't give me a lot of confidence about what's being reviewed. I love [H]ard|OCP to death, but honestly, when a video card review is posted, I read the introduction (which is always well written), go to AnandTech for a very well done, IMO, apples to apples comparison, then back to the [H] for the conclusion.

I may be in the minority, but my ideal situation would be to see some balanced apples to apples benchmarks with the addition of the "best playable settings" opinion. I want to be able to decide what would be best for me, not have it given to me by what someone else with different taste thinks.

Thank you for asking, Brent, and I honestly do appreciate all you do for the community, even though I may disagree with some parts of it.
 
I love your format and just wanted to ask that you NEVER stop including Everquest 2 in your evaluations. You’re the only review site that consistently gives good info on how new hardware affects EQ2. On that note, I would make one comment. EQ2 due to the nature of MMORPG's does not need mega FPS to be playable. I know that what is considered "playable" by one person may be a show stopper or another. I play at 1600x1200 2xAA (with adaptive AA on in performance mode), 8x AF trilinear filtering, in either High quality mode or Balanced depending on whether I am in town/group or soloing. I have found the game very playable at these settings and absolutely gorgeous on my X800XT PE AGP card in my IC-7G P4 system. I am overclocking the FSB to 880MHz.

I would challenge you to try tweaking the MANY settings for two different test profiles. One that aims for highest eye candy with minimum acceptable fps based on actual play not first person shooter standards. (my preferred method of playing) and one for best frame rate with minimum acceptable eye candy. Present a detailed list of your settings for each profile and the results of your testing on them.

I know that many EQ2 players use a simple mod that allows you to store settings profiles and load them on the fly in game. I do this and have a HQ mode for solo/small group play and a LQ mode for in group play where FPS is more important. this gives me the eye candy I crave and sacrifices it when performance is critical. I know I could use this information to help me decide what my next PC will be build with.

Thanks guys for doing a GREAT job on reviews!!!!! I have been a lurker in the [H]ard forums and the site for years and appreciate all you do.
 
jbailey said:
I'm sure I'm in the minority, but I would like to see more bar charts, and also more comparisons of the video card being reviewed with previously reviewed cards. I usually end up going to other websites for their reviewes because I can see a braoder spectrum of how that newer hardware fares. I realize that the rest of the review hardware will most likely not be the same, but it is still something that would be of great interest to me. It would help me determine which card will provide the performance I need at a price point I'm actually willing to pay.
Click to expand...

Oh yeah.... that too! :)

It would be awesome if you did a nice big round of of videocards, with a fast say FX-57 system (AGP & PCIe) and continue using that system for a while, even if they release an FX-60, so results will be comparable. So people with say an older 9800 Pro, could see what he would gain from an upgrade of everything the same but videocard.
 
Some folks might be getting the wrong idea about people asking for apples to apples results. Some of us wanting apples to apples (especially when graphing ;) ) does not mean we don't want the real world game play eval, this sets [H]ardOCP above the pack. We just want a bit [M]ore at the [H]ardest review site on the [P]lanet :p
 
I am looking for:

If I were to upgrade from the previous version video card. For example, lets say I have a 6800GT, and the 7800 series video cards came out. How much of an upgrade would it be?

Or if I had a 7800 and an 8800 video card comes out.....


See what I'm saying?
 
First, I am exceedingly glad that you guys stopped releasing your test demo recordings a year or two ago and have stuck to your guns on that issue. It removes optimizations from the equation as far as is possible without modifying the drivers. Kudos. :D

Second, I like that since the FX fiasco, you have been much more even handed when it is just a small mistake versus something major like the FX line and CrossFire. (I have a 5950 by the way and have been mildly impressed with it's abilities with the latest games. I often find I can turn all the settings to High and still play at 1024x768 in Quake 4 with an average frame rate of around the mid 40s.) In the past it seemed like if one thing was slightly less capable on one side of a comparison, then it was the red(or green)-headed step-child of Satan.

Now I do have a few detractions:
1) Graphs of non apples-2-apples comparisons are too confusing.
I don't know about others, but I don't want to see numbers in a comparison where one of the competitors is artificially limited more than another. With most of the recent reviews, I have left thinking that ATI had won out on a particular test because they had a higher framrate only to come back and find out that they had lower settings so they actually lost.
If a particular setup was playable (beat a certain fps limit) at certain setting, that is all you need to know. You don't need to see a graph of it. "Card A was playable with such and such turned on and card B was playable such and such turned on at a lower resolution," inherently proves that Card A is better than card B in that test. The graph just confuses things.
Suggestion: Only do framerate graphs and charts for apples-2-apples comparisons.

2) No help for generation leapers.
I usually skip a generation with my hardware upgrades if what I have is good enough. What I find is that usually 2 years is plenty of time for the hardware to have progressed far enough (2 generations usually) for me to shell out $2500 ($3200 w/monitor) for a new rig. However, because of this strategy, I have no easy frame of reference with your reviews, with regard to performance increase, seeing as I don't have any of the parts that are not under review in the test platform.
Suggestion: Throw in a top of the line part and a bottom of the line part from 2 years back (when possible) and we can figure out our particular part from the two. This allows us generation leapers to get some comparison with what we have even though we don't have the test bed system.
I understand that in some cases, especially CPUs, this is not possible. That's fine. But with most everything it is possible at some point.

Otherwise, thank you for the enlightening reviews. They are always a joy to read, even if I'm not currently in the market for that part.
 
I have one suggestion, you can take it or leave it.

Use multiple systems for the benchmarks. Use one relatively high end system, one "sweet spot" system, and one budget system on each review (although standardize them; don't change them everytime, maybe just once every year). Maybe even include one Dual Core system; that would show what folks would get with THEIR configuration. After all, I know that the 7800GTX is going to be faster than the X1600XL. However, how much fast on MY system (which at one time was high end, and is now "Budget")? That's useful info.. and would allow us to decide if the two extra fps on the system are worth the extra cost. It would also let us see what the sweet spots are easier...

Just a thought... but of course, it quadruples the actual amount of work to benchmark (though not video quality test) the card. Might also pick up some of the weird quirks, too.
 
You must log in or register to reply here.
Back
Top