Here's a video I stumbled upon which nicely showcases it.
http://www.youtube.com/watch?v=aQa97-ApWvc
The framerates aren't exactly 1:1 for comparison's sake but I think the problem is obvious to see. If you really need 1:1 check ~2:20 into the video where it zooms in on the dragon from above. Both cards running 22fps there and the difference is quite clear.
For once, I can actually see what you're trying to point out!
Anyway, while it's blatantly obvious in that video at that time point (that's all I looked at), is there any side-by-side with both cards at 60fps+?
I think Youtube only renders video in 30fps anyway. The difference would probably be lost there.
That's the thing, if the stuttering is lost at 60fps+, people with CF/SLI are in all likelihood playing their games at 60fps+ and aren't going to see this stuttering anyway.
I personally thought the stuttering was gone only at 60fps vsynced. Vsync disabled, it was there even at 60fps+ framerates. Obviously not that bad, but single card was noticeably smoother. But let's not get into that until one or the other can prove something.
Anyway, there's always games that mangle crossfire/SLI combos just as well. Let's take Metro 2033 or Stalker games, or SC: Conviction or Starcraft 2. Not to mention your GPU's will age just as any, and you'll have to upgrade quickly to keep running over 60fps or any other magical limit where microstuttering supposedly disappears.
Not bashing on Multi-GPU though. I was quite pleased with my 4890 crossfire overall. Just that there's some real problems involved with it that should be kept in mind, especially when recommending it to people who have no idea.
To me, the horrid input lag of v-sync far outweighs any supposed 60fps+ microstutter (I notice none).
People should be aware of this potential issue, but it's so subjective at 60fps+ that people must try it for themselves.
You're right about Metro 2033, that game taxed my cards harder than Crysis. Not that it wasn't using SLI (because it was fully utilizing both gpus), it was just that demanding of a game.
There's something wrong with the video, for GF is seems it doesn't matter what the FPS is and it's just as smooth always, while everyone knows that at under 20 FPS, it just isn't smooth no matter if it "microstutters" or not.
While the video is good for "illustrating microstuttering" or what it might look like, it's not good for anything else - 2 different video sources with different original FPS recorded, then encoded into 30 FPS (or whatever youtube uses) video just isn't going to show what you'd see on your screen (even if you'd record the original at 30 FPS with FRAPS, if the FPS on screen isn't dividable even (you know, 30/15/10/5 etc) it just won't show what you have on your screen like it should.
22 frames per second divided into 30 frames per second for example just doesn't work.
It's not strange CB, video is a perfect example, it's a good vid showing Microstuttering and interesting to see.
Regardless of video framerates etc. it's fair to assume both were recorded with the same method and went through the same encoding processes. The difference, I think, speaks for itself.
It's a good vid to show how it should look like - however it doesn't show actual microstutteringOriginally Posted by Silent_Killer
Regardless of what causes the anomalies etc, it's just simple video capturing / shooting facts that prove it won't show what's really happening on screen,
This is filmed on separate videocamera, meaning most likely 24, 25, 29.97 or 30 FPS (or Hz, whichever you prefer).
This camera, shooting at one of those rates, is videotaping screen running at 59.97 Hz (commonly mentioned as 59 (Win7) or 60 (earlier Windows versions) Hz), which has content running at anything from 0 to infinite FPS (real limit being of course 59.97 as the screen just won't show more)
In any occasion that what's happening on the screen is not dividable with the video refresh rate, it won't show correctly.
Then all this is of course as the last step encoded into 30 FPS video for youtube, which may screw it up even more
Posting Permissions
Reply With Quote
