Overclock.net › Forums › Industry News › Hardware News › [PCPer] Frame Rating Part 3: First Results
New Posts  All Forums:Forum Nav:

[PCPer] Frame Rating Part 3: First Results - Page 10

post #91 of 294
Thread Starter 
Quote:
Originally Posted by twitchyzero View Post

is there like a noob's guide to reading these frame time graphs? It's quite messy.

Like what's better and what's worse. As long as the discrepancy between framerate and frametime is minimal stuttering is reduced?

aka as long as frametime is close to 16ms while fps is 60fps you are good to go?

Here was one of the first articles on it (recently) that kind of explains what it is. What you would like to see, ideally, is all the frames taking the same time to render and display - which would be completely "smooth". It's the difference in frame time between adjacent frames that looks like stuttering to the eye. So a 16ms frame followed by a 40ms frame followed by another 16ms frame would look strange when you see it, as a kind of hitch or stutter. On the charts you would like to see a relatively smooth line, one that has a lot of spikes implies that there is a lot of difference in the frame times, and so stutter (whether it is perceptible or not is a different discussion).

http://techreport.com/review/21516/inside-the-second-a-new-look-at-game-benchmarking
Edited by Forceman - 2/23/13 at 12:51pm
post #92 of 294
My main issue with all this, is that we're already certain that what we see in this test is what we get. Personally, I'll wait to see similar methods being used on multiple setups (mobos, cpu's) before I start passing judgement on whats really going on here. By the looks of it all of the testing was done using the x79 based Tiatan review test setup. I'm looking forward to seeing how they expand on it and if this issue goes deeper than "Amd really screwed up"
Bubbles
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5 3570k @4.7 GA-z77x-ud5h Crossfire 7950's / Heatkiller GPU-X³ 79X0 16gb Samsung 30nm  
Hard DriveHard DriveCoolingCooling
Samsung F3 OCZ Vertex 3 XSPC Raystorm EK Coolstream 360 + XSPC RX240 
OSMonitorKeyboardCase
Windows 8 Pro Crossover Q27 CM Quick Fire Rapid (Reds) NZXT Switch 810 
Mouse
Roccat Savu  
  hide details  
Reply
Bubbles
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5 3570k @4.7 GA-z77x-ud5h Crossfire 7950's / Heatkiller GPU-X³ 79X0 16gb Samsung 30nm  
Hard DriveHard DriveCoolingCooling
Samsung F3 OCZ Vertex 3 XSPC Raystorm EK Coolstream 360 + XSPC RX240 
OSMonitorKeyboardCase
Windows 8 Pro Crossover Q27 CM Quick Fire Rapid (Reds) NZXT Switch 810 
Mouse
Roccat Savu  
  hide details  
Reply
post #93 of 294
I think this all boils down to the "if a tree falls in the forest...." argument..

Basically, if the frame time delta is so small, you can't notice it (ie: less than 1ms to 10ms), is it really an issue? Some may say yes, others say no.

Remember, 1ms is essentially 1000 FPS while 10ms is 100 FPS . So can you really see it when your framerate fluctuates between 100 to 1000? I know I can't.

On the other hand that, if that same 10ms delta translated into say 40ms (25fps) to 30ms (33fps), you can be sure there will be noticeable stuttering.

It's all about perspective and the framerate at which frame time variation occurs.

Personally, I think it's necessary to benchmark a gameplay experience or in other words, what's right there on the monitor.
post #94 of 294
Quote:
Originally Posted by SKYMTL View Post

I think this all boils down to the "if a tree falls in the forest...." argument..

Basically, if the frame time delta is so small, you can't notice it (ie: less than 1ms to 10ms), is it really an issue? Some may say yes, others say no.

Remember, 1ms is essentially 1000 FPS while 10ms is 100 FPS . So can you really see it when your framerate fluctuates between 100 to 1000? I know I can't.

On the other hand that, if that same 10ms delta translated into say 40ms (25fps) to 30ms (33fps), you can be sure there will be noticeable stuttering.

It's all about perspective and the framerate at which frame time variation occurs.

Personally, I think it's necessary to benchmark a gameplay experience or in other words, what's right there on the monitor.

A camera that capture the monitor is better solution to compare frame latency.
Workstation
(4 items)
 
  
CPUMotherboardGraphicsMonitor
Xeon E5-2690 Supermicro 2011 Nvidia GP100/ Vega FE Dell ultrasharp 4k 
  hide details  
Reply
Workstation
(4 items)
 
  
CPUMotherboardGraphicsMonitor
Xeon E5-2690 Supermicro 2011 Nvidia GP100/ Vega FE Dell ultrasharp 4k 
  hide details  
Reply
post #95 of 294
Quote:
Originally Posted by jomama22 View Post

It seems that a 0ms frame is like the dropped frame. When a frame take 0ms, usually there was a previous frame that took a much higher amount , let's say 45ms. So what's is happening is you get 2 quiclk successive frames then a lag of 45ms, then another quick 2 frames, then another 45ms lag


That is basically micro stutter.

Not necessarily. Anything below .5ms will look like 0 in those charts due to their scale.

On the other hand, 0ms for frames may mean they were pre-rendered from buffered data and were displayed immediately after the last one was rendered.

NVIDIA actually has an option in their control panel which can modify this value. For those interested, try messing around with the pre rendered frame value in the control panel and see the effect that has on fluidity.
post #96 of 294
I would love to hear some official AMD response to this article.
post #97 of 294
Quote:
Originally Posted by SKYMTL View Post

Not necessarily. Anything below .5ms will look like 0 in those charts due to their scale.

On the other hand, 0ms for frames may mean they were pre-rendered from buffered data and were displayed immediately after the last one was rendered.

NVIDIA actually has an option in their control panel which can modify this value. For those interested, try messing around with the pre rendered frame value in the control panel and see the effect that has on fluidity.

That makes more sense lol.
     
CPUMotherboardGraphicsRAM
intel 3960x @4.7 ghz asrock extreme9 x79 3x msi 7970 lightnings g.skill ripjawx 1600 cl8 
Hard DriveOptical DriveCoolingMonitor
ssd: kingston hyperx 120gb hdd: seagate barracu... bd-rwh water for cpu benq 2420t 24" 120hz 
Case
mountain mods extended asension 
CPUMotherboardGraphicsRAM
3960x Asrock Extreme9 2x Sapphire 6950 unlocked G.skill CL8 12800 x4 (4gb) 
Hard DriveHard DriveOptical DriveCooling
Kingston hyperx  Seagate barracuda xt 7200/ sata3/ 64 mb cache LiteOn BD-R x12 Watercooling 
OSMonitorPowerCase
win 7 x64 asus 23" 1080p 1000W Thermaltake toughpower HAX 932 Advanced (modified for Radiators on ins... 
MouseAudio
Razer Deathhadder 3.5g sennheiser PC 330, Game blaster with Extreme9 
  hide details  
Reply
     
CPUMotherboardGraphicsRAM
intel 3960x @4.7 ghz asrock extreme9 x79 3x msi 7970 lightnings g.skill ripjawx 1600 cl8 
Hard DriveOptical DriveCoolingMonitor
ssd: kingston hyperx 120gb hdd: seagate barracu... bd-rwh water for cpu benq 2420t 24" 120hz 
Case
mountain mods extended asension 
CPUMotherboardGraphicsRAM
3960x Asrock Extreme9 2x Sapphire 6950 unlocked G.skill CL8 12800 x4 (4gb) 
Hard DriveHard DriveOptical DriveCooling
Kingston hyperx  Seagate barracuda xt 7200/ sata3/ 64 mb cache LiteOn BD-R x12 Watercooling 
OSMonitorPowerCase
win 7 x64 asus 23" 1080p 1000W Thermaltake toughpower HAX 932 Advanced (modified for Radiators on ins... 
MouseAudio
Razer Deathhadder 3.5g sennheiser PC 330, Game blaster with Extreme9 
  hide details  
Reply
post #98 of 294
Quote:
Originally Posted by sugarhell View Post

A camera that capture the monitor is better solution to compare frame latency.

A high speed camera will help but once again, it may very well pick up things which will never be visible to the human eye.
post #99 of 294
Quote:
Originally Posted by SKYMTL View Post

A high speed camera will help but once again, it may very well pick up things which will never be visible to the human eye.

Or we can use this camera and burn the whole house down!!!

http://www.ted.com/talks/ramesh_raskar_a_camera_that_takes_one_trillion_frames_per_second.html

There would be people who will say they can notice it lol
post #100 of 294
Quote:
Originally Posted by airisom2 View Post


Yeah, I don't know about you guys but this video demonstrates exactly how I play my games like Crysis and BF3. Slowly panning from one side to the other looking for the tiniest of millisecond hitches so I can run onto the forums and claim the game was simply unplayable. rolleyes.gif Maybe I get a little more engrossed in my games than you guys do....
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [PCPer] Frame Rating Part 3: First Results