The Framerate Analysis Thread part 2

[brain_stew]

yeah but that were min and max come into play or even better standard deviation

Even that wouldn't help for examples like this. The gameplay is a perfect 60fps on both consoles, but none of those figures would tell you that. Forza's figures throw up similar oddities.

 

[grandmaster]

I think ps360's min/max/average is going to be the most manageable overall solution, but gameplay/cut-scenes stats should be kept separate - to be fair, ps360 does this.

 

[jaidek]

Yup, it's another example of how the "average frame rate" stat doesn't really work for me as an "at a glance" figure.

I agree, but if you watch their video or look at the gallery you will see they outline each clips frame rate and tear percents. As far as I know that is something that no one else has done. When I post their analysis' from now on I will reference the clip information. Here is what I mean:

WWE SmackDown vs. Raw 2010 Analysis

Clip 01: (in-game cinematic)
PlayStation 3 Avg. FPS: 29.87 / Xbox 360 Avg. FPS: 30.00
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 02: (in-game cinematic)
PlayStation 3 Avg. FPS: 30.00 / Xbox 360 Avg. FPS: 30.00
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 03: (gameplay)
PlayStation 3 Avg. FPS: 60.00 / Xbox 360 Avg. FPS: 60.00
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 04: (gameplay)
PlayStation 3 Avg. FPS: 60.00 / Xbox 360 Avg. FPS: 60.00
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 05: (gameplay)
PlayStation 3 Avg. FPS: 60.00 / Xbox 360 Avg. FPS: 60.00
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 06: (gameplay)
PlayStation 3 Avg. FPS: 60.00 / Xbox 360 Avg. FPS: 60.00
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 07: (gameplay)
PlayStation 3 Avg. FPS: 60.00 / Xbox 360 Avg. FPS: 59.66
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Global Average:
PlayStation 3 Avg. FPS: 47.79 / Xbox 360 Avg. FPS: 47.78
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%​

PS360's min/max/avg works really well because he deals with individual gameplay captures. That doesn't work so well when you have many different gameplay clips in one video. The method LoT chose allows the viewer to see each of the subclip stats individually so you can see where the console shines or fails for that particular clip. Then they provide a global average at the end for the entire analysis, which normally works well for games that stay around 30/60 fps.

 

[grandmaster]

I think it is fair to say that on this thread, people are just looking for accurate data which - if interesting - is worthy of clicking through to watch the full video.

That being the case, in the case of your work, simply lumping all your gameplay and cutscene captures into separate averages would probably be more helpful.

 

[jaidek]

Call of Duty: Modern Warfare 2 Analysis

Clip 1:
PlayStation 3 Avg. FPS: 51.88 / Xbox 360 Avg. FPS: 55.04
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 2:
PlayStation 3 Avg. FPS: 47.00 / Xbox 360 Avg. FPS: 54.45
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 3:
PlayStation 3 Avg. FPS: 50.17 / Xbox 360 Avg. FPS: 57.53
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 4:
PlayStation 3 Avg. FPS: 59.45 / Xbox 360 Avg. FPS: 59.96
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 5:
PlayStation 3 Avg. FPS: 58.78 / Xbox 360 Avg. FPS: 59.60
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 6:
PlayStation 3 Avg. FPS: 60.00 / Xbox 360 Avg. FPS: 59.98
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 7:
PlayStation 3 Avg. FPS: 59.10 / Xbox 360 Avg. FPS: 59.98
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 8:
PlayStation 3 Avg. FPS: 58.42 / Xbox 360 Avg. FPS: 59.65
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Clip 9:
PlayStation 3 Avg. FPS: 59.11 / Xbox 360 Avg. FPS: 59.91
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%

Global:
PlayStation 3 Avg. FPS: 53.73 / Xbox 360 Avg. FPS: 57.31
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%​

Lens of Truth: Call of Duty: Modern Warfare 2 Analysis

 

[assurdum]

Fps seems has same differences of the first...well, less drastical to be honest.
Global: PlayStation 3 Avg. FPS: 53.73 / Xbox 360 Avg. FPS: 57.31
PlayStation 3 Avg. Frame Tear: 0.0% / Xbox 360 Avg. Frame Tear: 0.0%
Plus bloom effects on 360. ot: I'm laugh about some comments in that comparison, the dithering artifacts in the shadows on the 360 version are exchanged for better rendering...oh jeez, come on....

 

[brain_stew]

Hmm, its interesting to see a game that's noted for its responsive controls and from a developer that stresses the importance of low input lag appears to be using triple buffering on both platforms.

 

[semitope]

Hmm, its interesting to see a game that's noted for its responsive controls and from a developer that stresses the importance of low input lag appears to be using triple buffering on both platforms.

triple buffering? Why not just vsync?

 

[Boxercide]

This argument has already come and gone.

 

[Alucardx23]

Hmm, its interesting to see a game that's noted for its responsive controls and from a developer that stresses the importance of low input lag appears to be using triple buffering on both platforms.

How can you notice when a game is using triple buffering instead of Vsync?

 

[brain_stew]

triple buffering? Why not just vsync?

The way the framerate fluctuates would suggest its using triple buffering for both versions. If it was straight double buffer v-sync, the framerate should "settle" at 30fps in a scene where it was consistantly failing to hit 60fps.#

Edit: See this post for a better explanation:

http://forum.beyond3d.com/showpost.php?p=1329602&postcount=516

 

[liolio]

You can't tell that easily between double or triple buffering. For example Uncharted solo is 30fps locked, how would you tell it's triple buffering against double buffering without ND speaking about their tech?

 

[nightshade]

^The answer to ur question is in Brain_stew's post above urs, Double buffered + Vsync means ur game will literally fluctuate between 30FPS & 20FPS (AS is the case with MGS4) whenever there is an FPS drop it'll go down to 20FPS.Having Triple buffering prevents this.

 

[grandmaster]

MW1 had the same latency as Burnout Paradise, which was definitely double-buffered, and I suspect it's the same with MW2...

 

[liolio]

^The answer to ur question is in Brain_stew's post above urs, Double buffered + Vsync means ur game will literally fluctuate between 30FPS & 20FPS (AS is the case with MGS4) whenever there is an FPS drop it'll go down to 20FPS.Having Triple buffering prevents this.

For your renderer yes, but a frame is issue to the display in the best case every 17ms (fps). So no matter the speed of the renderer will see frame-rate fluctuates between 60/30fps/20/15 fps.
I guess a devkit might tell you the speed of the renderer I'm unsure for a capture device does.
Triple buffering prevent you renderer to stall that's a neat gain. If the renderer runs fast enough (relatively to the display refresh rate) it even lowers input lag.

 

[PeterT]

Hmm, its interesting to see a game that's noted for its responsive controls and from a developer that stresses the importance of low input lag appears to be using triple buffering on both platforms.

Why is that surprising? Triple buffering (well, correctly implemented triple buffering) can only ever increase responsiveness compared to v-sync'ed double-buffering, not decrease it.

 

[ps360]

TEKKEN6 Cut-scene
http://zoome.jp/ps360/diary/450?no_bt=1

Avg. XBOX360(60.000) PS3(59.846)

 

[brain_stew]

Why is that surprising? Triple buffering (well, correctly implemented triple buffering) can only ever increase responsiveness compared to v-sync'ed double-buffering, not decrease it.

Well I know that, but there's, for whatever reason, a "propaganda machine" that seems to vehemantally argue otherwise. You wouldn't believe the hell I went through when I tried to let PC gamers at GAF know of the benefits of triple buffering and D3DOverrider, and that was even with an excellent comprehensive Anandtech article on the subject to back up my position.


Oh, and I never said those framerate graphs proved it was using triple buffering, just that the behaviour was consistent with it.

 

[ps360]

Analysis Tool Update
Add Rendering time Graph
Change Histogram Color

MODERN WARFARE 2

ACT1
Team Player
Avg.57.258 Min-Max.36.5-60.0
http://zoome.jp/ps360/diary/452/
http://zoome.jp/ps360/diary/453/

Cliffhanger
Avg.58.818 Min-Max.33.0-60.0
http://zoome.jp/ps360/diary/454/
http://zoome.jp/ps360/diary/455/

 

[jaidek]

Assassin's Creed II Analysis

PlayStation 3 Avg. FPS: 28.97 / Xbox 360 Avg. FPS: 29.76
PlayStation 3 Avg. Frame Tear: 16.36% / Xbox 360 Avg. Frame Tear: 5.06%​

Lens of Truth: Assassin's Creed II Analysis