The Framerate Analysis Thread part 2

[ps360]

WET DEMO

360 http://zoome.jp/ps360/diary/362/
PS3 http://zoome.jp/ps360/diary/352/

360 AVG(29.965) Tearing( 0.806%) Lowest(27.0)
PS3 AVG(29.319) Tearing(48.411%) Lowest(22.0)

 

[nightshade]

Wow thats a big difference, Almost half the frames are torn on PS3 while on 360 its almost nil.
We've started to see Multiplat games which either dont tear at all on the PS3 due to proper V-sync or tear like hell when its a case of v-sync off.

 

[Arwin]

I'm thinking this is actually a new skin to an old game, with no real engine enhancements though, but maybe I'm wrong. Isn't this basically Stranglehold? ps360, maybe you can test that game and see if you get similar results.

 

[I.S.T.]

According to Gamefaqs, no way. Both games were made by completely different people. Wet is by Artificial Mind and Movement, and Stranglehold was developed by Midway itself.

 

[Arwin]

According to Gamefaqs, no way. Both games were made by completely different people. Wet is by Artificial Mind and Movement, and Stranglehold was developed by Midway itself.

Does that automatically mean they are using a completely different engine?

 

[TheWretched]

Stranglehold used UE3, right?

 

[Dr. Nick]

Stranglehold used UE3, right?

Yes it did.

 

[ps360]

Wolfenstein - XBOX360
http://zoome.jp/ps360/diary/363/

 

[assurdum]

Wow thats a big difference, Almost half the frames are torn on PS3 while on 360 its almost nil.
We've started to see Multiplat games which either dont tear at all on the PS3 due to proper V-sync or tear like hell when its a case of v-sync off.

I have tried wet on the ps3, I haven't even noticed the tearing and I'm very sensible with that.

 

[AlNom]

Let's try to keep things technical... resorting to one's own visual acuity to negate the actual analysis is not for this thread...

 

[DJ12]

I have tried wet on the ps3, I haven't even noticed the tearing and I'm very sensible with that.

I played it for a while and the tearing is almost constant, but it's always at the top few lines, if your TV overscans by default it might not even be visible.

 

[grandmaster]

Yes, it is a "feature" of a few games and Unreal Engine ones in particular exhibit it mostly on PS3.

 

[assurdum]

Let's try to keep things technical... resorting to one's own visual acuity to negate the actual analysis is not for this thread...

I have only said I haven't noticed, no more. I don't try to negate

 

[assurdum]

Wolfenstein - XBOX360
http://zoome.jp/ps360/diary/363/

mmm I'm very interesting to ps3 at this point, it seems totally vsync (triple buffer) and I'm curious how much affect the fps to comparison.

 

[liolio]

Same player shoot again, how can you tell appart double buffering + v-sync and triple buffering?
You can't if the game doesn't "release" v-sync to preserve framerate.

 

[MazingerDUDE]

Same player shoot again, how can you tell appart double buffering + v-sync and triple buffering?
You can't if the game doesn't "release" v-sync to preserve framerate.

it's rather easy to detect when you understand its attributes

If your frame rate can stay between 20-30 or 30-60 with zero tearing, it means triple buffering is being used.

For example, Madden 10 is double buffering with dynamic v-sync (turns v-sync off when frame rate dips below 30) on 360, while triple buffering with v-sync on PS3, as a result, the game runs at either 60 (in game) or 30 (close ups) with slight bit of screen tearing when frame rate dips below 30 on 360, while PS3 runs 60 in game and can stay 30-40 for close ups with zero tearing even when frame rate dips below 30. DMC4 is a similar case as well.

for your reference

http://zoome.jp/ps360/diary/321/ Madden 10 PS3
http://zoome.jp/ps360/diary/319/ Madden 10 360

It's not so hard tell even when v-locked, if frame rate stays constant at 20 with zero tearing

http://zoome.jp/ps360/diary/332/ Kane & Lynch Dead Man 360

 

[rekator]

So, depend of the game is better a slow constant frame rate with zero tearing than a high and more "confortable" frame rate with tearing in top of the screen (who generally is in the overscan TV zone).

Personally, I'm more show the use of triple-buffering to tend to don't have inconsistant frame-rate when you can achieve a constant 30fps or 60 fps without a lot of tearing easily visible, in these case is better to have a constant slow frame rate, even is not the best for gaming feeling…
In a "slow" game no screen tearing is better, but in fast and reactive game high frame rate is preferable.

 

[liolio]

it's rather easy to detect when you understand its attributes

If your frame rate can stay between 20-30 or 30-60 with zero tearing, it means triple buffering is being used.

For example, Madden 10 is double buffering with dynamic v-sync (turns v-sync off when frame rate dips below 30) on 360, while triple buffering with v-sync on PS3, as a result, the game runs at either 60 (in game) or 30 (close ups) with slight bit of screen tearing when frame rate dips below 30 on 360, while PS3 runs 60 in game and can stay 30-40 for close ups with zero tearing even when frame rate dips below 30. DMC4 is a similar case as well.

for your reference

http://zoome.jp/ps360/diary/321/ Madden 10 PS3
http://zoome.jp/ps360/diary/319/ Madden 10 360

It's not so hard tell even when v-locked, if frame rate stays constant at 20 with zero tearing

http://zoome.jp/ps360/diary/332/ Kane & Lynch Dead Man 360

Thanks I actually asked earlier, I was wandering about a case (maybe hypothetical) where the render never exceed 30fps and this case it's tougher I guess. more of an assumption, no?
Because you may go lower the 30fps with double buffering without "releasing" v-sync, no?
Anyway in real case as the one you gave there can't be confusion.
Actually as I think more about it, it must be unlikely to don't find on a triple buffering game some frame where the engine is really relaxed and the frame rate goes up. So I guess the case I describe is pretty much unlikely to exist.

 

[liolio]

So, depend of the game is better a slow constant frame rate with zero tearing than a high and more "confortable" frame rate with tearing in top of the screen (who generally is in the overscan TV zone).

Personally, I'm more show the use of triple-buffering to tend to don't have inconsistant frame-rate when you can achieve a constant 30fps or 60 fps without a lot of tearing easily visible, in these case is better to have a constant slow frame rate, even is not the best for gaming feeling…
In a "slow" game no screen tearing is better, but in fast and reactive game high frame rate is preferable.

Actually from my understanding (and before Mazinger Dude response) triple buffering as no impact on prefs/frame rate. It's clearly a great technique with just some memory overhead.

 

[MazingerDUDE]

Thanks I actually asked earlier, I was wandering about a case (maybe hypothetical) where the render never exceed 30fps and this case it's tougher I guess. more of an assumption, no?
Because you may go lower the 30fps with double buffering without "releasing" v-sync, no?
Anyway in real case as the one you gave there can't be confusion.
Actually as I think more about it, it must be unlikely to don't find on a triple buffering game some frame where the engine is really relaxed and the frame rate goes up. So I guess the case I describe is pretty much unlikely to exist.

When a game is v-locked with double buffering, its frame rate will be locked at low common denominator of 60 devided by an integer (60, 30, 20, 15, 12 etc..) which means when the game can render a scene at constant 29 fps, it will be locked down to 20.

No matter how steady an engine can render, it can go below 30 at times and if you test to stress the engine you can easily tell whether it's v-locked or not.