The Framerate Analysis Thread part 2

So it goes from 39% to 2% on PS3 just by clipping out the 7 top/bottom pixels? So the vast majority of tearing is in the overscan area?
 
Yes - this is the reason that I pretty much always clip out the overscan on PS3 titles with soft v-sync - because it is in no way representative of actually playing the game.

Soft v-sync = pretty much constant tearing, but as it's basically invisible to the human eye it has no impact on the gameplay experience.
 
you guys play consoles games without overscan? i thought people typically try to preserve as many pixels as possible

also are torn frames counted towards the frame count? because i always wondered bout that especially if a frame is drawn from the top down a couple lines updated isnt really worth much.
 
you guys play consoles games without overscan? i thought people typically try to preserve as many pixels as possible

also are torn frames counted towards the frame count? because i always wondered bout that especially if a frame is drawn from the top down a couple lines updated isnt really worth much.


I certainly don't have any overscan on my HDTV, which is why I've always been a little uncomfortable discounting them.

Has an analysis been done of gameplay with the overscan area discounted? There was certainly a lot more than 0.2% tearing in the 360 demo during actual gameplay and it wasn't in the overscan area either.
 
you guys play consoles games without overscan? i thought people typically try to preserve as many pixels as possible.
I found only a couple of days ago that my 32" Sammy was in 16:9 mode voer HDMI and had overscan. Switched it to pixel mapping to recover the lost pixels. But that doesn't really affect the perceptability of soft-vsync tearing, because your eyes are rarely looking at that part of the screen.
 
Yeah, I have a 55" as well and sit about 10' away. I doubt I'd notice tearing if it was 7 pixels from the bezel unless I was looking for it.
 
The thing is that while position of the tear is of course crucial, of equal importance is the content of the framebuffer. If there's a tear in the top third of the screen, but the top third of the screen is basically a skybox, it's virtually unnoticeable.

Similarly if you're moving "into" the screen (for example driving straight ahead), a tear is far less noticeable than if you're turning sharp left/right, or if an explosion occurs.

I'd love to be able to put an "index" on the impact of screen tear but it's just not possible.

I am however a fan of clipping out the overscan area from analysis. The soft v-sync phenomenon occurs in a pretty big percentage of PS3 titles, making the torn frame percentage stat useless.

I shy away from average frame rates stats these days because context really is king. We can have game X at 29.5fps and its other console equivalent at 29.2fps and we can still see huge differences at crucial points: the longer your clip the more averages average out.

The problem I see with adjusting tear% for overscan means you would also want to adjust frame rate for overscan. Don't most games render top to bottom? So if the tear consistently appears at the top in the overscan area, that means you also won't see even a portion of the frame that was in the process of being rendered.

So while the apparent tearing in game will be less, the apparent frame rate is also going to be less.

That's why I dislike when people start to adjust numbers for this or that. Although the way PS360 did above for Tearing (adjusted tearing) is fine. But then I'd argue at that point you should perhaps have Frame Rate (adjusted framerate)

I still prefer when overall numbers are presented along with a video of what's being analysed and then people can make their own subjective judgements.

Regards,
SB
 
Soft v-sync almost always happens on 30fps titles (WipEout HD is the only example I can think of in a 60fps title) and making the overscan adjustment will not change the frame rate measurement there as in 2VBL games the tearing doesn't impact the frame-rate calculation.
 
So if the tear consistently appears at the top in the overscan area, that means you also won't see even a portion of the frame that was in the process of being rendered.

So while the apparent tearing in game will be less, the apparent frame rate is also going to be less.
I'm not sure I understand. Framerate is how many complete frames a game renders and sends to the screen, and not how many complete frame pixels are displayed. A 30 fps game not v-sync'd will be displaying 30 updates a second, only with a frame being rendered across multiple screen refreshes. You do see the whole frame, just the top half appears on the following frame.
 
I'm not sure I understand. Framerate is how many complete frames a game renders and sends to the screen, and not how many complete frame pixels are displayed. A 30 fps game not v-sync'd will be displaying 30 updates a second, only with a frame being rendered across multiple screen refreshes. You do see the whole frame, just the top half appears on the following frame.

In cases such as that where you maintain exactly 30 fps (unlikely without vsync or a FPS cap) there will be no tear as the frame should be displayed for 2 frames before the next frame is drawn.

With variable framerate is when you'll get a situation where frames will be displayed at different refreshes. Say 36 FPS. Hmm, yes, now I see what grandmaster is saying.

The situation I describe exists when your game is running higher than 60 FPS on a common LCD (where the screen refreshes 60 times a second) at which point you'll have frames overwriting other frames before they can complete. It's due to my coming from PC where it's quite common to have your game running at higher than 60 FPS. :p

Isn't the overscan area visible if you have your TV set to 1:1 pixel mapping?

Regards,
SB
 
The situation I describe exists when your game is running higher than 60 FPS on a common LCD (where the screen refreshes 60 times a second) at which point you'll have frames overwriting other frames before they can complete. It's due to my coming from PC where it's quite common to have your game running at higher than 60 FPS. :p
Yes, but that's a different phenomenum. It's quite possible for your game to render 200 frames in a second, but only display 60 refreshes with each frame made out of pieces of the rendered frames. In the case where you introduce tearing to accomodate frames not completed by the start of the next redraw, you get to see all of each frame, just chopped up over multiple refreshes. And that doesn't mean to say framerate is lower than the refresh rate - only it isn't synchronised. Instead, the backbuffer and frontbuffer are swapped after the refresh has started, with the top few lines already having been sent to the display.

eg. A 60 Hz display will have refresh timings in ms from start of...
Code:
0     16.7     33.3     50     66.7     83.3     100     116.7     133.3     150     166.7     183.3      200
A 30 fps vsync'd game will output the new frame when the next refresh starts at timings...
Code:
0              33.3            66.7              100               133.3             166.7                200
A non-vsync'd but still 30fps game with a 1ms offset will update at these times...
Code:
1              34.3            67.7              101               134.3             167.7                201
This'll mean the image tears 1/16th of the way down the screen for the first refresh, then the second refresh will draw the whole frame, then the next refresh is a new frame that starts drawing 1/16th down the screen again.

This is what PS3's soft vsync is, a change of buffers that's not synchronised to exactly the next refresh, but which occurs a tiny fraction of a second later. I don't really understand the whys and wherefores of this though!

Isn't the overscan area visible if you have your TV set to 1:1 pixel mapping?
Yes (or Full Screen mode for 768 panels), but it's mostly imperceptible.
 
XBOX360 Avg:29.108fps Min-Max:19.5-30.0fps Tear:13.611%(13.378%)
http://www.youtube.com/watch?v=ZqOHK4-B1aE

PS3 Avg:28.614fps Min-Max:16.5-30.5fps Tear:0.000%
http://www.youtube.com/watch?v=pgLosqRHOC8

QTE COMP
360 Avg:29.367fps Min-Max:27.0-30.0fps Tear:12.667%
PS3 Avg:28.933fps Min-Max:25.0-30.0fps Tear:0.000%
http://www.youtube.com/watch?v=K1G6GKOpyeQ

ya pretty much predicted this, noticed the slowdown on the ps3 version while fighting the boss. 360 version tearing is generally unnoticeable because of the low contrast of the scenes.

but judging from the framerate looks like computation cycles are practically maxing out on both consoles meaning neither console was really held back for parity both seem to need the reduction in resolution.
 
Thanks a lot ps360.

I wonder if the tearing is in the middle of the screen or at the border of the screen, because during playing the 360 demo, I cannot remember tearing...or probably the sometimes bright background helps to hide it?
 
Thanks a lot ps360.

I wonder if the tearing is in the middle of the screen or at the border of the screen, because during playing the 360 demo, I cannot remember tearing...or probably the sometimes bright background helps to hide it?


Me either. Which is why I was puzzled at PS360's vid. But I guess it's a contrast thing as others have noted.
 
Back
Top