The Framerate Analysis Thread part 2

[grandmaster]

Probably a stupid question, but why isn't their an analysis of the MP portion of games is it because its hard to tell the difference between latency and frames dropping?

There are too many unknown variables: quality of connection to PSN/XBL both server-side and client-side. What would be of interest to me would performance tests of both networks, but I have no idea how they would be accomplished.

 

[SG79]

grandmaster, I noticed some tearing on the upper side of the screen on the riverside and other similar parts in the trains mission. Basically whenever there is a body of water present.

A consistent tear on the top to be precise, and it's on the PS3. I'll get back to you on the exact location.

EDIT: A similar effect on the side at points and covers a larger area. It's not tearing but not sure what to make of it:

http://the-horror.com/forums/showpost.php?p=85828&postcount=219

 

[grandmaster]

Yes I do recall seeing that effect. It appears to be a bug. I would imagine that it's safely tucked into the overscan area so not a massive deal, but it's definitely 'there'.

 

[the ignoramus]

Has anyone analyzed Project Gotham Racing 4?
I believe it's one of the few "30 fps locked" games that really does not go below 30. Or at least 99% of the time it doesnt.
(Over a year and a half old and it's still one the top 5 best looking console games)

 

[Mr Deap]

Correct me if I'm wrong, but isn't the additional input lag at most less than a frame?

It's a lot more than a single frame.

There's no actual tool to mesure it, but If you can tell the difference, it's obviously more than a single frame.

 

[SG79]

With RE5 it's not noticeable, but there are games that the input lag hit you like a hammer. NGII fit the profile, add the slight lag associated with your average flat panel and the game was a nightmare at times.

 

[TheWretched]

Having now finished RE5 on PS3, I must say, the "worse framerate" is overblown, in my opinion. Only sometimes when too much is happening, it gets a bit lower, but the rest is steady as she goes. The worst part is the final boss... Then it REALLY starts dropping frames... I'd like to see, what framerates the PS3 gets there, compared to the 360 (not in video, stats are enough )

The switching MSAA/Quincunx actually isn't really noticable to me. If you look REALLY closely, you notice it, but in-game, you will never see it.

(I don't want to start a flame war or anything, these are just "my observations", nothing else)

 

[Chisholm]

Has anyone analyzed Project Gotham Racing 4?
I believe it's one of the few "30 fps locked" games that really does not go below 30. Or at least 99% of the time it doesnt.
(Over a year and a half old and it's still one the top 5 best looking console games)

I'm afraid it's not that rare to see the game go below 30fps, especially when there are rain effects going on; it's not too bad, but it's quite noticeable. PGR3's framerate is really smooth, though. Having said that, I completely agree that PGR4 is one of the best looking games this generation. It's a shame that it was so overlooked when it was far better looking -cockpit view aside- and playing than PGR3.

 

[Arwin]

I don't know ... the game looked pretty good, but not that great. What bummed me out a bit was that in order to get 'full' 720p (worth it!), they had to drop my favorite effects from PGR3 (HDR contrast and lighting). The rain conversely really didn't do it for me. It's like I had to trade the sun for rain ... that's never good! (ok unless the crops are at risk)

 

[Chisholm]

I don't know ... the game looked pretty good, but not that great. What bummed me out a bit was that in order to get 'full' 720p (worth it!), they had to drop my favorite effects from PGR3 (HDR contrast and lighting). The rain conversely really didn't do it for me. It's like I had to trade the sun for rain ... that's never good! (ok unless the crops are at risk)

Sure, I missed those effects too, but I felt that sometimes the HDR was a bit overdone. It was great to help immersion when in cockpit view but at times it was really distracting, especially when you were coming out of a tunnel.

For me, the resolution bump, the weather effects, the bikes, the new single player modes and the tighter controls accounted for a far more solid game. Having said that, it's obvious that it could have been even better with PGR3's HDR and lighting. I'm looking forward to Bizarre's new racing game to see if they can optimize their engine so it may work with HDR. They might already be hitting the limits of the eDRAM, though, given that they're already doing some postprocessing in the form of motion blur and 2x AA.

 

[Arwin]

Sure, I missed those effects too, but I felt that sometimes the HDR was a bit overdone. It was great to help immersion when in cockpit view but at times it was really distracting, especially when you were coming out of a tunnel.

There was a much more subtle implementation of HDR in PGR4, which was pretty good - but it was only active when you went to photo mode.

For me, the resolution bump, the weather effects, the bikes, the new single player modes and the tighter controls accounted for a far more solid game. Having said that, it's obvious that it could have been even better with PGR3's HDR and lighting. I'm looking forward to Bizarre's new racing game to see if they can optimize their engine so it may work with HDR. They might already be hitting the limits of the eDRAM, though, given that they're already doing some postprocessing in the form of motion blur and 2x AA.

It was definitely a more solid game, but for me actually the bikes helped kill it - I hate bikes! I don't know why but I can't help it. It's a bit like zombies.

 

[catisfit]

I think what is worthwhile to point out is that these features *are* subjective but they do give you the data to make your own determinations. If you want to disagree with what I think, that's absolutely fine. But the bottom line is that EG asks me to include recommendations for those who own both consoles and that's what I do. In the case of RE5 I didn't find this particularly difficult.

That's exactly my point-there is no data!

Unless I missed it somehow, I saw no torn frame percentages or average framerates written in either article. In the past these figures have usually been included.

 

[Dural]

I love PGR4 and also find it to still be fantastic looking a year and a half later, but the bug that removed the dynamic shadows on the cars in one of the updates that still hasn't been fixed really bothers me every time I play it.

 

[grandmaster]

That's exactly my point-there is no data!

Unless I missed it somehow, I saw no torn frame percentages or average framerates written in either article. In the past these figures have usually been included.

The video is more indicative of general performance and is certainly more useful than an average frame rate - it's the whole reason I visualised the data in the first place. For the record the screen tear percentage on Xbox 360 in the video in the article is 1.5%, though for the demo capture mentioned earlier in this thread I think it was something like 5%. The longer the sample, the lower the percentage becomes which is again why I prefer the videos for giving a better idea of performance.

 

[user542745831]

Gran Turismo 5 Prologue Video Performance Analysis

http://www.digitalfoundry.org/blog/?p=678

 

[grandmaster]

So a drop to around 50fps at worst on GT5. The impact would probably be far more intense in the online section where the coders can't control the amount of cars on-screen at any given point. The screen tear kills the IQ in those sequences but I'm guessing that Polyphony has opted to retain smoother controls over v-lock.

Yeah I'm quite keen on doing Killzone-style engine analysis on Halo 3 and MGS4, but if any
one has any suggestions for 'quick and dirty' tests I can do like user542745831's suggestion of GT5p, I'm all ears.

 

[SG79]

Ninja Gaiden II, patched of course. Though I suppose it'd be better to skip on that one until the PS3 version is out to make a proper head to head comparison.

 

[BadTB25]

Ninja Gaiden II, patched of course. Though I suppose it'd be better to skip on that one until the PS3 version is out to make a proper head to head comparison.

I don't think it would be a proper head-to-head since it was not developed at the same time by the same producer/director/team.

At the same time, this type of game is my favorite genre, so I may be inclined to pick it up if the improvements are robust enough. I don't really care about the co-op, but the additonal characters and hopefully missions may entice me. Hopefully it will be released for around $40 or so.

 

[SG79]

May be true, and though off topic, a lot of belated ports/versions aren't handled by the original teams yet they're compared (example: BioShock). In this case, some original team members are present.

 

[grandmaster]

The belated head-to-heads are worthy of even more scrutiny IMHO, especially if the original developer is not involved.