GT5's Game-Breaking Online Flaw (OP Updated: 11 Feb)

[dudeondacouch]

Must agree with dsgerbc.

If the seconds on the replay clock are ticking more quickly, it will arrive at a higher total time than an external stopwatch in the same given period.

 

[EDK]

During one minute of real time, the replay clock does more than one minute. I think this means that replay clock is "faster".

Well either way, it means the lap is slower, which is what I think we're trying to arrive at.

Understand your perspective, but the terminology "faster" would be confusing, in the least.

 

[Mac K]

Great article, great job done by all involved, hope it will bring additional attention and hopes it can actually be solved. Thank you guys for the effort in the mutual behalf 👍

I'll second this. Hopefully the upcoming update fixes this.

 

[dsgerbc]

Well either way, it means the lap is slower, which is what I think we're trying to arrive at.

Understand your perspective, but the terminology "faster" would be confusing, in the least.

It's sure confusing.
Question: second video in the post "Racing with 13 drivers using whatever peripherals ", is a live recording, not a recording of a replay playback in gt5? Cause clock seems in-sync with GT5's. I tried to see roughly how often there are "missing frames" in GT5 replay and how many are dropped at a time, only to find that this replay seems to be of a real-time race recording. Gonna try looking at Filip_Ovik's replay.

 

[EDK]

It's sure confusing.
Question: second video in the post "Racing with 13 drivers using whatever peripherals ", is a live recording, not a recording of a replay playback in gt5? Cause clock seems in-sync with GT5's.

All of the vids are live, as far as I know.

Speedy captured them live, and the replay clock differs from the live clock.

He did not make vids of the replays, but he did view them to verify that fact.

 

[Hydro]

It's sure confusing.
Question: second video in the post "Racing with 13 drivers using whatever peripherals ", is a live recording, not a recording of a replay playback in gt5? Cause clock seems in-sync with GT5's. I tried to see roughly how often there are "missing frames" in GT5 replay and how many are dropped at a time, only to find that this replay seems to be of a real-time race recording. Gonna try looking at Filip_Ovik's replay.

You are right.
Those videos were taken from live recordings.
We compared the live recordings to the replays in order to get a better idea of the time inconsistencies.
We also compared the two versions to check for differences in lag (ie: was there more in the replay, or vice-versa).

Edit: Much like EDK just said 👍

 

[dudeondacouch]

All of the vids are live... He did not make vids of the replays...

Far be it for me to question the validity of these tests (which, I must add, is assuredly genuine given the commitment and effort shown by the staff and participants), but doesn't eliminating the control group remove the ability of objective outsiders to accurately assess your findings?

 

[EDK]

Far be it for me to question the validity of these tests (which, I must add, is assuredly genuine given the commitment and effort shown by the staff and participants), but doesn't eliminating the control group remove the ability of objective outsiders to accurately assess your findings?

Not sure what you mean?

My post suggests that the slow clock/fast clock (whatever you want to call it) only presents during replays.

That said, the important part is that the achievable lap time is impacted in a significant way, where laps under the right conditions are 3 seconds slower than those in free run and while under other unaffected conditions.

 

[dudeondacouch]

I agree completely that replay times are irrelevant when it comes to the actual experience of racing, but with regard to this issue, and specifically the evidence showing proof of its existence, the replay data is very important and should be provided.

 

[dsgerbc]

I'm bored so did smth probably pretty useless. Compared gt5 clock deltas from the first three seconds of Filip_Ovik's video to video clock deltas.

About 80 frames. ~33 msec per frame. Vertical scale - time deltas, horizontal scale - frame #. Total discrepancy between video clock and gt5 clock by the end of three (GT5) seconds - .3seconds.

Not sure it's useful. Deltas seem a lot more choppier than I expected. But it might be interesting to repeat this with two replays (glitched/good) converted into video on the same hardware.

 

[dudeondacouch]

I don't think you're going to get any good data from replays with so small a sample.

Even watching a replay on your own console, you have to account for whatever compression comes into play when it is stored.

With that sample, you can see that the troughs are absolutely consistent, which I would submit is due to file compression. It seems that the other points do appear to vary toward the high side when comparing the final couple seconds in relation to the first, but a much larger sample would be necessary to garner anything that could be considered usable.

 

[dsgerbc]

^Larger sample needs image/number recognition software. Someone might have it, but I don't.

I agree. I was hoping to see some smooth deltas which are either consistently off the real-time ones or are generally in line, but waay off once in a while.

 

[bullet360]

I have always bin left wondering if even the online free run and the actual race physics differed. And apperently they do since they say free run is un-affected. Does this mean that we're gunna have to start tuning our vehicles to the "in race" physics as apposed to free run physics? Seeing l love tuning when it comes to suspension setups and LSD settings...

 

[CSLACR]

I'm bored so did smth probably pretty useless. Compared gt5 clock deltas from the first three seconds of Filip_Ovik's video to video clock deltas.

About 80 frames. ~33 msec per frame. Vertical scale - time deltas, horizontal scale - frame #. Total discrepancy between video clock and gt5 clock by the end of three (GT5) seconds - .3seconds.

Not sure it's useful. Deltas seem a lot more choppier than I expected. But it might be interesting to repeat this with two replays (glitched/good) converted into video on the same hardware.

So the clock is speeding up and then slowing back down that much, as shown on the high spikes, yes?

 

[dudeondacouch]

I have always bin left wondering if even the online free run and the actual race physics differed. And apperently they do since they say free run is un-affected. Does this mean that we're gunna have to start tuning our vehicles to the "in race" physics as apposed to free run physics? Seeing l love tuning when it comes to suspension setups and LSD settings...

This post makes me sad. Please learn English.

 

[CSLACR]

This post makes me sad. Please learn English.

I've seen a lot worse, the only problem here is "bin"

 

[AllinWitJT]

Bandwidth meter reading a 16 car race.

Mics. on

Mics. on

Mics. off pre-race qualify

Mics. off race

As you can see mics. on causes spikes in bandwitdh consistent with eratic patterns of lag.

The green is inbound data (downloading), and the red is outbound (uploading). The graph doesn't differentiate between cars and mics, it just logs total bandwidth in and out. When running with no mics, there is very little traffic outbound (less than 10 Kbs and around 25-30 Kbs inbound).

 

[dudeondacouch]

I've seen a lot worse, the only problem here is "bin"

Is "apposed" a word with which you're familiar?

 

[dsgerbc]

So the clock is speeding up and then slowing back down that much, as shown on the high spikes, yes?

Zeros you see are probably an artifact of video encoding. Basically a zero exactly every six frames compensates for two frames @50ms (33,33,33,50,50,0) to have all six average to 33ms. Positive spikes in delta above 50ms is where two clocks drift apart. Basically, if all deltas in the high-60s were 50ms, then clocks would average to the same.

 

[CSLACR]

So it's essentially possibly a graph of the bug's work in progress?

 

[griffinz]

Bandwidth meter reading a 16 car race.

Mics. on

Mics. on

Mics. off pre-race qualify

Mics. off race

As you can see mics. on causes spikes in bandwitdh consistent with eratic patterns of lag.

The green is inbound data (downloading), and the red is outbound (uploading). The graph doesn't differentiate between cars and mics, it just logs total bandwidth in and out. When running with no mics, there is very little traffic outbound (less than 10 Kbs and around 25-30 Kbs inbound).

Something is wrong here with your sampling. I have done a fair share of bandwidth monitoring of this issue, first your numbers are way to low, and second it doesn't pulsate like that.

This is a 7 car race on grand valley speedway for example, mics on.

 

[AllinWitJT]

Something is wrong here with your sampling. I have done a fair share of bandwidth monitoring of this issue, first your numbers are way to low, and second it doesn't pulsate like that.

This is a 7 car race on grand valley speedway for example, mics on.

Why is it wrong just because it's not like yours? 16 car race an hour long is my sample.

 

[griffinz]

Why is it wrong just because it's not like yours? 16 car race an hour long is my sample.

Its wrong because its wrong.

If its a P2P race U/D numbers should be nearly equal the entire race. If its a fixed host it would use a tremendous amount more bandwidth, my 14 player fixed host sample sustained over 2000kbit/s with a peak to 3450kbit/s on the upload.

 

[adellanotte]

I think it is also remarkable the raising amount of blackscreens and game freezes, wich to me is part of the same problem, as well as compatibility issues between historic-compatible gamers. I spend lots of time adjusting the cars and setting my best laptimes for further online racing, but when going online, all the previous work on car adjustment becomes totally useless. Going online, mates has become a nightmare. Hope they fix that soon. And just a reflexion: if the problem is ocurring so long ago, why people al PD don't focus on solving it rather that being releasing stupid (fo me) DLCs each month. And last, but not least, why PD has not make any statement regarding the problem.

 

[Speedy6543]

I think it is also remarkable the raising amount of blackscreens and game freezes, wich to me is part of the same problem, as well as compatibility issues between historic-compatible gamers.

Clearing your cache in GT5 should solve this problem.
I had the same thing going on right after the last patch. Cleared my cache and instantly it got much better.
GT life -> Options -> Online Options -> Clear temp. files

 

[stanley1664]

Sunday nights used to be filled with fun racing but due to the constant disconnect ive stopped playing online now, Haven't played online since October of last year.. Decided over the weekend to fire GT5 back up & purchase the DLC pack.. Big mistake nothing but hassle constantly being kicked from servers after about 45 seconds really big head ache coming on.. Last night decided to try and some real racing in with some friends..

14 in the lobby with the same recommended car the Tvr Tamora, circuit was high speed ring reverse.. started the race i think in 12th position.. after 2 laps noticed somthing rather odd.. My car just seemed overly quick like 2 seconds a lap quicker which against these guys is not rite.. However people starting moaning about this bug which I have never heard of before.. The cars I over took during the race seemed laggy and sluggish.. checked replay it was 15m 58 seconds but from my stop watch it was 15m 38 seconds a whopping diefference of 20 seconds.. so basically Im here to confirm this awful situation as some people haven't seen this yet or are living in denial. pls fix this PD,, game is null & void now..

 

[adellanotte]

Clearing your cache in GT5 should solve this problem.
I had the same thing going on right after the last patch. Cleared my cache and instantly it got much better.
GT life -> Options -> Online Options -> Clear temp. files

Txs Speedy
Believe when I say I'm cleaning cache all the time, but the problem persists.
Most of times i play just w one friend on his or mine private lounge, wich btw, is much more stable than online rooms. Lately we both have to restart both the Ps3 and the router at last twice a session.

Nightmare!
Cheers.

 

[windmill]

Clearing your cache in GT5 should solve this problem.
I had the same thing going on right after the last patch. Cleared my cache and instantly it got much better.
GT life -> Options -> Online Options -> Clear temp. files

Adellanotte and me race a lot togheter and we always before we go online empty our cache .

I am sure that its a part the ps3 and if you use usb or steering wheel like explained in https://www.gtplanet.net/complex-gt5-bug-causes-problems-for-online-racers/

However i am also very sure that its a lot to do also with the amount of servers Sony and PD have running for GT5 .

On certain days and certain times (weekends for example) the black screens and time and grip bug is LOTS worse , heavy traffic of PS3 users has lots to do with it !!

 

[jon-]

Txs Speedy
Believe when I say I'm cleaning cache all the time, but the problem persists.
Most of times i play just w one friend on his or mine private lounge, wich btw, is much more stable than online rooms. Lately we both have to restart both the Ps3 and the router at last twice a session.

Nightmare!
Cheers.

FWIW, we have huge problems with disconnects, black screens, and people not seeing each other in our championship.

Of 14-16 of us, there's always at least 1 person who can't see more than 10 people and a couple of people who just never see each other.

Likewise the game seems to make 2 people arch rivals so when one joins the room the other gets kicked.

We all clear our cache constantly.

It may or may not be related to this thread, but it's definitely not just you

 

[Famine]

Likewise the game seems to make 2 people arch rivals so when one joins the room the other gets kicked.

We used to get that in MW2, oddly. One guy always got dropped going from party to lobby, unless he hosted and then it was the other guy.