Hello, i would like to know if there's any way to play a replay backwards.

I know that we can pause it and play it forward in a multitude of different speeds, which is pretty cool and useful but i really need a backwards function.
Please tell me that i overlooked which keys is used to do it cause otherwise i'm really screwed* heh.



* I'm reviewing a lot of long replays that requires my attention on several cars on the same "minute", so, yeah, i've been restarting a lot the same damn replays till i get to see what i need. A backward function would be pure sex, to me. Please tell me that it exists..

Im sorry to tell that there isn't such feature.
It would be very usefull, but if I recall correctly it has something to do with physics and how they can't be calculated backwards, or something like that. But it shouldn't be impossible to make I guess

lmao love your choice of words ". A backward function would be pure sex, to me. Please tell me that it exists.."I supose if you render the replay's then your media player can play rhem backwards

Quote from Blackout :

Im sorry to tell that there isn't such feature.
It would be very usefull, but if I recall correctly it has something to do with physics and how they can't be calculated backwards, or something like that. But it shouldn't be impossible to make I guess

Oh Blackout, you!!
Even if the physics are little less accurate there can't be a backward function? Oh man, i'm in a world of pain!
Thanks for breaking in the bad news tho.

Quote from CSU1 :

lmao love your choice of words ". A backward function would be pure sex, to me. Please tell me that it exists.."I supose if you render the replay's then your medis player can play rhem backwards

Glad you liked it, by the way i'm brazilian and my english sucks (i'm sorry about that) but i guess that humour is an universal thing, eh?

I thought already about what you suggested here before, the problem is that... when i render a movie it will only "save" the camera at which i'm looking at the time it's being made (the movie).

So... it won't help me even a tad bit cause i need to fire up my delorean (i.e go back in time) to see other cars (hence, other cameras) at the "same time".

Well, if any of it made any sense to you.

It's really obvious and stuff but i can't word it any better (at least not now, but hey i just pulled an all nighter here, it's 08:37 AM).

Cheers from Brazil.

You'll probably never be able to truly "rewind" a replay, simply because a replay doesn't contain raw position data, but just controller input. If you watch a replay, it's like a bunch of AI cars that do the same inputs you did, resulting, as long as the physics engine has not changed, in the same actions you did originally. This is ofcourse a bit different in multiplayer replays, but the same principle applies (afaik).

The only thing that *could* happen sometime is storing physics keyframes, so while you cannot fluently rewind, you can atleast jump back to a certain point in the replay, like to the beginning of each lap, or each checkpoint, for example.

Android, so you are telling that a replay is just a banch of controller inputs data?

It means that there is no intermediate position of all the cars, their velocities and directions stored in any point that would possibility to start a replay from the midle?

But as you suggested, that does not look so difficult to store intermediate position data, things like one save each split time place.

-----

Just talking, considering that each time we play the replay we have a new fresh game happening, the replay could happen in a different way each time due some rounding in the calculation, cause everything depends on the first start of the race?

As an example, if we have an accident with many cars, each one reacting with others in many directions and performing many damages, each time I play the replay the calculations and all that interation will be the same or could be inserted some errors that would make the replay be different?

how? it takes the inputs and makes a calculation, the answer can't change, its like this

First replay run:
1+1=2
2nd replay run:
1+1=2
3rd replay run:
1+1=1.5

Quote from Speed Soro :

As an example, if we have an accident with many cars, each one reacting with others in many directions and performing many damages, each time I play the replay the calculations and all that interation will be the same or could be inserted some errors that would make the replay be different?

If you'd like to take a replay, and make two videos where the action in them is not absolutely identical, I'd be interested to see that.

SPRs and MPRs are made differently. SPRs do contain only controller input data (apart from the obvious car settings/track/laps info). MPRs are different because they do rely on packets of positional data and controller inputs, and interpolate between those packets using physics calculations to project the flow of motion/events.

It would be good to have this feature.

btw. Menie why are you always have this sad smile: :o

LMAO NIKI! I saw what you typed before you fixed it! Was that a spell checker that did that to you? hehehe!

Quote from SamH :

If you'd like to take a replay, and make two videos where the action in them is not absolutely identical, I'd be interested to see that.

SPRs and MPRs are made differently. SPRs do contain only controller input data (apart from the obvious car settings/track/laps info). MPRs are different because they do rely on packets of positional data and controller inputs, and interpolate between those packets using physics calculations to project the flow of motion/events.

I could do that if you count "lag"

It's my pet annoyance about LFS replays. Prime example is the new start thing where the car in front will not lag, great, so the car in front of me when I'm playing doesn't lag anywhere near as much as the others. In the replay I'll watch the start from a driver's view 1 car back (so looking at the driver one spot in front of me, both cars to the right for example), now the car that lagged quite a lot in "real time" doesn't lag anywhere near as much, but it will if I watch the start from my own car again. Why?

You also see your own car lag a bit in replays with a lot of cars, obviously LFS recorded its position perfectly, so why?

And finally, you often see cars in the distance do that "snap, snap, snap" lag around corners (particularly if there are several other cars close around you), and yet, ride onboard in a replay and they're rock solid, therefore LFS has (one way or another) recorded exactly where they are, so why the snap snap?


Anyhoo, I know it's all about physics priorities and simple physics for less important cars it's just annoying (to me)....so I guess that's what I would like changed/fixed before a rewind button.

Quote from Jakg :

how? it takes the inputs and makes a calculation, the answer can't change, its like this

First replay run:
1+1=2
2nd replay run:
1+1=2
3rd replay run:
1+1=1.5

Not exactly, I said like this:

First replay run:
1.00000001 + 1.00000001 = 2.00000002
2nd replay run:
2.00000002 + 1.00000001 = 3.00000003
n replay run:
n.999999999 + 1.00000001 = n+2 (where should be n+1)

I don't know if it could happen, it just curiosity.

Quote from SamH :

If you'd like to take a replay, and make two videos where the action in them is not absolutely identical, I'd be interested to see that.

SPRs and MPRs are made differently. SPRs do contain only controller input data (apart from the obvious car settings/track/laps info). MPRs are different because they do rely on packets of positional data and controller inputs, and interpolate between those packets using physics calculations to project the flow of motion/events.

Sure I won't do this, but probable it happens

If it confirms that the result will be same in 1000 times I play the replay, I would be induced to believe that the physics is something made in gesso, I guess maybe it is missing something natural, some aleatory dependence of the chaos, what would guide me to beleive that is possible to a very trained simdriver to reach the perfection once he had found all the exactly points and times to make the right inputs, what is impossible to reach in the real driving due the unpredictable nature of the things...what a trip... forget

If I understood it correctly, at least .sprs are only controller inputs, as Android said... I'm not sure about .mprs though position corrections are only possible if there is a position stored at some point (warping, lag aso.)...

Quote from Speed Soro :

Not exactly, I said like this:

First replay run:
1.00000001 + 1.00000001 = 2.00000002
2nd replay run:
2.00000002 + 1.00000001 = 3.00000003
n replay run:
n.999999999 + 1.00000001 = n+2 (where should be n+1)

I don't know if it could happen, it just curiosity.

The values wouldn't change between runs. All x86 CPUs do math the same way (consistently), the engine inputs are exactly the same (both in timing and values), the engine runs at the same frequency -> results will be the same. There are no random values in the physics calculation.

what i understand, is that basically LFS's replay is replacing what you do with a fake driver. therefore, LFS replays still load environments, etc. what the replay data stores is what you do, and then lfs just LITERALLY replays the outcome.

Quote from wien :

The values wouldn't change between runs. All x86 CPUs do math the same way (consistently), the engine inputs are exactly the same (both in timing and values), the engine runs at the same frequency -> results will be the same. There are no random values in the physics calculation.

Thank you! I couldn't be bothered to think hard enough to type that (I have an awful cold), but I totally agree. The result will always be the same. If anyone wishes to contend this (Speed Soro, perhaps?) with alien intervention perhaps, or some twist in the space/time continuum, they should provide evidence. Otherwise I'm confident that the logic is the logic it is. Logically speaking, of course.

Quote from sinbad :

It's my pet annoyance about LFS replays. Prime example is the new start thing where the car in front will not lag, great, so the car in front of me when I'm playing doesn't lag anywhere near as much as the others. In the replay I'll watch the start from a driver's view 1 car back (so looking at the driver one spot in front of me, both cars to the right for example), now the car that lagged quite a lot in "real time" doesn't lag anywhere near as much, but it will if I watch the start from my own car again. Why?

CPU time management. The cars nearer to you are more important. The physics applied to them are running at a higher resolution than cars further away. The same data from the replay is being used, but with particular/granular attention where it matters. To calculate ALL cars - even the ones out of sight, at the other side of the track - would be either a heavy burden, or an impossible burden for your processor. With cars that are not within your visual periphery or immediate proximity, almost no physics calculations, or the bare minimum physics calculations, are being applied. This makes closer racing much more possible in LFS than if all cars were having the same level of physical calculations applied to them.

Quote from sinbad :

You also see your own car lag a bit in replays with a lot of cars, obviously LFS recorded its position perfectly, so why?

Actually, in an MPR, your own car's positional information and inputs are not stored in any more detail than anyone else's. That's why, when you watch any car in a replay - including your own - you will see the same jerkiness in the steeringwheel positions that you do with anyone else's. The amount stored is dependent on the server you're on. The STCC servers all run 6 packets per second resolution. Some others run 4 packets per second, which does result in a slightly poorer/more jerky online experience and replay. Less packets = more margin for error.

Quote from sinbad :

And finally, you often see cars in the distance do that "snap, snap, snap" lag around corners (particularly if there are several other cars close around you), and yet, ride onboard in a replay and they're rock solid, therefore LFS has (one way or another) recorded exactly where they are, so why the snap snap?

Same as mentioned above.

Quote from sinbad :

Anyhoo, I know it's all about physics priorities and simple physics for less important cars it's just annoying (to me)....so I guess that's what I would like changed/fixed before a rewind button.

I wish I'd read this bit before typing the other stuff above. Well, hopefully someone will appreciate the dissertation

Quote from SamH :

LMAO NIKI! I saw what you typed before you fixed it! Was that a spell checker that did that to you? hehehe!

What is LMAO? And it was writing mistake

LMAO = Laughing My ASCII Off