PDA

View Full Version : Laser/drone scanning



Leper Messiah
24-04-2017, 15:20
Any chance I can get some accurate info on what tracks are laser scanned for PC2 (I know some UK tracks were laser scanned for PC). Also this Drone scanning doohicky, does that system laser scan as well? Or is it just for photos? Quite a bit of confusion in the epeen wanging crowd about this and I want correct info before I make mistakes (may have already on YT!! said all tracks were laser and drone scanned for PC2).

Roger Prynne
24-04-2017, 15:25
Yes all tracks are laser scanned for PC2 with a drone that takes pictures as well as very accurate laser scanning, which includes a lot of the surrounding area as well.

It's the best thing since sliced bread :D

Leper Messiah
24-04-2017, 15:33
Yes all tracks are laser scanned for PC2 with a drone that takes pictures as well as very accurate laser scanning, which includes a lot of the surrounding area as well.

It's the best thing since sliced bread :D

thanks for the info Rog, does that include all OLD tracks moving across from PC?

Roger Prynne
24-04-2017, 15:36
I think so but don't quote me just yet.

Leper Messiah
24-04-2017, 15:38
I think so but don't quote me just yet.

that would be stellar, I hope it's true!!

Roger Prynne
24-04-2017, 15:45
Check this out in the mean time...


http://www.youtube.com/watch?v=1wcH-ABa3xY

Mad Al
24-04-2017, 15:48
The drone scan is not laser scanning...

The drone scans use 3D Photogrammetry to create 3D from the 2D photos

- go check out FlyPT's http://forum.wmdportal.com/showthread.php?11094-Cheap-alternative-to-laser-scan thread on the old WMD forum.. it's basically the same sort of thing.. and you may remember some of those experiments Casey did with scanning model cars to make 3D reference models using photography then stitching them together to form a 3D model

Roger Prynne
24-04-2017, 16:02
What he said ^

Yeah I knew that :barbershop_quartet_

Mad Al
24-04-2017, 16:11
What he said ^

Yeah I knew that :barbershop_quartet_


It's your age Roger.. probably needed a longer kip this afternoon ;)




also, just out of interest for those that like this sort of thing..

https://www.raspberrypi.org/blog/pi-3d-scanner-a-diy-body-scanner/

some people just have too much time on their hands...

Roger Prynne
24-04-2017, 16:32
Yeah the nurse is going to give me a bed bath in a moment and I was getting confused thoughts.

rosko
24-04-2017, 16:56
All of them? They originally said all new tracks & some of the older ones. I know Andy Tudor said that all tracks are having laser scan & drone scan but i thought it was a mistake/slip of the tongue on his part. Any chance of confirming that?

dault3883
24-04-2017, 17:02
was Le Mans done this way on PC1

Mahjik
24-04-2017, 17:38
was Le Mans done this way on PC1

It was not. Only 4 tracks were scanned in PC1.

RacingAtHome
24-04-2017, 17:47
It was not. Only 4 tracks were scanned in PC1.

Was just the MSV circuits. Brands, Snetterton, Oulton and Cadwell.

dault3883
24-04-2017, 17:55
It was not. Only 4 tracks were scanned in PC1.

I thought that it was narrow from after Mulsanne Corner all the way through to the Ford Chicanes. I have a hard time keeping the car straight in that length.

Mad Al
24-04-2017, 18:01
All of them? They originally said all new tracks & some of the older ones. I know Andy Tudor said that all tracks are having laser scan & drone scan but i thought it was a mistake/slip of the tongue on his part. Any chance of confirming that?

All new tracks are scanned as per what Andy said... the only exception is any track that is new to the game that is either a historic track or is a fictional track (both for obvious reasons, you can't scan what doesn't exist)

Yes there is at least one track that has been redone using laser scan data that was in PC1..

RacingAtHome
24-04-2017, 18:02
All new tracks are scanned as per what Andy said... the only exception is any track that is new to the game that is either a historic track or is a fictional track (both for obvious reasons, you can't scan what doesn't exist)

Yes there is at least one track that has been redone using laser scan data that was in PC1..

And that's the Nurburgring/Nordschleife.

Mad Al
24-04-2017, 18:04
And that's the Nurburgring/Nordschleife.

I can neither confirm nor deny anything... (except that Ian already confirmed that over on GTP)

Mahjik
24-04-2017, 19:59
I thought that it was narrow from after Mulsanne Corner all the way through to the Ford Chicanes. I have a hard time keeping the car straight in that length.

Even though it may not have been scanned, every track had professionals who have driven the real tracks verify their authenticity. The main areas scanning helps with is elevation changes (i.e. getting the correct angles).

The way to verify if something is off width-wise is to get a screenshot with cars 2-3 wide (depending on the track). Then compared that to a real life photo to see how much room is taken. We had people swearing Bathurst was off until similar comparisons were done.

rosko
24-04-2017, 20:59
I'm really not that bothered about scanned tracks as long as they flow well, but i'm surprised they didn't just redo them all considering their budget & the crazy amount of detail they are putting everything.

Roger Prynne
24-04-2017, 21:39
I'm really not that bothered about scanned tracks as long as they flow well, but i'm surprised they didn't just redo them all considering their budget & the crazy amount of detail they are putting everything.

Game development is just like any other business, a certain amount of the budget for this and a certain amount for that etc etc.

dault3883
24-04-2017, 22:27
Even though it may not have been scanned, every track had professionals who have driven the real tracks verify their authenticity. The main areas scanning helps with is elevation changes (i.e. getting the correct angles).

The way to verify if something is off width-wise is to get a screenshot with cars 2-3 wide (depending on the track). Then compared that to a real life photo to see how much room is taken. We had people swearing Bathurst was off until similar comparisons were done.

Im sorry if you thought i was saying it was narrower than it is in real life thats not what i was trying to say. what i meant to say is its narrower than im used to it being. That being said im used to the way they have the track on the PS2 game Le mans 24 hours.Which may or may not be accurate to the life size track. 237510

breyzipp
25-04-2017, 07:04
I'm really not that bothered about scanned tracks as long as they flow well, but i'm surprised they didn't just redo them all considering their budget & the crazy amount of detail they are putting everything.

Yeah I agree it's a bit of an odd decision. There are so many new tracks to the game which all have been done with drone/laser scanning. The 4 old ones from PC1 (Brands, Snetterton, Oulton and Cadwell) were already laser scanned and the Nurburgring has been redone as well.

That leaves us with a lot of famous and really important tracks which are not scanned (Le Mans, Spa, Silverstone, Hockenheim, Bathurst, Road America, ...) which would make a big deal for me if they were scanned as well. But I understand this is probably the backfiring of PCARS 1 already having a big track roster resulting in a lot of work to be redone which does not immediately offer something new to the table for the mass crowd. Well, I sure hope these tracks will be scanned in the future as well and be updated by DLC or else PCARS3. It would be so nice to have every real life track scanned and accurate to the cm.

hkraft300
25-04-2017, 07:16
I thought all the pc2 tracks, including existing pc1 tracks, will be drone scanned, going by Tudor's interviews.

Leper Messiah
25-04-2017, 08:08
right so I'm actually more confused now!

Drone scanning is NOT laser scanning, all new tracks are laser scanned, nords has been laser scanned (does that include GP track?), add the 4 MSV trackjs laser scanned from PC....is that info right?

Sankyo
25-04-2017, 08:40
right so I'm actually more confused now!

Drone scanning is NOT laser scanning, all new tracks are laser scanned, nords has been laser scanned (does that include GP track?), add the 4 MSV trackjs laser scanned from PC....is that info right?
New tracks are drone scanned.

hkraft300
25-04-2017, 09:15
New tracks are drone scanned.

But the SMS drone has laser+ photogrammetry onboard, no?
From the PC gamer presentation by Tudor.

https://youtu.be/Uw9ezoqLuqA

Mad Al
25-04-2017, 09:26
New tracks are drone scanned.

except for those that were laser scanned ;)
(which aren't always mentioned in the track build notes, but definitely LB was.. Mark put in loads of info about ref collecting. Daytona is actually using both laser and drone)

@Leper, basically it will make little difference to the end result if they were laser or drone scanned, the data is used as a reference to get the basic shape, gradients and camber correct and even the data used for the surface (which is not part of the visible rendered track) will be an interpolated version of the raw data..
Some circuits are now having scans done so they can supply developers with the data directly (and would probably insist on it's use as part of the license), which may mean the scan can be a couple of years old, but would still be correct as far as basic road detail goes (tracks change slightly year on year but unless you build on a marsh, it should stay basically the same...)

Honestly, it's making a mountain out of a molehill to worry about which circuit was scanned with which technique, as you really won't know one way or the other when driving on them.

Zpectre87
25-04-2017, 10:26
except for those that were laser scanned ;)
(which aren't always mentioned in the track build notes, but definitely LB was.. Mark put in loads of info about ref collecting. Daytona is actually using both laser and drone)

@Leper, basically it will make little difference to the end result if they were laser or drone scanned, the data is used as a reference to get the basic shape, gradients and camber correct and even the data used for the surface (which is not part of the visible rendered track) will be an interpolated version of the raw data..
Some circuits are now having scans done so they can supply developers with the data directly (and would probably insist on it's use as part of the license), which may mean the scan can be a couple of years old, but would still be correct as far as basic road detail goes (tracks change slightly year on year but unless you build on a marsh, it should stay basically the same...)

Honestly, it's making a mountain out of a molehill to worry about which circuit was scanned with which technique, as you really won't know one way or the other when driving on them.

Would it be accurate to assume drone scanning is better for topography with laser scanning being strictly for the track's surface?

IMO a drone is just that much more convenient to use than a cumbersome laser scanner, which is why it took a while for long tracks such as the Nordschleife to be scanned, but I could be wrong as I'm not a specialist. :D

Of course, if the drone's sensors have long enough range and high enough sensitivity, SMS could just use drones for the whole thing while saving on time and resources.

hkraft300
25-04-2017, 10:36
...

Of course, if the drone's sensors have long enough range and high enough sensitivity, SMS could just use drones for the whole thing while saving on time and resources.

Seems it's all done with the drone. Laser mounted on drones can do surface and topography, I don't see why it can't.
It's also helped SMS photography staff avoid run-ins with American police too lol

Mad Al
25-04-2017, 11:25
http://www.expouav.com/news/latest/figuring-aerial-surveying-drone-instead-arguing-photogrammetry-vs-lidar/

A drone is able to do both, it doesn't mean that it did always/ever use both in this case, some tracks are pure photo based, some are laser (and that may be ground based) and at least one is a combination.. as far as accuracy goes, for photo stuff, it's down to flying height, image resolution and how much processing power you are prepared to throw at the problem... and ultimately how much resolution you actually need, at 50mph you are sampling every 3.7 cm if you have a physics engine running at 600Hz, so by 200mph that's gone up to about 15 cm..

I'll also add, that the scan will include all the other stuff around a circuit and can help with accurate placing (and scaling) of buildings, fences, Armco, trees.. so it's not just about the track surface

dault3883
25-04-2017, 17:22
Yeah I agree it's a bit of an odd decision. There are so many new tracks to the game which all have been done with drone/laser scanning. The 4 old ones from PC1 (Brands, Snetterton, Oulton and Cadwell) were already laser scanned and the Nurburgring has been redone as well.

That leaves us with a lot of famous and really important tracks which are not scanned (Le Mans, Spa, Silverstone, Hockenheim, Bathurst, Road America, ...) which would make a big deal for me if they were scanned as well. But I understand this is probably the backfiring of PCARS 1 already having a big track roster resulting in a lot of work to be redone which does not immediately offer something new to the table for the mass crowd. Well, I sure hope these tracks will be scanned in the future as well and be updated by DLC or else PCARS3. It would be so nice to have every real life track scanned and accurate to the cm.

If we have already paid for the tracks we shouldn't have to pay for them again just because they have been "fixed". So i would hope the DLC would be a free one, or else just include the fixed tracks in an update or something.

RacingAtHome
25-04-2017, 21:15
If we have already paid for the tracks we shouldn't have to pay for them again just because they have been "fixed". So i would hope the DLC would be a free one, or else just include the fixed tracks in an update or something.

It will be Project CARS 3 when it would be updated.

dault3883
25-04-2017, 21:21
It will be Project CARS 3 when it would be updated.

so basically in another 3-4 years

Roger Prynne
25-04-2017, 21:33
If we have already paid for the tracks we shouldn't have to pay for them again just because they have been "fixed". So i would hope the DLC would be a free one, or else just include the fixed tracks in an update or something.

Without trying to sound rude, you have no idea the amount of work/time/money that goes into these tracks, even the ones you call fixed.
Also don't forget that even the tracks from pCARS1 have had a hell of a lot of 'new' work done to them, as well as all the new LiveTrack 3 stuff.

dault3883
25-04-2017, 21:39
Without trying to sound rude, you have no idea the amount of work/time/money that goes into these tracks, even the ones you call fixed.
Also don't forget that even the tracks from pCARS1 have had a hell of a lot of 'new' work done to them, as well as all the new LiveTrack 3 stuff.

Your right i have no idea. That being said i wasnt implying that they HAD to be free or that thats what they SHOULD do, what i said was i HOPE. Can you blame a guy for hoping for something free. I would just hope that they dont have us buy those tracks in a dlc, then turn around and laser scan them. Then make a dlc again essentially having us buy two dlc's for the same thing. Im not saying i wouldnt buy it. Im just saying it wouldnt be my most happy buy. Im trying to say this without coming across like im entitled to it cause im not.

FlyPT
25-04-2017, 21:53
some people just have too much time on their hands...


Wish I had more... Maybe I could have concluded PVS2.



If anyone wants a taste of something similar to photogrametry, chek my SIG.


https://www.youtube.com/watch?v=uy49bTGuanE

breyzipp
26-04-2017, 06:38
Your right i have no idea. That being said i wasnt implying that they HAD to be free or that thats what they SHOULD do, what i said was i HOPE. Can you blame a guy for hoping for something free. I would just hope that they dont have us buy those tracks in a dlc, then turn around and laser scan them. Then make a dlc again essentially having us buy two dlc's for the same thing. Im not saying i wouldnt buy it. Im just saying it wouldnt be my most happy buy. Im trying to say this without coming across like im entitled to it cause im not.

So far on PCARS1 DLC I have not even had the feeling once that I was being ripped off. The on demand program was great with some awesome tracks like Ruapuna and Bannochbrae added to them. Hands down the best DLC was the classic Lotus expansion with 2 layouts of Rouen-Les-Essarts, historical versions of Silverstone and Hockenheim (yes that is 4 new tracks in a car pack DLC) and a handful of classic Lotus cars added to the mix of which one was free for everyone. All that for what? 7 EUR or something? Over on the other side at Turn 10 studios I have feel ripped off A LOT more, paying for some cars twice, a Volvo V60 that's basically the same as the S60 I already payed for. And their 7 EUR car pack DLCs never include any tracks, they call that an expansion and it costs 15-25 EUR...

SMS DLC policy has been outstanding for PCARS1 so let's give them the benefit of the doubt for PCARS2 shall we?

I would like a 100% laserscanned game with over 60 locations as well but these things happen to cost a lot of money to create.

hkraft300
26-04-2017, 07:56
Pc1 dlc was very reasonably priced. Hope they continue to buck the trend of rip-off dlc. Pretty sure it's been announced the cost of pc2 dlc will rise, but I doubt it will be T10 money.
Even then, buy what you want. Every pc1 dlc has had at least 1 car I wanted the dlc for, but turned out to have 2-3 cars that are also amazing to drive.
Also, SMS didn't and hasn't done BS car+location counts.

breyzipp
26-04-2017, 08:46
Pc1 dlc was very reasonably priced. Hope they continue to buck the trend of rip-off dlc. Pretty sure it's been announced the cost of pc2 dlc will rise, but I doubt it will be T10 money.
Even then, buy what you want. Every pc1 dlc has had at least 1 car I wanted the dlc for, but turned out to have 2-3 cars that are also amazing to drive.
Also, SMS didn't and hasn't done BS car+location counts.

Ian posted something about DLC on the GT Planet thread a while ago. IIRC then PCARS2 would have less frequent DLC (not monthly) but each DLC would have more content too. So yeah it's normal the price goes up then.

Ryzza5
26-04-2017, 09:30
You'll find some info on this in this technical demonstration

https://youtube.com/embed/Uw9ezoqLuqA

F1_Racer68
26-04-2017, 12:37
Here's some infor about photogrammetry. It's a technique that has been around since the invention of the camera and the hot air balloon. It's essentially what cold war era spy planes such as the U2 and SR71 were used for.

https://en.m.wikipedia.org/wiki/Photogrammetry

It is most commonly used in Forensics, Archeology, map making, and movie making (The Matrix being one of the first).

It uses 2 or more cameras to capture an object from multiple angles simultaneously. It then uses those images to establish reference points which software can then use to recreate a digital version of the object. This is essentially exactly what laser scanning does as well. If done with a high enough resolution camera system it is nearly impossible to tell the difference in the finished product.

As has been mentioned as well, the photogrammetry technique provides data for a wider filed of view than most of the laser systems. This can be beneficial in regards to the track surroundings. Where laser scanning is focused on the track surface and edges, photgrammetry would include data about trackside structures, trees, specific objects, etc.

rosko
26-04-2017, 13:25
Do you have any evidence that its impossible to tell the difference from a laser scan track surface & photogrammetry in the finished product? I'm not talking about how it looks but im talking about the surface detail of the track.

Mad Al
26-04-2017, 13:54
Do you have any evidence that its impossible to tell the difference from a laser scan track surface & photogrammetry in the finished product? I'm not talking about how it looks but im talking about the surface detail of the track.
Can you tell the difference between laser and non laser in the original ?

rosko
26-04-2017, 14:12
Can you tell the difference between laser and non laser in the original ?

That's a question to a different question. I've never driven on any of those tracks so how can i?

beetes_juice
26-04-2017, 14:33
http://lanmarservices.com/2014/11/07/laser-scanning-vs-photogrammetry/

rosko
26-04-2017, 14:47
http://lanmarservices.com/2014/11/07/laser-scanning-vs-photogrammetry/

So according to that its not as accurate. Just want to make sure we don't make claims about this technology that is misleading.

Anyway i'm pretty sure they use both methods for each new track, I'm pretty sure Andy said that, the photogrammery is used more for the surroundings & laser to get the track detail.

Mad Al
26-04-2017, 15:00
So according to that its not as accurate. Just want to make sure we don't make claims about this technology that is misleading.

Anyway i'm pretty sure they use both methods for each new track, I'm pretty sure Andy said that, the photogrammery is used more for the surroundings & laser to get the track detail.

the point is both are accurate enough... there is sod all point measuring down to a micron of accuracy if it's not going to be noticed in game.

F1_Racer68
26-04-2017, 15:01
Do you have any evidence that its impossible to tell the difference from a laser scan track surface & photogrammetry in the finished product? I'm not talking about how it looks but im talking about the surface detail of the track.

Nice try.... If you want to address my comments, then please at least have the decency to quote me accurately. My exact words were "nearly impossible".

But to answer your question:


https://www.youtube.com/watch?v=0SLQvxyV_Vw

The example in the video was done using a 6MP camera. A higher resolution camera would yield more data in the image and therefore more accurate detail in the end product.

There is also this comparitive study:

237521

So, as I said, in the end result, it is NEARLY impossible to tell the difference.

And for those interested in understanding the Photogrammetry vs laser (LiDAR) scanning technologies, there is this video:


https://www.youtube.com/watch?v=XYgtBuUusDM

Roger Prynne
26-04-2017, 15:30
Wow! you started something here Leper.... good info, keep it coming guys.

rosko
26-04-2017, 16:20
the point is both are accurate enough... there is sod all point measuring down to a micron of accuracy if it's not going to be noticed in game.

Some people want the best accuracy though, I mean i would not know if a car was accurately represented but i still want it to be accurate. I'm not bothered so much about tracks personally but some people are, just go onto the ac or iracing forums & some will not accept anything other than a laser scan. So my point being in that case it's important to establish if photogrammy is as accurate.

rosko
26-04-2017, 16:22
Nice try.... If you want to address my comments, then please at least have the decency to quote me accurately. My exact words were "nearly impossible".

But to answer your question:


https://www.youtube.com/watch?v=0SLQvxyV_Vw

The example in the video was done using a 6MP camera. A higher resolution camera would yield more data in the image and therefore more accurate detail in the end product.

There is also this comparitive study:

237521

So, as I said, in the end result, it is NEARLY impossible to tell the difference.

And for those interested in understanding the Photogrammetry vs laser (LiDAR) scanning technologies, there is this video:


https://www.youtube.com/watch?v=XYgtBuUusDM

My mistake nothing intentional there, simply misread it. And thanks for the links.

Mahjik
26-04-2017, 16:35
lol, I love the idea of using an article from a laser scanning company to prove whether the two technologies are equivalent. ;) Of course their own article is going to drive people to their business and say the other is inferior.

F1_Racer68
26-04-2017, 16:47
lol, I love the idea of using an article from a laser scanning company to prove whether the two technologies are equivalent. ;) Of course their own article is going to drive people to their business and say the other is inferior.

In all fairness, LiDAR scanning is more accurate, but that level of accuracy is negligible. Every article I have read, or video I have watched about the photgrammetry method will state the same thing. And the level of accuracy is dependant on the quality of cameras used (for obvious reason).

The truth is, for the level of detail required in a racing sim, Photogrammetry is more than sufficient. After all, we don't need ot have every single gap between pieces of aggregate in the tarmac mapped. It's enough to knwo that the gap is htere and acts as it should. It doesn't need to be micrometer accurate in size, depth or geolocation to the real track. If you are after THAT level of detail, then I might suggest that you seek professional help :D

Mahjik
26-04-2017, 16:54
There is only so much that gets transferred from the imagining to the 3D. I'm sure you know this F1_Racer68, but others probably do not. The question is whether photogrammetry can capture enough detail for what is used in sim racing, and the answer is yes. I will say, those skeptical will not be able to tell between the SMS tracks which are laser scanned or which were done via photogrammetry.

For those wondering, yes you still have track surface detail with photogrammetry (i.e. micro-bumps as the sim community has been fond of naming them).

F1_Racer68
26-04-2017, 16:59
There is only so much that gets transferred from the imagining to the 3D. I'm sure you know this F1_Racer68, but others probably do not. The question is whether photogrammetry can capture enough detail for what is used in sim racing, and the answer is yes. I will say, those skeptical will not be able to tell between the SMS tracks which are laser scanned or which were done via photogrammetry.

For those wondering, yes you still have track surface detail with photogrammetry (i.e. micro-bumps as the sim community has been fond of naming them).

Yup, EXACTLY.

Here's how I look at it....

If Photgrammetry is good enough for archaeologists, and criminologists to use for investigative purposes and archiving purpses, then I have no doubt it is good enough for a sim racer. The level of detail captured by it is more than enough to give us what we need for the track and surrounding details. Couple it with GPS mapping, and it is a highly accurate reflection of the real location.

Leper Messiah
27-04-2017, 19:53
except for those that were laser scanned ;)
(which aren't always mentioned in the track build notes, but definitely LB was.. Mark put in loads of info about ref collecting. Daytona is actually using both laser and drone)

@Leper, basically it will make little difference to the end result if they were laser or drone scanned, the data is used as a reference to get the basic shape, gradients and camber correct and even the data used for the surface (which is not part of the visible rendered track) will be an interpolated version of the raw data..
Some circuits are now having scans done so they can supply developers with the data directly (and would probably insist on it's use as part of the license), which may mean the scan can be a couple of years old, but would still be correct as far as basic road detail goes (tracks change slightly year on year but unless you build on a marsh, it should stay basically the same...)

Honestly, it's making a mountain out of a molehill to worry about which circuit was scanned with which technique, as you really won't know one way or the other when driving on them.

I just made the mistake of engaging with an iRacing fanboy on Youtube so if it's not laser scanned it's simcade and all that guff!


Wow! you started something here Leper.... good info, keep it coming guys.

I did didn't I but top info as you say, really interesting stuff, cheers for the info everyone! :)

hkraft300
27-04-2017, 22:45
I just made the mistake of engaging with an iRacing fanboy on Youtube so if it's not laser scanned it's simcade and all that guff!


:hopelessness:

Zpectre87
28-04-2017, 14:40
Some people want the best accuracy though, I mean i would not know if a car was accurately represented but i still want it to be accurate. I'm not bothered so much about tracks personally but some people are, just go onto the ac or iracing forums & some will not accept anything other than a laser scan. So my point being in that case it's important to establish if photogrammy is as accurate.

I try to stay away from these kinds of people as much as possible.

There's a lot of placebo effect, misbeliefs and marketing going on in sim racing, which is why I prefer to pick my games based on the overall experience. I maintain that Assetto Corsa is a boring game, the same way rFactor was (I was firmly in SimBin camp last gen). And I remember reading a recent post here from a member stating that he prefers playing F1 2016 because the official ruleset makes the game more immersive than others despite its shortcomings. Since I'm not made of money to try iRacing out, and I've found rFactor 2's graphics a sad joke by current standards, it didn't take me much convincing to adopt pCARS as my main sim.


In all fairness, LiDAR scanning is more accurate, but that level of accuracy is negligible. Every article I have read, or video I have watched about the photgrammetry method will state the same thing. And the level of accuracy is dependant on the quality of cameras used (for obvious reason).

The truth is, for the level of detail required in a racing sim, Photogrammetry is more than sufficient. After all, we don't need ot have every single gap between pieces of aggregate in the tarmac mapped. It's enough to knwo that the gap is htere and acts as it should. It doesn't need to be micrometer accurate in size, depth or geolocation to the real track. If you are after THAT level of detail, then I might suggest that you seek professional help :D

According to the article Mad Al posted, the disadvantages of photogrammetry compared to laser scanning shouldn't really matter when trying to map a track's surface. Photogrammetry is the inferior method when dealing with surfaces blocked by various kinds of objects, which is obviously not the case with race tracks unless we're speaking of the scenery.

bradleyland
28-04-2017, 23:53
Nice try.... If you want to address my comments, then please at least have the decency to quote me accurately. My exact words were "nearly impossible".

But to answer your question:

...

Allow me to paraphrase:

https://media.makeameme.org/created/quotfactsquot.jpg

Ryzza5
29-04-2017, 00:02
So what happens to these players if/when other racing sims begin using other methods instead of laser scanning because the cost-benefit ratio favours it?

Leper Messiah
01-05-2017, 14:36
So what happens to these players if/when other racing sims begin using other methods instead of laser scanning because the cost-benefit ratio favours it?

I felt a great disturbance in the simcommunity, as if millions of voices suddenly cried out in terror and were suddenly silenced.

Chimildo
02-05-2017, 08:37
People try not they do or do not to get a Star Wars quote in

hkraft300
02-05-2017, 14:06
People try not they do or do not to get a Star Wars quote in

That's a terrible Yoda impersonation.

Cheesenium
03-05-2017, 07:06
Some people want the best accuracy though, I mean i would not know if a car was accurately represented but i still want it to be accurate. I'm not bothered so much about tracks personally but some people are, just go onto the ac or iracing forums & some will not accept anything other than a laser scan. So my point being in that case it's important to establish if photogrammy is as accurate.

At the end, it does not matter much because when the artist is building the track into the game, many of the details of the tracks are lost in the process of converting a point cloud to a polygon. The modeler needs the general shape of the track, not the tiny minute details of each pore on the track where having milimeter level of accuracy isn't as useful as it should.

Even at it's worst with drone laser scanning or photoscanning, the accuracy is still at centimeter level of uncertainty which is more than adequate for the usage in a game. Stationary LIDAR will be far more accurate as you get your far denser point clouds at milimeter level. However, how much of the details of the point cloud will be converted to the track polygon? Not to mention, a denser point cloud would make track modelling more difficult to work due to more computational power is needed to manipulate the data and more time needed to complete while stationary LIDAR is tremendously more expensive to hire. Drone based laser scanning cost about a few thousand dollars to scan a 5km road but you'll lose some accuracy. Stationary LIDAR that are far more precise will cost you tens of thousands for the same thing. Thats the figures I got from the industry.

There are many people out there that think LIDAR is the only way to build tracks in a game which is false. Other methods such as referencing, CAD drawings, aero mapping and so on still can produce good results in many games. Meanwhile, among all the racing games in the market, I think it is only iRacing that owns a surveyor level LIDAR equipment to scan their tracks. Other games such as AC does not and their accuracy (2cm) is not that great compared to photoscanning or drone LIDAR. Still, I think SMS did have a really good surveyor for Brand Hatch and a few other UK tracks that provide a good mesh to work with.

Most people just got their heads caught up with the term laser scan that sounds cooler without much understanding on how it works. If laser scanning is that great, surveyors would had dump every other surveying methods (leveling etc) and adopt laser scan. Laser scan is a more convienient method of making tracks but it is not be all, end all methodology.

Moreover, there is a car in Pcars (hint, it is one of the DLC car) that was originally laser scanned but the laser scanned point cloud was discarded because the contractor did such a poor job that the point cloud does not even look like the car. SMS had to model the car with photographs which the car looks pretty much like the real car. The use of laser scan does not guarantee good results and it is not the holy grail of reference collection. At the end, best accuracy means nothing if you do not have the right people to manipulate.

bradleyland
03-05-2017, 15:41
All great points regarding the technology used. You can give an poor workman the finest tools available, and the result still be garbage. What's important to me isn't the tool used, or even the finite technical specification of the scan, but rather what the outcome is. It boils down to this basic question:

Would someone who has driven the real circuit hundreds of times find all the subtleties and nuances of the real circuit in the game?

For me personally, it's about the following (in order of priority):


Track surface nuances - The little bumps, humps, and dips that a track regular knows and incoporates in to their driving.
Trackside details - Signs, corner distance markers, railings, gaps in rails, paint on railings, kerbs, paint on kerbs, run-offs, run-off surface materials, run-off locations, etc. A racer uses all these tiny details as braking, turn-in, apex, and exit targets.
Track atmosphere/environment - The general feel of the track. You needn't replicate every single building, but some elements that aren't trackside are still important to provide a feeling of being there. For example, driving past the Touristenfahrten parking lot and entry point on the Nordschleife, or the tall tree to the left of the entry to Metzgefeld.
Only one of the three on my list have anything to do with the track scanning method, and even with millimeter accurate scanning technology, the artist must know what areas they can discard detail, and where they should keep it. That requires the consult of someone familiar with the track, and at that point, you really don't need the mm detail; you just need to know all the little subtleties.

Leper Messiah
03-05-2017, 18:52
What I find makes SMS tracks above and beyond the competition is how used and alive the tracks look. Most other games tracks look too sterile and perfect....SMS track artists are brilliant IMO, combine that with the accuracy of modern techniques and I'm getting rather hyped for PC2. :)

dault3883
03-05-2017, 23:20
What I find makes SMS tracks above and beyond the competition is how used and alive the tracks look. Most other games tracks look too sterile and perfect....SMS track artists are brilliant IMO, combine that with the accuracy of modern techniques and I'm getting rather hyped for PC2. :)

arent we all

hkraft300
04-05-2017, 01:24
Trackside details - Signs, corner distance markers, railings, gaps in rails, paint on railings, kerbs, paint on kerbs, run-offs, run-off surface materials, run-off locations, etc. A racer uses all these tiny details as braking, turn-in, apex, and exit targets.
Track atmosphere/environment - The general feel of the track. You needn't replicate every single building, but some elements that aren't trackside are still important to provide a feeling of being there. For example, driving past the Touristenfahrten parking lot and entry point on the Nordschleife, or the tall tree to the left of the entry to Metzgefeld.



What I find makes SMS tracks above and beyond the competition is how used and alive the tracks look.

Things photogrammetry is good for.

Alan Dallas
08-05-2017, 07:27
I just made the mistake of engaging with an iRacing fanboy on Youtube so if it's not laser scanned it's simcade and all that guff!



I did didn't I but top info as you say, really interesting stuff, cheers for the info everyone! :)

*Puts on Morpheus sunglasses and changes his voice* "What if I told you that not all of the tracks iRacing 'claims' to be laser scanned actually are?"

konnos
08-05-2017, 10:52
Some people just want to see the world burn...

RacingAtHome
08-05-2017, 11:31
Some people just want to see the world burn...

We do.

breyzipp
08-05-2017, 13:42
What I find makes SMS tracks above and beyond the competition is how used and alive the tracks look. Most other games tracks look too sterile and perfect....SMS track artists are brilliant IMO, combine that with the accuracy of modern techniques and I'm getting rather hyped for PC2. :)

I agree and not only the track itself but also the weather and lighting. Sunset in PCARS1 looks amazing with the warm colour palette and the long shades of buildings and trees. Even real life depressing rain/clouds/fog weather looks depressing in the game. :)

Gran Turismo for example feels too perfect with its clean tarmac and colourful surroundings.

Leper Messiah
08-05-2017, 14:15
*Puts on Morpheus sunglasses and changes his voice* "What if I told you that not all of the tracks iRacing 'claims' to be laser scanned actually are?"

Really? (takes blue pill)


I agree and not only the track itself but also the weather and lighting. Sunset in PCARS1 looks amazing with the warm colour palette and the long shades of buildings and trees. Even real life depressing rain/clouds/fog weather looks depressing in the game. :)

Gran Turismo for example feels too perfect with its clean tarmac and colourful surroundings.

Indeed, no pre baked lighting for Pcars and it's DAMN good looking. I still maintain that Bannochbrae is the most beautiful gaming environment I have had the pleasure to feast my eyes upon. Makes me want to get out of the car and walk around first person to soak up the details.

F1_Racer68
08-05-2017, 14:15
All great points regarding the technology used. You can give an poor workman the finest tools available, and the result still be garbage. What's important to me isn't the tool used, or even the finite technical specification of the scan, but rather what the outcome is. It boils down to this basic question:

Would someone who has driven the real circuit hundreds of times find all the subtleties and nuances of the real circuit in the game?

For me personally, it's about the following (in order of priority):


Track surface nuances - The little bumps, humps, and dips that a track regular knows and incoporates in to their driving.
Trackside details - Signs, corner distance markers, railings, gaps in rails, paint on railings, kerbs, paint on kerbs, run-offs, run-off surface materials, run-off locations, etc. A racer uses all these tiny details as braking, turn-in, apex, and exit targets.
Track atmosphere/environment - The general feel of the track. You needn't replicate every single building, but some elements that aren't trackside are still important to provide a feeling of being there. For example, driving past the Touristenfahrten parking lot and entry point on the Nordschleife, or the tall tree to the left of the entry to Metzgefeld.

Only one of the three on my list have anything to do with the track scanning method, and even with millimeter accurate scanning technology, the artist must know what areas they can discard detail, and where they should keep it. That requires the consult of someone familiar with the track, and at that point, you really don't need the mm detail; you just need to know all the little subtleties.

Totally agree with you on these three points, as well as their order of importance (I think most here would also agree). THe thing is, that all of htese can be achieved with photogrammetry, and in fact, it is in some ways the better choice for addressing the second and third bullets.