17 replies to this topic

[burnabao]

When people are using LIDAR data to construct real-world courses, how much manual terrain sculpting is typically necessary?  I am working on le Golf National and although my LIDAR data is at 1m points, there seem to be much bigger undulations to the terrain in real life (from TV footage/photos) than I can see in Unity.  I'm happy to sculpt by hand to get the right look, but just wondered what people's experiences were on this.

 

 

[clubcaptain]

TV camera angles often distort undulations.

[burnabao]

True, although from previous experience I thought TV had a habit of flattening the undulations so that they weren't so obvious - this seems to be the other way around.  Will have to see if I can get some hard data to compare with my heightmap.  Thanks

[shimonko]

Both TV and PP reduce the look of undulations - with the wider field-of-view PP recently brought in, theoretically that would make uphills look even less undulating as the holes will look longer than previously.

 

It's worth however doing a sanity check on  your height value in terrain settings - as it directly affects the size of undulations.

 

A quick way of doing this is to find the height difference between two points in Unity and compare against the same height difference on Google Maps. It won't be exact as Google's heights aren't the best, but it will be close enough as a sanity check.

Many ways of finding a height difference in Unity - e.g. find a corner of a tee box at a high place on the course, one at a low place. Drop a little cube at each point. Work out the difference in vertical positions of the cubes.

And in Google Maps, find the difference in elevations of those two tee box corner points at https://www.daftlogi...nd-altitude.htm.

[DoGgs]

The only sculpting i do is when a FW is not holding a drive, the fw may be sloping left or right in the landing area and the ball just rolls off into the rough every time.  You simply have to go in and remedy this, there are a few courses out there where this happens ( pasatiempo 10th hole being the most recent one i have seen )  With Baltusrol upper i had to use the ramp tool to slightly lift the lower portion of the landing areas on most of the front 9 holes as they are set on the lower part of the mountainside, thankfully its only a subtle adjustment and you would never notice it.

[Kablammo11]

OP, are you sure that your terrain is set to the correct height? If you set too much height, all elevations on the height map will be exaggerated. 

[burnabao]

K11 I believe the heightmap is all set up correctly (LarryKuh made it for me and I have no reason to doubt him!).  It looks like "reality" in a lot of places but then in others it seems to be much flatter vs real life images.  Maybe it's an optical illusion from TV/photo pictures; I'll grab a couple of photos from the IRL course and the view from the same location on my very early build to show what I mean.  Thanks for the input.

[worrybirdie]

Remember that long camera lenses (as used in golf coverage) compress the perspective, also making undulations more dramatic looking.

[Justin9926]

K11 I believe the heightmap is all set up correctly (LarryKuh made it for me and I have no reason to doubt him!).  It looks like "reality" in a lot of places but then in others it seems to be much flatter vs real life images.  Maybe it's an optical illusion from TV/photo pictures; I'll grab a couple of photos from the IRL course and the view from the same location on my very early build to show what I mean.  Thanks for the input.

Have you built it and played it in PG yet? Sometimes what you see in unity looks totally different in game. I thought one or two that I was doing looked to be flat, but once you build it and play it in PG, it looks totally different.

[jt83]

You did well to find French lidar!

[M Rose]

I'll lay the course out right on top and not touch anything. The first thing I'll do is flatten tee boxes and I'll use the elevation from the center point of those tees.

 

Generally I'll re-sculpt most of the bunkers a little bit and maybe exaggerate some of the terrain around them a little bit as I find the lidar can be a little too broad. Fairways and greens will get a very light smooth unless there is some stair-stepping involved, but this is rarely a problem with good lidar.

 

I don't aim for 100 percent accuracy because I'm not convinced it can be done with the current tools... at some point I have to trade a little bit of it for expedience as I nearly always burn out spectacularly after working on a single real course for too long.

 

The current course I'm working on is a really good test because it's a course that I have played more than 200 times in real life, and although some time has now passed since the last time I was there, I've also been able to see it on television.... as far as I can tell nothing has changed except some native grass and trees have grown longer and larger.

[Stingreye]

Definitely toss an OB spline (1 and only 1), tee box, 1 tee, easy medium and hard pins, and green spline then build it and play it.

You will notice that things look VERY different in game.  I thought I dug bunkers that looks realistic in unity but in game were double overhead. MY benches and ball washes were bigger than the avatar    The point is, you need to compile it and see.  ALso see how the ball rolls.  It should be fairly accurate.

[burnabao]

Thanks for all the input. Will restore the original heightmap and do some in game testing then!

[Cintigolfer]

I've found that if there are small hills, 1m lidar flattens them out. I have almost finished creating a course where I was a member for 20+ years and the small hills/mounds were only about 6-12" tall from the lidar data. The real hills/mounds were about 36" tall.

For larger/more gradual slopes, the lidar is more accurate.

[DPRoberts]

I think, at this time, that what you describe is about the best we can hope. I believe slewin (classic renditions) has mentioned using 0.25m data but most are using 1m at best.

Connecting dots from 1m to 1m would mean anything in between is a line, right? We know that would not be accurate. Then, the CF meshes are much like a thick blanket placed over the top of this grid with error. The net effect is a general smoothing or lessening of height.

But, it's a heck of a lot more convenient and accurate than any of us could manufacture by hand. Unfortunately, I also doubt the entities using the Lidar for environmental purposes will care to increase the accuracy of the technology beyond 0.15m. So, once 0.25m data is widely available, I think that'll be as good as it gets.

[burnabao]

@everyone thanks very much for the advice.  Restored heightmap to the LIDAR data and built/played the course in PG and it looks way better.  I think it also helps that I painted the terrain a plain texture which shows the relief bit better than the satellite image overlay.  Need to build/sculpt tees and do some smoothing on/around the greens, lower bunkers etc, but nothing too major at first glance.

[Joe Habiger]

Working with 1 meter lidar in Minnesota I hardly have to do any sculpting, as with Mr. Rose above I just flatten tee boxes, sculpt some bunkers that are not perfect and maybe smooth out and area a bit, sculpt in my water and sculpt a bit to flatten the areas where the houses will go but I really haven't done much to that 1 meter elevation because it was so nice coming in. I am working on Hazeltine National and I am close to letting people have a beta of it but it's all the small work now except the cart paths which will be a major pain in my butt but I am saving that for last.

[prodgers]

Also please remember that unless the course is on the ocean, the elevation is NOT zero. Too many overlook that issue.