Terrain tools help???

[rkk01]

So the DEM is not that accurate, such that using accurate gradient data will locally result in deeper cuttings / higher embankments than should be present...

I tend to have an OS 1:25k contour map / satellite photo layer to hand whilst editing terrain to fit - setting the "Level" tool to whatever 10m contour I'm following and then smoothing the terrain between the fixed contour elevations. It seems to work well, but is very time consuming as every mesh intersection has to be manually adjusted up and down

So, which of the terrain tools would be most useful?

The "Smooth" tool seems to just cause a huge land slip / slump...
Or do I just have the tool settings wrong???

Any help much appreciated

[brysonman46]

I find using the "smooth" tool with a large brush, lightly used, to be the best. It seems to take a little time (many hundredths of a second) for the function to activate, so a rotating, rubbing motion is best (like polishing a table). Do not just place it in a spot, or you will get the "landslip".

[AndiS]

You could lay track along the contour lines and then use "snap to track" to form cuttings / embankments. Of course, you will choose rather wide slopes. And you will have to adapt the slope to the steepness of the terrain. Needless to say you take the track out afterwards.

I used that for toy-style settings long ago. A recent effort to create a similar "separation wall" yielded very mixed results.

[rkk01]

Thanks for the advice.

I hadn't considered using track as a contour - probably more accurate than the level tool, but still with the need to blend between each of the 5m of 10m benches.

I do use a similar technique for banks with a constant gradient - set 10-20 parallel tracks at the required gradient and then snap to track. Very good for making earth dams on reservoirs!

[longbow]

Which DEM are you using? The OS Panorama Landform DEM available from UKTS is pretty accurate.

[rkk01]

I'm using the DEM that on the UKTS file archive. The elevation data is pretty poor, IMHO, maybe +/-5 at a guess?

IIRC that OS DEM dataset is based on a 50m grid. In terrain with any strong relief features, averaging the elevation data over a 50x50 tile is bound to result in some variance from "actual". Lay a track through on a known elevation and gradient and you are always going to get some cut / fill features that don't match the actual size / presence of cuttings and embankments on the ground...

[AndiS]

Here is another approach to the issue: I would lay track in parallel to the real on (spaced a bit only for easy selection). Then you set gradient breaks at your heart's content and drag them up and down so that the match the height of the terrain that is shown in the map or on photos. You could start by placing the dummy track using the offset tool and pushing it down by 5 m where you want to see a 5 m embankment in the end.

Then, you use snap-to-track on the dummy track with the widest settings you can get (assuming that they don't destroy nearby slopes). Like 100 m wide top and a very slight gradient of the slope. That should get the near terrain close to what it should be and you can always decorated the farther parts so it does not hurt as much.

Thanks for the information on the OS Panorama Landform DEM (to both of you). Sounds high quality going by the website but 50 m grid is not so much of an improvement over 90 m (or less in the north).

I wondered about the feasibility of getting their 10 m contour line data into Blender but the download is only offered as a single 1.5 GB chunk. Blender outperforms the World Editor in such tasks so much, but all the data conversion is a major pain indeed.