I’m having problems with the chooser.
If I use an image as a mask - e.g. a black circle with white outer area, and try to use that as a mask for the chooser to select between a constant (set say - to 0) and a perlin map.
The result dosn’t cleanly choose between the two - it seems to be doing some strange sort of averaging which means i’m getting terrain i don’t want - or vice versa.
It worked fine in early versions (prior to 1.0).
is this a known bug - how can I fix it.
The Chooser’s behaviour appears normal and consistent to me in .99, 1.0 and 1.2, unless I’m missing something…
Are you working in WM 1.0, or in the recently released 1.2 ?
Make sure it is not set to to “Add B to A”
Also make sure that the heighfield used for choosing is “totaly white” in the white area, and completely black in the black area. You can Bias/Gain to punch more contrast into it if it’s not.
Perhaps the problem is the mask and not the chooser… how are you creating the mask?
Thanks for the reply
The mask is being created in photoshop in 24 bit colour and saved as a bmp.
I’ve tried using both settings in the chooser (apart from add b to A) always one gives the correct result - the other merely seems to do a partial merge - with the terrain being reduced in height - but not completely replaced by the chosen channel.
I tried with version 1 and updated to 1.2 - same result. I’d add some images - but I don’t know how to do it to this forum.
hm… Check the preview for the inputed file… See if WM really interprets your BMP file as it should be. You can try to compare it with the output of a Constant set to 0, and another set to 1. If the heights match, then you shouldn’t have any problem. I have a feeling it’s WM that is interpreting what your BMP calls black as dark grey, and/or BMP’s white as light grey.
If so, a workaround is to use the bias/gain device to give it more contrast.
The other is to experiment with different file formats. I usually used 24bit TGA, making sure that the colour really is desaturated.
Yep the input file is interpreted correctly.
The black and white components of the input image are the same as a constant of 0 and 1 and I think you’re right about what its doing:
(i.e. I have a feeling it’s WM that is interpreting what your BMP calls black as dark grey, and/or BMP’s white as light grey)
Interestingly I just tried it with a tga image - and it works fine. Seems the problem may be limited to bmp and terragen inputs.
Bizarre.
Thanks for your help.
Hi Terry,
I’d be interested in seeing the problem files for further analysis. You should be able to use the Clamp device set to Expand mode, and then pull each slider back from the extreme a bit, to make sure that any input mask spans the entire 0…1 range.
Mail the files to support@world-machine.com and I’ll take a look at them
Cheers