Re: [R] RGL 'front' upwards default in rendering image

From: Duncan Murdoch <>
Date: Tue 03 Oct 2006 - 18:32:43 GMT

On 10/3/2006 2:27 PM, Joe Byers wrote:

> Thanx for the help.  Other comments are posted inline below.
> Joe
> Duncan Murdoch wrote:

>> On 10/3/2006 11:31 AM, Joe Byers wrote:
>>> thanks for the reply.
>>> I wish I could create a self contained example of the problem, but I
>>> can't.
>>> My task is displaying the surface generated from a simulation of
>>> Forward prices for a selected trajectory. The x axis is time from
>>> 9/25/06 to 3/31/2007 or 133X1 vector of integers. The z axis is the
>>> expiration month of the forwards as integers (1-36), the y axis is
>>> price. For each day(x) there are 36 prices (y). surface3d displays
>>> the surface but displays the z axis in reverse order and the x axis
>>> in reverse order as well. I can rotate the image 180 degrees but
>>> time on the x axis is running from right to left instead of left to
>>> right. My z axis is now in ascending order.
>>> I have read the documentation for rgl and much of the supported
>>> documents on neosciences web site, but I can not understand how to
>>> generate the image with the displaying the x axis left to right and z
>>> axis in ascending order. I think that rotation matrix or a setting
>>> in rgl.materials will do it, but I can't figure out which one. This
>>> mathematical conversions is beyond me.
>> Generally I'd suggest using surface3d, which uses the same conventions
>> as persp and other R functions, rather than rgl.surface, which uses
>> the OpenGL/computer graphics conventions.
>> If you do that, you would want x and y to be the fixed variables (rows
>> and columns) and z to be the response.
> surface3d instead of rgl.surface seems to have worked.  I need to work 
> out issued with the axis labels and placement, but you made some 
> suggestions below that I will try.  I think I have my y and z ases 
> turned around.  The documentation for rgl.surface uses y and surface3d 
> uses z in their examples, but the examples do the same thing.
> One note is surface3d uses a white background instead of black that 
> rgl.surface has.  I did  see the method that I will experiment 
> with to see how this works.

You can play with a variable called r3dDefaults to set your own defaults for par3d(), background, and material. The open3d() function (called by most *3d functions) looks here for parameters that you haven't specified.

Duncan Murdoch


>>> Here is my code
>>> library(rgl);
>>> #set x and z vectors
>>> x<-as.numeric(,order='ymd'));
>>> z<-as.numeric(colnames(a)[1:ncurve]); #only use the portion of the
>>> curve with data, exclude NaN's
>>> res<-par3d('zoom'=5);
>>> #attempt to set the image/window size, constrained to
>>> 0,0,256,256 wish I could change this to 0,0,640,640
>> That's on my wish list too. All I can suggest is that in the current
>> release in Windows in MDI mode, the Windows|Tile command works.
>> (The difficulty is that rgl is portable to 3 incompatible windowing
>> systems. Things like setting the size of a Window thus take much more
>> work than you'd expect. More than three times as much, because I
>> don't think there's anybody who is familiar with all 3 systems.)
> I do understand.  I know windows, and have a linux server running that 
> is used for a data warehouse and web server at home.  Here at the 
> university of Tulsa, I am stuck with Windows only in our college.  But I 
> am not a OS programmer, and for that matter not a very good high level 
> programmer either:).

>>> res<- rotationMatrix(pi,0, 1, 0) # attempt to set rotation, does
>>> not make a difference
>> That just stored a matrix. You need to use view3d to set the
>> userMatrix to that value.
>>> #create a x labels vectors that somewhat matches in the image, trial
>>> and error here
>>> labels<-as.character( could just use rownames(calday.1)
>>> labels<-labels[c(1,n%/%6,n%/%4,n%/%3,n%/%2,(n*2)%/%3,(n*3)%/%4,(n*5)%/%6,n)]
>>> i=1;
>>> y<-ez[,,i]; # set y array
>>> #create the y axis labels,again trial and error to fit, using 7
>>> slots from 0 to the max(y)
>>> labelsy<-as.character(format(max(y)*rep(1/7,8)*(seq(1:8)-1),digits=2))
>>> #pretty colors thank you rgl docs.
>>> ylim <- range(y)
>>> ylen <- ylim[2] - ylim[1] + 1
>>> colorlut <- terrain.colors(ylen) # height color lookup table
>>> col <- colorlut[ y-ylim[1]+1 ] # assign colors to heights for
>>> each point
>>> #generate the surface
>>> rgl.surface(x,z,y,ylim=c(0,max(ez[,,i])),back='lines',color=col)
>>> title3d(xlab='Date',zlab='Contract Month');#,ylab='Price $');
>>> #sub='Forward curve Trajectory 1',
>>> axis3d('x--',labels=labels,nticks=9)
>>> axis3d('y+-',labels=labelsy,nticks=8)
>>> axis3d('z--',labels=as.character(z),nticks=ncurve);
>> You can use the "at" argument to set ticks at some location and
>> "labels" to set the text to something else. For example, you might
>> want to reverse one of the axes if the direction isn't the one you want.
>> Duncan Murdoch
>>> Thank you
>>> Joe
>>> Duncan Murdoch wrote:
>>>> Joe Byers wrote:
>>>>> The documentation for surface3d and rgl.surface in the package RGL
>>>>> states
>>>>> "'surface3d' always draws the surface with the `front' upwards
>>>>> (i.e. towards higher 'z' values). This can be used to render
>>>>> the
>>>>> top and bottom differently; see 'rgl.material' and the example
>>>>> below."
>>>>> Is there a way to override this default? I have search all the
>>>>> related methods help and the documents on RGL's website.
>>>> There are lots of ways to override it: as the docs say, rgl.surface
>>>> is more flexible than surface3d, and you can draw triangles or quads
>>>> arbitrarily. What sort of thing do you want to do?
>>>> Duncan Murdoch
>>> ------------------------------------------------------------------------
>>> ______________________________________________
>>> mailing list
>>> PLEASE do read the posting guide
>>> and provide commented, minimal, self-contained, reproducible code.
>> mailing list PLEASE do read the posting guide and provide commented, minimal, self-contained, reproducible code. Received on Wed Oct 04 04:45:58 2006

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.1.8, at Tue 03 Oct 2006 - 19:30:06 GMT.

Mailing list information is available at Please read the posting guide before posting to the list.