R help - Jan 2012 - bmp() shifts the image (Windows XP)

When using bmp() under Windows XP, I find that the saved image is a
shifted version of the correct image. Try this:

n<-5
fn<-"01.bmp"
x<-matrix(runif(n*n),nrow=n)
image(x,col=gray(0:255/255),axes=F,frame.plot=F)
bmp(filename = fn,width = n, height = n, units = "px")
par(mar=c(0,0,0,0),pty="s")
image(x,col=gray(0:255/255),axes=F,frame.plot=F)
dev.off()

The image 01.bmp is like this:
22 23 24 25 w
32 33 34 35 w
42 43 44 45 w
52 53 54 55 w
w w w w w w
Where 22 represents x[2,2], etc; w represents a white pixel.

For my application, the image has to be .bmp format. The same shifting
behaviour is seen for large values of n. It is not just due to the
small n value.

For my application, this shifting is important and has to be
eliminated. Please help.

On an unrelated note, I found out that the bmp() code is "smart"
enough to write my image as 8-bit using a palette instead of 24-bit
with 0:255 grey levels if the image being saved does not use all 256
grey levels. I would love to hear it if somebody knows a good way to
make bmp() stupid and always save as 24-bit. My kludge, using 256x256
pixel images, is to tack on an extra row with grey levels 0:255. Then
when displaying, I have to crop the image to get rid of that bogus
row.

Thanks very much for any help!
Bill

On 12-01-01 9:05 AM, William Simpson wrote:
> When using bmp() under Windows XP, I find that the saved image is a
> shifted version of the correct image. Try this:
The image() function isn't designed to be able to do pixel-level 
addressing, so it's not too surprising that some rounding error 
somewhere leads to this.  You could look through the Windows graphics 
device code to fix it.

However, if you really need pixel level addressing, you should be using 
raster objects.  I don't know if someone has written code to output a 
.bmp file, but it's a very simple format, so it shouldn't be too hard, 
especially if you only need a limited range of pixel formats (e.g. 
grayscale).

Duncan Murdoch
>
> n<-5
> fn<-"01.bmp"
> x<-matrix(runif(n*n),nrow=n)
> image(x,col=gray(0:255/255),axes=F,frame.plot=F)
> bmp(filename = fn,width = n, height = n, units = "px")
> par(mar=c(0,0,0,0),pty="s")
> image(x,col=gray(0:255/255),axes=F,frame.plot=F)
> dev.off()
>
> The image 01.bmp is like this:
> 22 23 24 25 w
> 32 33 34 35 w
> 42 43 44 45 w
> 52 53 54 55 w
> w w w w w w
> Where 22 represents x[2,2], etc; w represents a white pixel.
>
> For my application, the image has to be .bmp format. The same shifting
> behaviour is seen for large values of n. It is not just due to the
> small n value.
>
> For my application, this shifting is important and has to be
> eliminated. Please help.
>
> On an unrelated note, I found out that the bmp() code is "smart"
> enough to write my image as 8-bit using a palette instead of 24-bit
> with 0:255 grey levels if the image being saved does not use all 256
> grey levels. I would love to hear it if somebody knows a good way to
> make bmp() stupid and always save as 24-bit. My kludge, using 256x256
> pixel images, is to tack on an extra row with grey levels 0:255. Then
> when displaying, I have to crop the image to get rid of that bogus
> row.
>
> Thanks very much for any help!
> Bill
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.