read.bmp() mismatch between predicted and actual number of bytes in the file

Question

I'm trying to read a BMP file in R using the read.bmp function in the bmp package. However, I get the following error:

Error in read.bmp("tiger.bmp") :mismatch between predicted and actual number of bytes in image

I am not sure how can I get past this error. Any and all help is highly appreciated.

Answer 1

That message means that read.bmp is expecting a certain byte depth, 4 bytes per integer in the binary, with each integer composing a part of the pixel, which you can see in this section of code. The script is defining and testing the dimensions of the file based on the color bit depth and header definitions.

If you look at this code which is under the hood of the function:

  bytes_pixel=h$depth / 8
  row_width_bytes=h$width * bytes_pixel
  # bmp rows are written to lenth of nearest 4 bytes
  rounded_row_width_bytes = ceiling(row_width_bytes/4)*4
  bytes_to_trim = row_width_bytes %% 4
  bytes_to_read=rounded_row_width_bytes * h$height
  if(h$bmp_bytesz==0) {
    if(Verbose) warning("invalid byte size information for image")
  }
  else if(h$bmp_bytesz != bytes_to_read)
    stop("mismatch between predicted and actual number of bytes in image")

h$bmp_bytesz=4

The h$ fields are all defined header settings which set the byte depth of 1L to 4 for various dimensions of the image. This section is checking to see that the file is coming in as expected. It is either 8, 24 or 32 bit because it made it past the first warning above this section. It stopped because there is a problem with the encoded dimensions of the file.

If the file is correctly formatted and not damaged, with 3x8 channels, it should come in properly. (or black and white with 1x8 channels)

Try running:

is.bmp('tiger.bmp') to see if it returns that it is a viable Windows BMP file.