These data input functions are deprecated and they will be removed in the next release of hyperspec package. Now functions in package (link) should be used as the alternatives.

Old description:

This function allows ENVI data import as hyperSpec object.

read.ENVI.Nicolet() should be a good starting point for writing custom wrappers for read.ENVI() that take into account your manufacturer's special entries in the header file.

  file = stop("read.ENVI: file name needed"),
  headerfile = NULL,
  header = list(),
  keys.hdr2data = FALSE,
  x = 0:1,
  y = x,
  wavelength = NULL,
  label = list(),
  block.lines.skip = 0,
  block.lines.size = NULL,
  pull.header.lines = TRUE

  file = stop("read.ENVI.HySpex: file name needed"),
  headerfile = NULL,
  header = list(),
  keys.hdr2data = NULL,

  file = stop("read.ENVI: file name needed"),
  headerfile = NULL,
  header = list(),
  x = NA,
  y = NA,
  nicolet.correction = FALSE



complete name of the binary file


name of the ASCII header file. If NULL, the name of the header file is guessed by looking for a second file with the same basename as file but hdr or HDR suffix.


list with header information, see details. Overwrites information extracted from the header file.


determines which fields of the header file should be put into the extra data. Defaults to none.

To specify certain entries, give character vectors containing the lowercase names of the header file entries.

x, y

vectors of form c(offset, step size) for the position vectors, see details.

wavelength, label

lists that overwrite the respective information from the ENVI header file. These data is then handed to initialize()

block.lines.skip, block.lines.size

BIL and BIP ENVI files may be read in blocks of lines: skip the first block.lines.skip lines, then read a block of block.lines.size lines. If block.lines.NULL, the whole file is read. Blocks are silently truncated at the end of the file (more precisely: to header$lines).


currently unused by read.ENVI, read.ENVI.Nicolet hands those arguements over to read.ENVI


(internal) flag whether multi-line header entries grouped by curly braces should be pulled into one line each.


see details


a hyperSpec object


ENVI data usually consists of two files, an ASCII header and a binary data file. The header contains all information necessary for correctly reading the binary file.

I experienced missing header files (or rather: header files without any contents) produced by Bruker Opus' ENVI export.

In this case the necessary information can be given as a list in parameter header instead:

samplesintegerno of columns / spectra in x direction
linesintegerno of lines / spectra in y direction
bandsintegerno of wavelengths / data points per spectrum
data typeformat of the binary file
11 byte unsigned integer
22 byte signed integer
34 byte signed integer
44 byte float
58 byte double
916 (2 x 8) byte complex double
122 byte unsigned integer
header offsetintegernumber of bytes to skip before binary data starts
interleavedirections of the data cube
"BSQ"band sequential (indexing: [sample, line, band])
"BIL"band interleave by line (indexing: [sample, line, band])
"BIP"band interleave by pixel (indexing: [band, line, sample])
byte order0 or "little"little endian
1 or "big"big endian
"swap"swap byte order

Some more information that is not provided by the ENVI files may be given:

Wavelength axis and axis labels in the respective parameters. For more information, see initialize().

The spatial information is by default a sequence from 0 to header$samples - 1 and header$lines - 1, respectively. x and y give offset of the first spectrum and step size.

Thus, the object's $x colum is: (0 : header$samples - 1) * x [2] + x [1]. The $y colum is calculated analogously.

Nicolet uses some more keywords in their header file. They are interpreted as follows:

descriptiongiving the position of the first spectrum
z plot titleswavelength and intensity axis units, comma separated
pixel sizeinterpreted as x and y step size (specify x = NA and y = NA)

These parameters can be overwritten by giving a list with the respective elements in parameter header.

The values in header line description seem to be microns while the pixel size seems to be in microns. If nicolet.correction is true, the pixel size values (i.e. the step sizes) are multiplied by 1000.


  • read.ENVI.HySpex():

  • read.ENVI.Nicolet():


This function was adapted from caTools::read.ENVI():

Jarek Tuszynski (2008). caTools: Tools: moving window statistics, GIF, Base64, ROC AUC, etc.. R package version 1.9.


C. Beleites, testing for the Nicolet files C. Dicko