Block diagram
Microsoft Windows uses specific file format to store applications icons - ICO. This is a container that contains one or more image files (effectively, DIB parts of BMP files or full PNG files are contained inside).
This page hosts a formal specification of Microsoft Windows icon file using Kaitai Struct. This specification can be automatically translated into a variety of programming languages to get a parsing library.
meta:
id: ico
title: Microsoft Windows icon file
file-extension: ico
xref:
justsolve: ICO
mime:
- image/x-icon
- image/vnd.microsoft.icon
pronom: x-fmt/418
wikidata: Q729366
tags:
- windows
license: CC0-1.0
endian: le
doc: |
Microsoft Windows uses specific file format to store applications
icons - ICO. This is a container that contains one or more image
files (effectively, DIB parts of BMP files or full PNG files are
contained inside).
doc-ref: https://learn.microsoft.com/en-us/previous-versions/ms997538(v=msdn.10)
seq:
- id: magic
contents: [0, 0, 1, 0]
- id: num_images
-orig-id: idCount
type: u2
doc: Number of images contained in this file
- id: images
-orig-id: idEntries
type: icon_dir_entry
repeat: expr
repeat-expr: num_images
types:
icon_dir_entry:
-orig-id: ICONDIRENTRY
seq:
- id: width
-orig-id: bWidth
type: u1
doc: Width of image, px
- id: height
-orig-id: bHeight
type: u1
doc: Height of image, px
- id: num_colors
-orig-id: bColorCount
type: u1
doc: |
Number of colors in palette of the image or 0 if image has
no palette (i.e. RGB, RGBA, etc)
- id: reserved
-orig-id: bReserved
contents: [0]
- id: num_planes
-orig-id: wPlanes
type: u2
doc: Number of color planes
- id: bpp
-orig-id: wBitCount
type: u2
doc: Bits per pixel in the image
- id: len_img
-orig-id: dwBytesInRes
type: u4
doc: Size of the image data
- id: ofs_img
-orig-id: dwImageOffset
type: u4
doc: Absolute offset of the image data start in the file
instances:
img:
pos: ofs_img
size: len_img
doc: |
Raw image data. Use `is_png` to determine whether this is an
embedded PNG file (true) or a DIB bitmap (false) and call a
relevant parser, if needed to parse image data further.
png_header:
pos: ofs_img
size: 8
doc: |
Pre-reads first 8 bytes of the image to determine if it's an
embedded PNG file.
is_png:
value: png_header == [137, 80, 78, 71, 13, 10, 26, 10]
doc: True if this image is in PNG format.