Vtf2img
A Python library to convert Valve Texture Format (VTF) files to images
Install / Use
/learn @julienc91/Vtf2imgREADME
vtf2img
A Python library to convert Valve Texture Format (VTF) files to images.
Installation
With pip
$ pip install vtf2img
From source
$ git clone https://github.com/julienc91/vtf2img
$ cd vtf2img
$ python setup.py install
QuickStart
As a library
from vtf2img import Parser
vtf_file = "materials/models/weapons/customization/paints/custom/workshop/ak47_asiimov.vtf"
parser = Parser(vtf_file)
header = parser.header
print(f"VTF version: {header.version}, Image size: {header.width}x{header.height}")
# VTF version: 7.5, Image size: 2048x2048
image = parser.get_image()
image.show()
image.save("ak47_asiimov.png")
The result of get_image is an instance of Pillow's Image class,
which allows exporting in many different formats (PNG, JPG, BMP, ...).
As a CLI command
If you installed the package as instructed in the previous section, a vtf2img command should
be available.
$ vtf2img dragon_awp.wtf awp_dragon_lore.png
Otherwise, you can execute the package's __main__.py file with:
$ python -m vtf2img dragon_awp.wtf awp_dragon_lore.png
Limitations
- many image formats supported by the VTF format are not (yet?) supported by this program (see the section below)
- this program only considers the hi-res image contained in the VTF file
- all the flags that might have been set by the VTF's creator are purely ignored
VTF compatibility
Works with VTF version 7.2+.
Compatibility of image formats:
| Format | Supported | |-------------------|:---------:| | A8 | ✗ | | ABGR8888 | ✓ | | ARGB8888 | ✓ | | BGR565 | ✗ | | BGR888 | ✓ | | BGR888_BLUESCREEN | ✗ | | BGRA4444 | ✗ | | BGRA5551 | ✗ | | BGRA8888 | ✓ | | BGRX5551 | ✗ | | BGRX8888 | ✗ | | DXT1 | ✓ | | DXT1_ONEBITALPHA | ✗ | | DXT3 | ✗ | | DXT5 | ✓ | | I8 | ✗ | | IA88 | ✗ | | P8 | ✗ | | RGB565 | ✗ | | RGB888 | ✓ | | RGB888_BLUESCREEN | ✗ | | RGBA16161616 | ✗ | | RGBA16161616F | ✗ | | RGBA8888 | ✓ | | UV88 | ✗ | | UVLX8888 | ✗ | | UVWQ8888 | ✗ |
Most VTF Files are using either DXT1 or DXT5.
Note: DXT1 and DXT5 decoding algorithms were adapted from Harry Jeffery's implementation on vtf2png.
