Hello @xzrlalala,

You can use any of the arithmetic operations to add or cast the black image. For example:

    if (vips_black(&t1, 100, 100, "bands", 3, NULL) ||
        vips_linear1(t1, &t2, 1.0, 255.0, NULL) ||
        vips_cast_uchar(t2, &t3, NULL) ||
        vips_copy(t3, &t4, "interpretation", VIPS_INTERPRETATION_sRGB, NULL))
        error ...

The linear will make a float image, so you need to cast to uchar if you want an 8-bit image. You need to set the interpretation to sRGB (it'll be GAA by default).

I meant to add, you can make large images more quickly by making a 1x1 pixel image, then expanding it. This is what new_from_image does:

https://github.com/libvips/libvips/blob/master/libvips/iofuncs/image.c#L2420-L2459