gdkpixbuf.pixbuf

A pixel buffer.

[gdkpixbuf.pixbuf.Pixbuf] contains information about an image's pixel data, its color space, bits per sample, width and height, and the rowstride (the number of bytes between the start of one row and the start of the next).

Creating new [gdkpixbuf.pixbuf.Pixbuf]

The most basic way to create a pixbuf is to wrap an existing pixel buffer with a [gdkpixbuf.pixbuf.Pixbuf] instance. You can use the [gdkpixbuf.pixbuf.Pixbuf.newFromData] function to do this.

Every time you create a new [gdkpixbuf.pixbuf.Pixbuf] instance for some data, you will need to specify the destroy notification function that will be called when the data buffer needs to be freed; this will happen when a [gdkpixbuf.pixbuf.Pixbuf] is finalized by the reference counting functions. If you have a chunk of static data compiled into your application, you can pass in NULL as the destroy notification function so that the data will not be freed.

The [gdkpixbuf.pixbuf.Pixbuf.new_] constructor function can be used as a convenience to create a pixbuf with an empty buffer; this is equivalent to allocating a data buffer using malloc() and then wrapping it with [gdkpixbuf.pixbuf.Pixbuf.newFromData]. The [gdkpixbuf.pixbuf.Pixbuf.new_] function will compute an optimal rowstride so that rendering can be performed with an efficient algorithm.

As a special case, you can use the [gdkpixbuf.pixbuf.Pixbuf.newFromXpmData] function to create a pixbuf from inline XPM image data.

You can also copy an existing pixbuf with the [gdkpixbuf.pixbuf.Pixbuf.copy] function. This is not the same as just acquiring a reference to the old pixbuf instance: the copy function will actually duplicate the pixel data in memory and create a new class@Pixbuf instance for it.

Reference counting

[gdkpixbuf.pixbuf.Pixbuf] structures are reference counted. This means that an application can share a single pixbuf among many parts of the code. When a piece of the program needs to use a pixbuf, it should acquire a reference to it by calling [gobject.object.ObjectWrap.ref_]; when it no longer needs the pixbuf, it should release the reference it acquired by calling [gobject.object.ObjectWrap.unref]. The resources associated with a [gdkpixbuf.pixbuf.Pixbuf] will be freed when its reference count drops to zero. Newly-created [gdkpixbuf.pixbuf.Pixbuf] instances start with a reference count of one.

Image Data

Image data in a pixbuf is stored in memory in an uncompressed, packed format. Rows in the image are stored top to bottom, and in each row pixels are stored from left to right.

There may be padding at the end of a row.

The "rowstride" value of a pixbuf, as returned by [gdkpixbuf.pixbuf.Pixbuf.getRowstride], indicates the number of bytes between rows.

last row in the pixbuf may not be as wide as the full rowstride, but rather just as wide as the pixel data needs to be; that is: it is unsafe to do memcpy (dest, pixels, rowstride * height) to copy a whole pixbuf. Use [gdkpixbuf.pixbuf.Pixbuf.copy] instead, or compute the width in bytes of the last row as:

last_row = width * ((n_channels * bits_per_sample + 7) / 8);

The same rule applies when iterating over each row of a [gdkpixbuf.pixbuf.Pixbuf] pixels array.

The following code illustrates a simple put_pixel() function for RGB pixbufs with 8 bits per channel with an alpha channel.

static void
put_pixel (GdkPixbuf *pixbuf,
          int x,
	   int y,
	   guchar red,
	   guchar green,
	   guchar blue,
	   guchar alpha)
{
 int n_channels = gdk_pixbuf_get_n_channels (pixbuf);

 // Ensure that the pixbuf is valid
 g_assert (gdk_pixbuf_get_colorspace (pixbuf) == GDK_COLORSPACE_RGB);
 g_assert (gdk_pixbuf_get_bits_per_sample (pixbuf) == 8);
 g_assert (gdk_pixbuf_get_has_alpha (pixbuf));
 g_assert (n_channels == 4);

 int width = gdk_pixbuf_get_width (pixbuf);
 int height = gdk_pixbuf_get_height (pixbuf);

 // Ensure that the coordinates are in a valid range
 g_assert (x >= 0 && x < width);
 g_assert (y >= 0 && y < height);

 int rowstride = gdk_pixbuf_get_rowstride (pixbuf);

 // The pixel buffer in the GdkPixbuf instance
 guchar *pixels = gdk_pixbuf_get_pixels (pixbuf);

 // The pixel we wish to modify
 guchar *p = pixels + y * rowstride + x * n_channels;
 p[0] = red;
 p[1] = green;
 p[2] = blue;
 p[3] = alpha;
}

Loading images

The GdkPixBuf class provides a simple mechanism for loading an image from a file in synchronous and asynchronous fashion.

For GUI applications, it is recommended to use the asynchronous stream API to avoid blocking the control flow of the application.

Additionally, [gdkpixbuf.pixbuf.Pixbuf] provides the [gdkpixbuf.pixbuf_loader.PixbufLoader] API for progressive image loading.

Saving images

The [gdkpixbuf.pixbuf.Pixbuf] class provides methods for saving image data in a number of file formats. The formatted data can be written to a file or to a memory buffer. [gdkpixbuf.pixbuf.Pixbuf] can also call a user-defined callback on the data, which allows to e.g. write the image to a socket or store it in a database.

Fluent builder for [gdkpixbuf.pixbuf.Pixbuf]