.DOCX File Format – Pictures and drawings

Office Open XML Document

Pictures and shapes sit inside runs. Modern ones are w:drawing elements in DrawingML; older ones are w:pict in VML, which files keep alongside DrawingML for compatibility. The picture’s bytes are a part of their own, usually in word/media/, reached by a relationship.

A picture in DrawingML

<w:r>
  <w:drawing>
    <wp:inline distT="0" distB="0" distL="0" distR="0">
      <wp:extent cx="1828800" cy="1371600"/>
      <wp:docPr id="1" name="Picture 1" descr="A chart of sales"/>
      <a:graphic>
        <a:graphicData uri="http://schemas.openxmlformats.org/drawingml/2006/picture">
          <pic:pic>
            <pic:nvPicPr>
              <pic:cNvPr id="0" name="image1.png"/>
              <pic:cNvPicPr/>
            </pic:nvPicPr>
            <pic:blipFill>
              <a:blip r:embed="rId4"/>
              <a:stretch><a:fillRect/></a:stretch>
            </pic:blipFill>
            <pic:spPr>
              <a:xfrm><a:off x="0" y="0"/><a:ext cx="1828800" cy="1371600"/></a:xfrm>
              <a:prstGeom prst="rect"><a:avLst/></a:prstGeom>
            </pic:spPr>
          </pic:pic>
        </a:graphicData>
      </a:graphic>
    </wp:inline>
  </w:drawing>
</w:r>
An inline picture, 2 inches by 1.5 inches. The relationship rId4 points to the image part.
ElementMeaning
wp:inlineThe picture sits in the line of text, like a large character.
wp:anchorThe picture floats, placed relative to the page, margin or paragraph: wp:positionH, wp:positionV, wrapping elements, behindDoc.
wp:extentThe size: cx and cy in EMU, 914,400 to the inch.
wp:docPrIdentity and alternative text: name, descr, title.
a:blipThe image: r:embed for a part in the package, r:link for an external file.

Common image parts are PNG, JPEG, GIF, TIFF and BMP. Word also writes EMF and WMF pictures, and SVG with a PNG beside it as an extension of the blip. An r:link points outside the file, and the viewer on this site never loads it.

VML

VML is the older drawing language, in the namespaces with prefixes v and o. It appears in older documents, and in the fallback of mc:AlternateContent.

<w:pict>
  <v:shape style="width:144pt;height:108pt" type="#_x0000_t75">
    <v:imagedata r:id="rId5" o:title="Picture"/>
  </v:shape>
</w:pict>
A VML picture: v:imagedata names the image by relationship ID.

Shapes, text boxes and charts

Word 2010 and later shapes use wps:wsp inside w:drawing, and are usually in the Choice of an mc:AlternateContent with VML in the Fallback. A text box holds w:txbxContent, whose children are ordinary block-level content: paragraphs and tables. A chart has a c:chart element whose relationship points to a chart part, normally in word/charts/, which has an embedded workbook of its data. SmartArt uses dgm:relIds and several diagram parts.

Embedded objects

w:object holds a preview picture in VML and an o:OLEObject element whose ProgID names the program and whose r:id points to the data, usually a part in word/embeddings/. Embedded Office documents are whole files of their own: a .xlsx or .docx inside the package. A .bin part holds other OLE objects.