System.OutOfMemoryException and System.ApplicationException when signing/rendering a word doc

We have two issues with a word document. The document has a lot of large images in it ((14 large images, each on a seperate page)).

  1. When we try to sign the word document, Aspose library throws OutOfMemoryException. After trying to sign a few times, it might succeed at some point.
  2. Then we tried to render the signed word document to PDF file, but it throws below exception: "System.ApplicationException: Image top or bottom margin is too large 1.65,0 image file:C:\WINDOWS\TEMP\Aspose.Words.5351c6aa-89e5-4f89-8e17-8536b10d2de9.005.png"

Can you please have a look at the two issues here and let us know what could have been the cause of the issue?


Thanks for your request.

  1. Could you please provide me sample document and code that will allow me to reproduce the problem?
  2. Which versions of Aspose.Words and Aspose.Pdf are you using? Also please attach your document for testing.
    Best regards.

Signing process description:

  1. Client/user post Base64 encoded word document in XML format to Aspose server
  2. Aspose server receives the xml string and decode it to be be raw bytes and keep it in memeory stream
  3. Aspose server sign the document - replace a predefined token in word doc with a signature image
    Below is the simplified code scriptlets for signing the attached word document, which throws OutOfMemory exception.
xml = "base64 encoded string of the bytes content of the word document to be signed"
XPathNavigator docNav = CreateDocNavigatoe(xml);
protected static XPathNavigator CreateDocNavigatoe(String xml)
    byte[] bytesXml = ASCIIEncoding.ASCII.GetBytes(xml);
    MemoryStream memoryStream = new MemoryStream(bytesXml);
    XPathDocument doc = new XPathDocument(memoryStream);
    return doc.CreateNavigator();
private MemoryStream GetSigningConfirmStream()
    byte[] base64EncodedDocBytes = Convert.FromBase64String(GetDocumentContent());
    return new MemoryStream(base64EncodedDocBytes);
private String GetDocumentContent()
    return GetNodeValue(docNav, _xpath_RequestDetail_Document_Content);
private Signature[] GetSignatures(bool isWordDocType)
    XPathExpression expr = docNav.Compile(_xpath_RequestDetail_Signatures_signature);
    XPathNodeIterator iter = docNav.Select(expr);
    Signature[] signatures = new Signature[iter.Count];
    int i = 0;
    while (iter.MoveNext())
        XPathNavigator cloneNav = iter.Current.Clone();
        signatures[i++] = CreateSignatureFromRequest(cloneNav, isWordDocType);
    return signatures;
private Document _asposeDocument; //memory stream of the word document
private bool SignAtToken(Signature[] _signatures, int index)
    String signatureToken = GetTokenOrBookMark(_signatures, index, SIG_TOKEN_VALUE);
    if (!(signatureToken.StartsWith("@") && signatureToken.EndsWith("@")))
        return false;
    // Get all of the Run nodes in the body (includes tables)
    NodeList runNodes = _asposeDocument.FirstSection.Body.SelectNodes("//Run");
    bool signed = false;
    int iBookMarkPosition = 0;
    int latestMatchTokenPosition = -1;
    // Check each Run for the token. If found, replace it with signature.
    foreach (Run run in runNodes)
        string text = run.Text;
        int iMatchTokenPosition = index;
        // Check to see the signature
        int iToken = text.IndexOf(signatureToken);
        if (iToken != -1)
        // if (text.Equals(signatureToken) )
            DocumentBuilder builder = new DocumentBuilder(_asposeDocument);
            // Move to where the token is
            // Clear the token (leave the run), so when user resend the request, it will not get signed again.
            run.Text = "";
            // Start a bookmark to contain the signature token (so we can unsign it later)
            // We use the signatureMark here so we can unsign the doc if we need to.
            // if we find a new token, then start the book mark position from 0 again,
            // So when we unsign it will be easier.
            // use latestMatchTokenPosition to flag a new token found here.
            if (latestMatchTokenPosition != iMatchTokenPosition)
                iBookMarkPosition = 0;
            String signatureMark = SIG_TOKEN_BOOKMARK_PREFIX + (iBookMarkPosition++) + GetTokenOrBookMark(_signatures, iMatchTokenPosition, SIG_TOKEN_VALUE);
            // Insert the number of space before the token if any.
            // Insert the image in place of the token
            // End the bookmark
            signed = true;
            latestMatchTokenPosition = iMatchTokenPosition;
    return signed;

For the second issue, looks like there was a similar issue reported before. I attached a new document 07ML05789_assign_1.3.doc that is causing the same kind of exception "“System.ApplicationException: Image top or bottom margin is too large 1.65,0 image file:C:\WINDOWS\TEMP\Aspose.Words.5351c6aa-89e5-4f89-8e17-8536b10d2de9.005.png”. "

Thank you for additional information.

  1. I cannot reproduce OutOfMemory exception on my side. I use the latest versions of Aspose.Words (6.0.1) and Aspose.Pdf (3.9.0) for testing.
  2. Regarding the second problem, I managed to reproduce the problem. However, the error occurs on Aspose.Pdf side. Therefore, I think you should ask this question in Aspose.Pdf forum.
    In additional, In the current version of Aspose.Words there are two ways to convert Word document to PDF.
  3. Direct conversion to PDF using Aspose.Words (without using Aspose.Pdf). See the following link for mor einformation:
    here is code:
Document doc = new Document(@"Test012\in.doc");
  1. Another way is using Aspose.Words and Aspose.Pdf (old method):

Best regards.