Writing an image to a text file as binary data C#

Question

Hi,

I need to create a file that embeds an image as text within some records. I'm having some trouble writing the images as text. What I'm doing is gathering the image as a byte array from a SQL database (image type) then I'm writing that image to a text file by going through each byte and writing that byte's ASCII equivalent to the file.

Before I can write that image to a text file, I must convert it to a TIFF (it was formerly a jpeg) in CCITT4 format. To double check that this is being done correctly, I also save the stream as a TIFF and view it in "AsTiffTagViewer," which shows that the compression is correct. I AM able to view the tiff in a proper viewer; however, when gathering the text from the file, I am unable to view the image.

Here's the code:

  byte[] frontImage = (byte[])imageReader["front_image"];
    MemoryStream frontMS = new MemoryStream(frontImage);
    Image front = Image.FromStream(frontMS);
    Bitmap frontBitmap = new Bitmap(front);
    Bitmap bwFront = ConvertToBitonal(frontBitmap);
    bwFront.SetResolution(200, 200);
    MemoryStream newFrontMS = new MemoryStream();
    bwFront.Save(newFrontMS, ici, ep);
    bwFront.Save("c:\\Users\\aarong\\Desktop\\C#DepositFiles\\" + checkReader["image_id"].ToString() + "f.tiff", ici, ep);
    frontImage = newFrontMS.ToArray();   
    String frontBinary = toASCII(frontImage); 

        private String toASCII(byte[] image)
        {
            String returnValue = "";
            foreach (byte imageByte in image)
            {
                returnValue += Convert.ToChar(imageByte);
            }

            return returnValue;
        }

It is frontBinary that's being written to the file. Does anyone have an idea as to what is wrong? The tiff that's saved is correct, yet the exact same byte array, when written as ASCII text, is not being written correctly.

Thank you.

EDIT This issue has been corrected by using a BinaryWriter(byte[]) to correctly write the images as text. Thank you all for your help!

Answer 1

Well ASCII is only seven-bit, for one thing. However, I don't believe your code actually uses ASCII. It sort of uses ISO-8859-1, implicitly.

Never treat text as binary or vice versa. It will always lead to problems.

The best way of converting binary to ASCII text is to use Base64:

string text = Convert.ToBase64String(frontImage);
byte[] data = Convert.FromBaseString(text);

Also note that if your code did work, it would still be painfully inefficient - read up on StringBuilders and then consider that your code is semi-equivalent to

Encoding.GetEncoding(28591).GetString(data);

However, base64 is definitely the way to go to convert between text and binary data losslessly. You'll need to convert it back to binary in order to view the TIFF again, of course.

Note that you haven't shown how you're saving or loading your data - you may have problems there too. In fact, I suspect that if you were able to save the string accurately, you might have been lucky and preserved the data, depending on exactly what you're doing with it... but go with base64 anyway.

Answer 2

One approach to taking binary data and converting it to text data is to use a StreamReader and provide the desired encoding. Like Jon mentioned above it is unwise to use ASCII, but in case any one DOES want to stream binary to some other text encoding, here is some code to do it.

public static String GetString(System.IO.Stream inStream)
{
    string str = string.Empty;
    using (StreamReader reader = new StreamReader(inStream, System.Text.ASCIIEncoding.ASCII)) // or any other encoding.
    {
        str = reader.ReadToEnd();
    }
    return str;
}

Answer 3

Is there a specific reason why you use text instead of a binary file?

Storing binary data in text files is always a bad idea since encodings may convert the bytes to another representation and special characters like linefeed may also be treated specially and converted.

Either store the data as byte array in a binary file or use proper binary to ascii conversion like Jon's Base64 proposal or maybe a list of comma separated hex-values is also possible.

Answer 4

If you are writing only the image data to the file, you should not write it as text at all, but as binary data.

If you are mixing text and binary data in the file, you should not convert the binary data to text. It might work with some specific encodings to convert it back and forth, but it certainly doesn't work with any encoding to convert it to unicode characters (using Convert.ToChar).

Do it the other way around. Encode the text into binary data using the GetBytes method of the proper Encoding object, so that you only have binary data to write to the file.

Answer 5

You're probably reading the database back as Unicode, which will alter some of the binary values in the image.

You can use methods on the System.IO.File class to read/save as binary and text. These might help along with the Base64 options mentioned above.