In this lesson you
- Saw why a scanned page is silent
- Let Proof recognize the text, then read it
- Judged the scan's contrast by eye
Transcript
A page that is only a picture
Scanned documents and readable text. What does a screen reader do with a scanned letter? Usually, nothing at all. This letter came from an office scanner. It looks like text, but every page is a picture. To a screen reader, a picture of words isn't words. Proof says so first. Its card reads Image-only, every fix waits on this. Nothing else can be checked until the words exist as text. Before you fix a scan, ask whether the original file still exists. If it does, exporting it again from Word or InDesign gives better text than any scan can.
Recognizing the text
Here, the scan is all we have. So choose Remediate this document. It runs Proof's automatic pass, which starts by reading the text off the page. Proof recognizes the text on each page, in the document's own language. Then it tags that text, and gives the file a title. For a one-page letter it takes well under a minute. When it finishes, the receipt lists what changed. The page that was only an image now has text. It is tagged, and it has a title.
What a screen reader will read
To check the result, open the Reading step. These are the words a screen reader will now read, in the order it will read them. Each line came from the picture of the page. Read them against the page, because a poor scan gives poor text. If a word was misread, go to the Tags tab, select that line on the page, and choose Edit. There you can correct the words a screen reader speaks. Recognized text also tends to come back with too many top-level headings. Check those on the Tags tab too. The lesson on tags and headings shows how.
Contrast, by eye
Two questions are left, in the Visibility step. On a scan, the text a reader sees is still part of the image. The selectable words are an invisible copy laid over it, so Proof can't measure their contrast. Look at the page yourself. If the text is dark and clear against the paper, it passes. If it's faint or blotchy, choose Fails. A color change can't reach the pixels of a picture, so that page needs a better scan, or the original file. This letter is clear, so choose Passes. The second is use of color. Proof can't read colors from a picture, so look for anything the letter says only with color, like a red word meaning urgent. This letter has none, so choose I checked every flagged place. Both answers are recorded, and Visibility is done. The letter is ready to certify.
When only some pages are scanned
Some documents are mostly real text, with a few scanned pages inside. For those, the Pictures and tables step says how many pages are scanned, and offers Recognize text. It runs on its own, because it re-images those pages. Now you can tell when a page is only a picture, turn it into readable text, check what was read, and judge a scan's contrast by eye.
Recap
That's the end of this lesson.
Maple Ridge Community Health, its documents and its people are fictional, made for these lessons.