TL;DR:
An assistant reads a picture of your screen. Small text in that picture is made of very few pixels, so letters blur together.
Zooming in before you ask gives each letter more pixels, which makes similar characters easier to tell apart.
Zoom to the part that matters. More resolution on the right area beats more of the page.
A scanned PDF has a table of numbers in tiny print. You ask an assistant to read it and it gets one number wrong, a 6 where there is an 8. You can read it fine when you lean toward the screen. So why does the assistant stumble? Because it is not leaning in. It sees the picture exactly as captured.
The short answer
The assistant does not read your screen the way you do. It gets an image: a grid of coloured pixels. A letter in small text may be only a handful of pixels tall, and at that size an 8, a 6 and a 3 are almost the same blob.
Zooming in does not add new information to the page, but it changes how many pixels the area occupies in the picture. More pixels per letter means clearer shapes, and clarity means fewer misreads. That is the whole idea: more resolution where it counts.
What happens step by step
The screen is captured as a picture with a fixed number of pixels.
Large pictures are often shrunk to a manageable size before they are read, so fine detail can be lost.
Text becomes a few dozen pixels per character. Smaller text means fewer pixels, and fewer pixels mean more guessing.
When you zoom the page, the same letters fill more of the picture, so the capture holds more detail about each one.
The reading improves, especially for digits and similar-looking letters.
Walk-through: tiny text in a scanned PDF
Say the PDF is a US Letter page, 11 inches tall, and the table is printed in 8 point type. A point is 1/72 of an inch, so 8 point is 8 / 72 ≈ 0.111 inch. That is the full height of the type's em box. Digits are somewhat shorter.
Show the whole page fit on a screen where the page area is 900 pixels tall. That is 900 / 11 ≈ 82 pixels per inch. The text is then 0.111 × 82 ≈ 9 pixels tall. Nine pixels for a character with a loop, a bar and a curve is thin. If the picture is also shrunk before reading, it is less.
Zoom the PDF viewer so the table fills the screen instead. With the viewer set to 3× zoom, each inch of page covers 82 × 3 ≈ 246 pixels, so the text is 0.111 × 246 ≈ 27 pixels tall. Now an 8 has a visible gap in its middle and a 6 clearly has an open top. The assistant now gets triple the pixels along each side of a character.
Check the ratio: 27 / 9 = 3, matching the zoom of three. One caution: the scan itself has a limit. If the scan was made at low quality, zooming in only enlarges the blur. Zoom helps up to the quality of the original, and not beyond it.
Zooming also helps with a second problem: layout. In a scanned table, a faint column line can make two neighbouring numbers look like one. Larger type shows the gap between them, so "4 812" is not read as "4812". When the answer is a number you will act on, read the zoomed picture against the original yourself.
Zoom, then ask a narrow question
Zoom to the part you care about, not the whole page. A smaller area at larger size beats a full page at tiny size.
Ask about one thing at a time. "Read the third column" is easier to check than "read the whole table".
Scroll so no important text is cut off by the edge of the window.
Use the app's own zoom rather than your eyes' squint. The browser or viewer zoom re-draws the text at the new size.
What it can't do
Invent detail the scan lacks. Blurred or faded print stays unclear at any zoom.
Be perfect on digits. For anything that matters, check numbers yourself against the page.
See beyond the part you show. Zoom into one corner and the rest of the page is out of view.
Tell you what it cannot read, every time. Sometimes it guesses. If an answer looks odd, zoom and ask again.
Privacy and control
Zooming narrows what is on screen, which can also help privacy: if you only need one paragraph, zoom to it so little else is in the capture. If the PDF holds sensitive data, check what the zoomed view still shows around the edges.
DuckyHelper looks at your screen only while the pill is open and for that answer, so a zoomed view is what it reads that moment.
Try it yourself
Find a PDF or web page with small print. Ask an assistant to read one line of numbers with the page small, then zoom the viewer in a long way and ask again. Compare both answers with the page. If the zoomed read is better, you have seen resolution at work, and you now have an easy habit for hard-to-read screens.
Frequently asked questions
Does zooming in always fix a misread?
No. It helps when the problem is too few pixels. If the original scan is faint or blurry, the extra size only makes the blur bigger.
Should I zoom with the app or with my screen magnifier?
The app's zoom is usually better, since it re-draws the text sharply at the new size.
Get started
