Resolution is the only decision that really matters
A PDF page has no pixels. It stores instructions — draw this text in this font at this position — which a viewer renders at whatever size you ask for. Converting to an image means choosing that size, and the number that controls it is dots per inch.
At 72 dpi a US Letter page becomes 612 × 792 pixels: right for a web page or an email, too soft to print. At 150 dpi you get 1275 × 1650, which is the sensible default for almost everything — sharp on a high-density screen and acceptable on paper. At 300 dpi you get 2550 × 3300, which is genuine print resolution and produces files roughly four times larger.
There is no benefit to going higher than the use requires. A 300 dpi export of a page you are going to post on social media is a slower conversion, a bigger download and an identical result once the platform has resized it.
PNG or JPG
PNG is lossless. Every pixel comes out exactly as rendered, which keeps text edges crisp and flat colours clean, and it supports transparency. It is the right choice for pages that are mostly text, tables, line drawings, logos or screenshots.
JPG throws away detail the eye is least likely to miss, in exchange for files that are often five to ten times smaller. That trade is excellent on photographs and awful on small text, where it produces the grey fringing around letters that makes a document look cheap. If the page is a scanned photograph, choose JPG; if it has words on it that someone needs to read, choose PNG.
What you lose by converting
An image of a page is not a page. The text stops being text: nobody can select it, search it, copy a phone number out of it, or have a screen reader read it aloud. Links stop working. Form fields become a picture of form fields. The file usually gets bigger, sometimes by a lot, because a compact instruction like "draw this paragraph in Helvetica" is replaced by several million pixels.
That is a perfectly good trade when you need an image — a thumbnail, a preview, a slide, something to paste into a document that will not accept a PDF. It is a bad trade when what you actually wanted was a smaller PDF, or a PDF that opens on a device that was struggling with the original.
If your goal is a smaller file, compressing the PDF keeps it a PDF and usually beats converting it to images. If your goal is to stop someone copying the text, be aware that this only inconveniences them: optical character recognition turns your images back into text in seconds.
Long documents and phones
Each page is rendered to a canvas, encoded, and released before the next one begins, so memory stays flat rather than climbing with page count. Even so, 300 dpi across a hundred pages is real work — a few minutes on a laptop and considerably longer on a phone, where the browser also caps how much canvas memory a page may hold at once.
If you only need certain pages, delete the rest before converting. If you need all of a very long document at print resolution, a desktop browser will finish sooner and is less likely to run out of memory partway through. Progress is shown page by page, and you can cancel at any point without leaving a partial file behind.