图片转 PDF

这款工具的作用

This places your images into a PDF, one to a page or several to a sheet. A JPEG or a PNG is embedded byte for byte, because a PDF can carry both encodings natively and decoding a photo only to re-compress it on the way in would cost quality for nothing. WebP, HEIC, AVIF, TIFF, GIF and BMP have to be converted, since a PDF cannot hold them, and that conversion happens exactly once.

Reach for it when a set of pictures has to travel as one document: photographs of a signed contract taken on a phone, an expense claim where every receipt arrived as its own JPG, screenshots for a bug report, or a scanner app's output that an upload form will only accept as a single attachment.

工作原理

1. Drop your images onto this page, or use the file picker. Up to 200 at once, and the list you see is the page order.
2. Drag a row to move an image, or use the arrow buttons if you are working from a keyboard.
3. Choose the page size. Match the image gives every page the shape of the picture on it; A4 or Letter puts them all on standard paper.
4. Set the fit. Contain shows the whole image, cover fills the page and crops whatever hangs over, stretch distorts it to fill exactly.
5. Turn off one image per page if you want a contact sheet, then say how many images go on each sheet.
6. Press Convert. The download starts on its own.

JPEG and PNG go in untouched. Both are encodings a PDF carries natively, so the bytes of your photograph are copied into the document rather than decoded and re-compressed, and a 4 MB JPEG contributes about 4 MB. Most converters decode everything and re-encode it as a JPEG at some quality setting, which is why pictures that went in sharp come out soft. The quality slider here reaches only the formats a PDF cannot hold, and it applies to each of them once.

Phone photographs are the reason auto-rotate exists. A camera almost always stores its picture landscape and writes an EXIF orientation tag saying which way up it belongs, and browsers apply that tag inconsistently when they decode. The tag is read and applied here rather than left to the decoder, which is why a page of photographed documents comes out upright instead of lying on its side.

An image that carries no resolution in its metadata is placed at 96 dpi, the CSS reference resolution. That is what makes a screenshot arrive at roughly the size it appeared on screen rather than blown up to fill a sheet of A3. A scan that records its own dpi is placed at that instead, so a 300 dpi page of A4 lands as a page of A4.

Contain is the default fit for the same reason it is the default everywhere on this site. Cover fills the page and crops whatever will not fit, and quietly trimming the top of somebody's signature is not a trade a tool should make on its own. Stretch fills the page exactly and distorts the picture to do it, which is occasionally right for a background and almost never right for a document.

Nothing is uploaded. Decoding, rotation and page composition all happen in Web Workers inside your browser, so a folder of 200 photographs never leaves the device. That is worth more here than for most file types, because a photograph carries a timestamp and often a location alongside the picture itself.