Convert CBZ to PDF

Rename one to .zip and any archive tool opens it. Inside is one image per page and usually nothing else — no manifest, no metadata standard, no text, no encoding of its own. The images are stored rather than re-compressed by the container, so a CBZ of JPEG pages is about the size of those JPEGs added together.

That simplicity is why comic readers standardised on it and why this conversion is unusually predictable. There is no markup to interpret and no styling to lose. Converting to PDF is essentially placing the pictures onto pages in order and adding the document structure around them.

Every other e-book conversion on this site comes with the same warning: the source reflows to fit the screen and a PDF does not, so converting freezes the type at one size forever. It is the single most important thing to say about converting a novel.

Here it does not arise. A comic page is a fixed composition — panels, lettering and art laid out together — and it was never going to reflow. Nothing is being given up by moving it into a fixed-page format, which makes CBZ to PDF one of the few conversions in this family that costs almost nothing in principle.

A CBZ has no index saying which image is page one; readers sort the filenames and display them in that order. Pages named 1.jpg to 12.jpg therefore sort with 10 immediately after 1, and the reader is not wrong — the naming is.

A converter does the same sort, so a badly named archive produces a badly ordered PDF, and fixing it afterwards means re-ordering pages in a document rather than renaming files in a folder. Zero-padding is the entire remedy: 001.jpg, 002.jpg, and so on. Check the order before converting, not after.

A ZIP holding already-compressed images adds almost nothing to their combined size. A PDF adds a page object, a resource dictionary and cross-reference data for every page, and it cannot compress a JPEG any further than it already is.

The increase is modest per page and noticeable across a two-hundred-page volume. Nobody should convert a comic archive expecting to save space — the conversion buys compatibility, and it charges for it in bytes. If the archive needs to be smaller, that has to happen to the images before they go into either container.

The lettering in a comic is part of the artwork. A converted PDF therefore contains no text at all: nothing can be selected, copied, searched, read aloud by a screen reader or indexed by a desktop search tool. It is a stack of pictures with page numbers.

Turning the speech into searchable text is optical character recognition performed on the images, which is a separate operation with its own accuracy problems — comic lettering is stylised, often hand-drawn, and frequently the hardest kind of text to recognise. It is not something a format conversion does or should pretend to do.

A dedicated comic reader has a mode for it: pages advance right to left, and in a two-page view the right-hand page is the earlier one. That setting lives in the reader, not in the CBZ, because the archive has nowhere to record it.

PDF has somewhere to record it and this conversion does not use it. The format carries a /Direction entry in its viewer preferences, set to R2L, which is what Acrobat calls Binding: Right Edge and what tells a reader which way a spread pairs. Nothing here writes that entry, so the file comes out with the default and a two-page view puts the earlier page on the left.

The pages themselves keep the order they had, so single-page reading is entirely correct. For manga read one page at a time on a tablet this is a non-issue; on a wide screen it is the thing that will look wrong first, and the fix is a document-properties change in Acrobat or any tool that can set viewer preferences — once per file, not once per session.

Many scans store a spread as one wide image. In a comic reader that is handled — the page is fitted to the screen height and read across. In a PDF it becomes one very wide page in a document whose other pages are portrait, and the reader zooms out to fit it.

That is a cosmetic irritation rather than a fault, but it decides whether a converted volume reads well. If the archive mixes single pages and spreads, look at a spread in the output early. Splitting spreads into two pages before converting is an option, and it is the kind of decision better made once for the whole archive than page by page later.

When it has to open somewhere a comic reader does not exist. No browser displays a CBZ, most e-reader firmware ignores it, no email client previews it, and a colleague sent one sees an unfamiliar extension and assumes the file is broken. A PDF opens on all of them.

Printing is the other case, and it is where the fixed page is a genuine advantage rather than a compromise. Beyond those, a dedicated reader gives a better experience than any PDF viewer will — fitting pages to screen height, handling spreads, and remembering where you were across devices.

High-resolution scans are what make comic archives large, and neither container compresses them. A volume scanned at 300 dpi for archival purposes can be several hundred megabytes, which is well past the 25 MB free ceiling for a server conversion here and awkward on a tablet regardless.

The fix is to resize and re-encode the images first, keeping the archival scans separately. A page around 1,600 pixels tall is comfortable on almost any screen and a fraction of the size. Do it to the images inside the CBZ, then convert — a PDF built from reduced pages is small; a PDF built from full scans and then compressed is usually worse-looking and barely smaller.