Convert PPT to PDF

The .ppt container dates from 1990 and was superseded in 2007. Everything still arriving in it is therefore at least nineteen years old, and usually comes from one of a small number of places: an internal training library nobody was funded to migrate, a lecturer’s accumulated course material, a conference archive, or an attachment on an email thread that has outlived the people on it.

That age is the whole subject of this page. The deck is not broken, but it depends on an environment that no longer exists — particular fonts, particular clip art, particular embedded components — and every year that passes makes the original less reliable to open. Converting it to a PDF fixes what it looks like while there is still software that can render it correctly. That is the argument for doing it now rather than when somebody needs the material. A conversion performed while a knowledgeable person can look at the result is a migration; the same conversion performed in a hurry two years later, by somebody who never saw the original, is a hope.

The export produces a page for each slide, in the aspect ratio the deck was designed in. For files of this generation that is usually 4:3, which is why the result has margins down the sides when it is opened on a modern widescreen display and why it prints comfortably on A4 or Letter without scaling.

The text remains real text rather than a picture of text, so the PDF can be searched, copied from and indexed. That matters for an archive more than for a presentation: a folder of PDFs is findable by full-text search, where a folder of binary decks needs a program that can open each one to answer the same question. It also means the slides can be quoted. Somebody assembling a new course can copy a definition out of a 2004 lecture without retyping it, which is the difference between an archive that gets used and one that merely gets stored.

Slides are the worst case for font substitution because their text is sized to fit a box exactly. A title that filled one line in the original wraps onto two if the replacement face is a fraction wider, and the second line lands on top of the subtitle. Nothing about that is subtle once you see it, and nothing warns you before you do.

The machine performing this conversion carries metric-compatible stand-ins for the common Microsoft faces — Liberation for Arial, Times New Roman and Courier New, Carlito for Calibri, Caladea for Cambria — plus DejaVu and Noto for wider coverage. Those cases hold their spacing. A deck set in a font somebody bought in 1999 will be substituted with something approximate, so look at the title slides. Title slides first because they use the largest type and the tightest boxes, and section dividers second for the same reason. Body bullets have room to breathe and almost never show the problem, which is why a deck can look fine on slide four and be broken on slide one.

Decks from this period lean heavily on the drawing objects of their time: grouped autoshapes, gradient fills, bevelled WordArt headings, the clip art gallery. Those features were reimplemented rather than carried forward when the format changed, so they are the most likely part of the file to be rendered approximately.

Approximately is usually good enough for an archive and occasionally not good enough for reuse. If the intention is to lift diagrams out of the old deck and put them into a new one, look at the PDF pages holding them before assuming the shapes came through cleanly, because a flowchart that lost its connectors is still a page and no longer information.

A slide of this vintage can contain material owned by another program — a Word table, a chart from a spreadsheet, a formula placed with Microsoft’s old equation component, which Microsoft itself removed from Office in 2018 after it turned out to be exploitable. Those objects are stored inside the file as separate embedded content.

How well each renders depends on how much of it the presentation stored alongside the object, and it varies by object. Technical and academic decks are the ones to check, because a mathematics lecture whose formulas came through as empty boxes is worse than useless — it looks complete. Scan for blanks where symbols should be before archiving the result and discarding the original. Where a formula has not survived, the recovery is usually to open the deck in an office suite that still reads the embedded object, screenshot the equation and place it as an image. That is laborious and it is the only route once the component that drew it no longer ships anywhere.

Every slide is drawn in its final state. Progressive reveals, motion paths, timed sequences and slide transitions have no representation on a printed page, so what the reader receives is the end of each build rather than the sequence that produced it.

For most archived material this is exactly right, because the deck is being read rather than performed. Where it is not right is a slide whose meaning depended on the order — a quiz that revealed the answer, a diagram assembled piece by piece. Those slides are worth splitting into several before converting, if the deck is going to be used for teaching rather than merely kept.

The conversion runs the office engine with no export options set, which produces the slide pages and nothing else. Whatever the presenter typed into the notes pane stays in the original file.

That is a genuine loss for training material, where the notes often hold the script and the slides hold only the headings. Before retiring a deck, open it and check whether the notes carry information that exists nowhere else. If they do, the deck is not really a set of slides and converting it as one loses the half that mattered.

The volume case here is a directory rather than a file: an induction course, a decade of seminar slides, a product training set. Drop the entire folder — each deck is converted in turn, keeps its filename with the extension replaced, and everything comes back as a single ZIP ready to be dropped into a document store.

Sample the results by risk rather than by position in the list. Open the longest deck, the most graphical one, and anything with equations or embedded charts, because those are where the conversion actually struggles. Twenty plain bullet decks will all be fine and tell you nothing about the twenty-first. It is also worth keeping the originals rather than replacing them. A binary deck is small, and holding it alongside the PDF means a better conversion route in five years is still an option — which it is not once the only surviving copy is a page.

Rendering a presentation to pages needs a full office engine, so this pair runs on a server rather than in your browser. The file is sent over an encrypted connection to a container that runs LibreOffice headless and holds nothing else — which the page states rather than leaving you to infer that everything here is local.

The container has no outbound internet route. Each job gets its own scratch directory and its own profile, and that directory is deleted the moment the job ends, whatever the outcome. Anything still running after sixty seconds is killed, because a damaged binary deck is more likely to make the engine hang than to fail cleanly. The free ceiling is 25 MB per file.

A PDF is the right answer when the deck is going into an archive, onto an intranet, into a handout pack or into an email to somebody who may not own an office suite. It fixes the appearance, removes the dependence on fonts and components that are disappearing, and cannot execute anything.

It is the wrong answer if somebody has to present or edit the material again, because recovering an editable deck from a PDF is reconstruction rather than conversion. In that case convert the .ppt into the modern presentation format instead and keep it as a working file, with the PDF alongside as the version that was circulated.