File conversion glossary
Short definitions of the terms you meet when you convert or compress a PDF, a photo, a song or a video. Each one says what the term is in two sentences, why it matters for the conversion and what ConvCC does about it.
Updated:
Inside a PDF
- A PDF text layer
- A PDF's text layer is the real, selectable text behind what you see on the page: characters with positions, fonts and sizes. A PDF made on a computer has one; a scan has only a picture of the page unless OCR adds a text layer.
- A scanned PDF
- A scanned PDF is a PDF whose pages are photographs or scans of paper: each page is one image, with no real text inside. You can read it, but you can't search, select or copy its words.
- An encrypted PDF
- An encrypted PDF is protected by a password. With an open password, it can't be read at all without it; with only a permissions password, it opens but printing, copying or editing is restricted.
- A ToUnicode map
- A ToUnicode map is the table inside a PDF font that says which real character each drawn glyph stands for. Without it, a PDF can look perfect on screen while its text is unreadable to a computer.
- A PDF object stream
- An object stream is a compressed container that stores many small PDF objects (page descriptions, fonts' settings, links) together, instead of one by one in plain text. PDF 1.5 introduced them to make files smaller.
- A4 vs Letter?
- A4 (210 × 297 mm) is the standard paper size in most of the world; US Letter (8.5 × 11 in, 216 × 279 mm) is the standard in the United States and Canada. Letter is slightly wider and shorter.
Reading the document
- Reading order
- Reading order is the sequence in which a person reads the text on a page: down each column, title before body, captions with their figures. A PDF doesn't store it; it only stores where each piece of text is drawn.
- A multi-column layout
- A multi-column layout sets the text of a page in two or more side-by-side columns, as in scientific papers, newspapers, magazines and legal registers. Each column is read top to bottom before the next one.
- A running header
- A running header (or footer) is the line repeated at the top or bottom of every page: the document title, a chapter name, a date, the page number. It belongs to the page, not to the text.
- End-of-line hyphenation
- End-of-line hyphenation splits a word across two lines with a hyphen («conver-» / «sion») so justified text spaces evenly. The hyphen exists only because of where the line ended.
- Table extraction
- Table extraction is finding the tables on a PDF page and recovering their rows, columns and cells as data. A PDF draws a table as text at positions and some lines; it doesn't store it as a grid.
- A merged cell
- A merged cell spans two or more columns (or rows) of a table, usually a group header such as «2025» over «Q1 Q2 Q3 Q4». In HTML it is a colspan or rowspan; in Excel, a merged range.
- OCR
- OCR (optical character recognition) turns a picture of text into real text by recognising each letter. It is what makes a scanned page searchable, copyable and editable.
- A searchable PDF
- A searchable PDF is a scanned PDF with an invisible text layer under each page image, placed word by word over the picture. It looks exactly like the scan, but search, select and copy work.
Output formats
- DOCX
- DOCX is the Word document format: a ZIP of XML files describing paragraphs, styles, tables and images. Word, Google Docs, LibreOffice and Pages all open it.
- XLSX
- XLSX is the Excel workbook format: sheets of cells, each holding a number, text, a date or a formula. Excel, Google Sheets, Numbers and LibreOffice open it.
- Markdown
- Markdown is plain text with light marks for structure: # for headings, - for lists, | for tables, [text](link) for links. It reads well as is and converts cleanly to HTML.
- JSON
- JSON is a text format for structured data: objects with named fields, lists and values, readable by every programming language. It is the usual way programs exchange data.
- Reflowable text
- Reflowable text has no fixed lines: the app that shows it breaks it into lines to fit the screen, window or font size. Word, HTML, Markdown and e-books are reflowable; a PDF is not.
Images and file size
- DPI
- DPI (dots per inch, also PPI) is how many pixels an image has for each inch of the page. A Letter page is 1,275 pixels wide at 150 dpi and 2,550 at 300 dpi.
- Lossy compression
- Lossy compression makes files smaller by discarding detail the eye barely notices, as JPEG does; lossless compression, as in PNG, shrinks files without changing a single pixel.
- EXIF orientation
- EXIF orientation is a flag a camera or phone stores in a photo to say which way is up, instead of rotating the pixels. Photo apps read the flag; many other programs ignore it.
- HEIC
- HEIC is the photo format iPhones use by default: a HEIF file whose picture is compressed with the HEVC video codec. It stores the same photo in about half the space of a JPG.
- WebP
- WebP is an image format made by Google for the web, with lossy and lossless modes, transparency and animation. A WebP photo is usually 25 to 35% smaller than a JPG that looks the same.
Audio and video
- A codec
- A codec is the method used to compress and decompress audio or video: H.264, HEVC, VP9 and AV1 for video; AAC, MP3 and Opus for sound. The codec, not the file extension, decides whether a device can play a file.
- A container format
- A container format is the file that wraps video, sound and subtitles together: MP4, MOV, MKV, WebM and AVI are containers. It says how the streams are stored, not how they are compressed; that is the codec's job.
- Bitrate
- Bitrate is how much data a second of audio or video uses, in kilobits or megabits per second. File size is simply bitrate multiplied by length.
- H.264
- H.264, also called AVC, is the most widely supported video codec: nearly every phone, browser, TV and editor made in the last fifteen years plays it. It is the video inside most MP4 files.