Glossary
The terms you meet when editing photos, publishing audio and tracking search visibility, each explained in two or three sentences.
A
- AI answer engine
-
A search or assistant service that replies to a question with a written answer generated by a language model, usually listing the sources it drew on, instead of or above a list of links.
- AI citation
-
A link to one of your pages listed as a source in an AI-generated answer. Tracked alongside brand mentions, citations show which of your pages the engines trust and how often you appear in answers to the questions that matter to you.
- Alpha channel
-
An extra layer of information that stores how opaque each pixel is, from fully transparent to fully solid. It is what lets a cutout sit on any background. PNG and WebP support it; JPG does not, so transparent areas turn white or black when a cutout is saved as JPG.
- Audiogram
-
A short video made from an audio clip: cover art or a still image, an animated waveform and captions. It lets you share a podcast moment on social feeds that only play video, where most people scroll with the sound off.
C
- Canonical URL
-
The address declared as the main version of a page, with a rel="canonical" link, when the same content can be reached at several addresses. Search engines then concentrate ranking signals on that one URL instead of splitting them.
- Compression artifacts
-
Visible flaws left by lossy compression such as JPG: blocky squares in flat areas, halos along edges and color banding in skies. Every new save adds a little more, which is why it is best to edit from the original file rather than a downloaded copy.
- Core Web Vitals
-
Three metrics defined by Google to measure the real-world experience of a page: loading (LCP), responsiveness (INP) and visual stability (CLS). A page passes when at least 75% of visits meet the “good” threshold for each of the three.
- Cumulative Layout Shift (CLS)
-
A score for how much content moves unexpectedly while a page loads, such as text jumping down when an image or a banner appears above it. Good is 0.1 or less; over 0.25 is poor. Reserving space for images and embeds prevents most shifts.
- Cutout mask
-
A grayscale map that says which pixels belong to the subject: white keeps, black removes, gray keeps partly, which is how fine hair stays natural. Editing the mask instead of the photo means any correction can be undone without losing quality.
H
- HEIC
-
High Efficiency Image Container, the default photo format on iPhones since 2017. It stores a photo in roughly half the space of a comparable JPG, but some older software cannot open it, which is why HEIC photos are often converted before sharing.
- hreflang
-
An attribute that tells search engines which language and regional versions of a page exist, so each searcher gets the right one. Every version must list all the others and itself, and an x-default entry can catch visitors in every other language.
I
- Interaction to Next Paint (INP)
-
How quickly a page responds to clicks, taps and key presses, measured as the delay before the screen visibly updates. Good is 200 milliseconds or less; over 500 milliseconds is poor. It replaced First Input Delay as a Core Web Vital in March 2024.
L
- Largest Contentful Paint (LCP)
-
The time it takes for the largest image or block of text in the viewport to appear. Good is 2.5 seconds or less; over 4 seconds is poor. Heavy hero images, render-blocking scripts and slow servers are the usual causes.
- LUFS
-
Loudness Units relative to Full Scale, the unit used to measure perceived loudness over a whole program. Podcasts commonly aim for about −16 LUFS in stereo, so an episode plays at a similar volume to the others in a listener’s queue.
P
- PPI and DPI
-
Pixels per inch (ppi) measures how densely an image is printed; printers often call it dpi, for dots per inch. 300 ppi is the usual target for photo prints seen up close and 150 ppi for posters seen from a few steps away. Changing the value alone never adds a single pixel.
R
- Rank tracking
-
Recording the position of your pages for a set of keywords, day after day, in a given country, language and device. Positions are measured signed out from each market, so they do not depend on anyone’s location or search history.
S
- Search intent
-
What a person searching actually wants: to learn something, reach a specific site, compare options before buying, or buy now. Pages rank best when they match the intent behind a query, not just its words.
- SERP
-
Search engine results page: everything shown after a search, from the classic list of links to ads, maps, product carousels, videos and AI-generated answers. Ranking positions are counted among the organic results.
- Speaker diarization
-
Working out who speaks when in a recording, so that each passage of a transcript is labeled with its speaker. It is most reliable when the number of speakers is known and when voices do not overlap.
- SRT
-
SubRip Subtitle, the most widely supported subtitle file format. It is plain text: numbered subtitles, each with a start and end time written like 00:01:02,500 --> 00:01:05,000, followed by one or two lines of text.
U
- Upscaling (super-resolution)
-
Increasing the pixel dimensions of an image. Classic resizing stretches existing pixels and looks soft; AI upscaling predicts plausible detail from patterns learned on similar images, which looks sharper but can invent detail that was never there, so faces and text deserve a check.
V
- Visibility score
-
One figure that sums up how visible a site is for its tracked keywords, weighting each position by the share of clicks it typically receives. It moves more steadily than individual rankings, which makes trends easier to read.
W
- WebVTT (VTT)
-
Web Video Text Tracks, the subtitle format read natively by web video players. Close to SRT, it starts with a WEBVTT header, uses a dot instead of a comma before the milliseconds and supports positioning and basic styling.
- Word error rate (WER)
-
The standard measure of transcription accuracy: the words substituted, deleted or inserted, divided by the number of words actually spoken. A 5% word error rate means about one mistake every 20 words, usually on names, numbers and crosstalk.
No results. Try another word.