
Reviews
Part of How to evaluate accessible technology
Assistive technology compared: screen readers, magnifiers, voice, and switches
Assistive tools compared by input, output, learning needs, privacy, compatibility, and fit for real tasks across computers and mobile devices.
What to take away
- These tools solve different access problems and are often used together.
- The best comparison uses the same real tasks, files, apps, and environment for every candidate.
- Built-in availability does not guarantee that a third-party app or document will work well.
- Privacy, fatigue, training, and recovery from errors can outweigh speed in a short demonstration.
- A tool category is not a prescription. The individual user should drive the choice.
Screen readers, magnifiers, voice control, switch access, and captions are not competing versions of one product. They change different parts of the conversation between a person and a device. Comparing them starts with the barrier and the task. Writing that barrier and task down is the first step in the accessible technology guide.
The basic differences
Named examples
- screen reader
- NVDA (free, Windows), JAWS (paid, Windows), VoiceOver (free, Apple), TalkBack (free, Android)
- magnifier
- ZoomText, Windows Magnifier, macOS Zoom, iOS Zoom
- voice control
- Windows Voice Access, macOS Voice Control, iOS Voice Control, Dragon
- switch access
- iOS Switch Control, Android Switch Access, AbleNet switches, RJ Cooper switches
- captions
- YouTube auto-captions, Zoom live transcription, Otter.ai, CART services
Changes
- screen reader
- visual interface to speech or braille
- magnifier
- size, contrast, pointer, or visible area
- voice control
- speech to commands and text
- switch access
- one or more signals to scanning or selection
- captions
- speech and sound to visible text
Common strength
- screen reader
- nonvisual navigation and reading
- magnifier
- keeps visual context
- voice control
- hands-free input
- switch access
- alternative motor input
- captions
- access to spoken media and calls
Common constraint to test
- screen reader
- unlabeled controls or poor focus order
- magnifier
- limited field of view or panning fatigue
- voice control
- noise, privacy, correction, vocabulary
- switch access
- scanning time and accurate timing
- captions
- delay, accuracy, speaker identity
Some users combine a magnifier with speech, captions with visual alerts, or switches with a screen reader. Compatibility between tools is part of the test.
Assistive tools at a glance
Screen reader
- Changes
- speech or braille
- Strength
- nonvisual navigation
- Constraint
- unlabeled controls
Magnifier
- Changes
- size and contrast
- Strength
- keeps visual context
- Constraint
- panning fatigue
Voice control
- Changes
- speech to commands
- Strength
- hands-free input
- Constraint
- noise and privacy
Screen readers
A screen reader exposes interface structure, control names, states, text, and navigation through speech or braille. Good performance depends on the operating system, app, document structure, and user skill. Screen reader users should also review the personal device and account security checklist, since account recovery and permissions affect daily access.
Four name-brand options dominate. NVDA is free and open source on Windows, with a large add-on community. JAWS, from Freedom Scientific, is paid Windows software with a long history in workplaces. VoiceOver ships free on Apple platforms. TalkBack ships free on Android. Each has different braille display support and different web-browsing behavior.
Screen reader test points
- Headings and tables
- Forms and notifications
- Drag-based interfaces
- Authentication prompts
- Error recovery
Test headings, tables, forms, notifications, drag-based interfaces, authentication prompts, and error recovery. Those prompts differ a great deal between sign-in methods, as passwords, passkeys, and MFA compared shows. Speech rate in an experienced user's setup may sound extremely fast to a beginner, so do not compare comfort using someone else's settings.
VoiceOver, the screen reader built into Apple platforms, treats the interface as a hierarchy of elements navigated by keystrokes on a Mac and by location-sensitive gestures on a touchscreen, with spoken descriptions and support for many braille displays. That breadth illustrates why a screen reader is an operating method rather than a simple read-aloud button.
Magnification and visual adjustments
Magnification can enlarge a region or the entire screen. Text scaling changes layout, while zoom may enlarge rendered content. Contrast themes, color filters, pointer controls, and reduced transparency can be combined with either.
Name-brand choices include ZoomText, which adds speech and reading tools to magnification, and the free built-in magnifiers: Windows Magnifier, macOS Zoom, and iOS Zoom. ZoomText is paid software; the built-ins cost nothing but vary in maximum zoom level and how they track the cursor.
Magnification test points
- Enlarged text is clipped
- Dialogs open outside visible area
- Pointer remains easy to find
- Movement between menus and documents
- Panning fatigue over time
Check whether enlarged text is clipped, dialogs open outside the visible area, and the pointer remains easy to find. Changing one output control at a time is how setting up accessibility features keeps that testable. Test frequent movement between menus, documents, and notifications. Excessive panning can make a technically usable setup tiring.
Voice control
Voice control can open apps, select labeled items, dictate text, and run commands. It may reduce physical effort and provide an alternate path when touch or keyboard input is difficult.
Named tools cover a wide range. Windows Voice Access and macOS Voice Control are free and built in. iOS Voice Control is free on iPhone and iPad. Dragon, from Nuance, is paid dictation software with strong medical and legal vocabularies. Otter.ai transcribes meetings, with a free tier and paid plans.
Voice control test points
- Background noise
- Headset use
- Names and technical terms
- Correction commands
- Login screens
- Offline behavior
- Long sessions
Test background noise, headset use, names and technical terms, correction commands, login screens, offline behavior, and long sessions. Review where recordings or transcripts are processed. A shared office or confidential task may require another method even when recognition is accurate. Reviewing which apps hold microphone access is a recurring item in the smartphone upkeep and security checklist.
Switch access
Switch systems convert one or more reliable movements into selection. The interface may scan choices automatically or move focus step by step. External switches can connect through adapters or dedicated controllers.
Named options include iOS Switch Control and Android Switch Access, both free and built in. AbleNet makes switches such as the Jelly Bean and Big Red, widely used in schools. RJ Cooper sells switches and mounting hardware for desktop setups. Prices for external switches typically run from about $50 to $300, depending on size, mounting, and connection type.
Measure missed selections, scan speed, rest position, mounting, cable routing, battery status, and the path out of an unwanted screen. A fast scan is not better if it increases errors or strain. Changes in posture, strength, or fatigue across the day may call for more than one profile.
Captions and visual sound cues
Captions convert speech and relevant sound into text. They can be supplied by the media creator, generated live, or produced through a human captioning service. Quality depends on the method, audio, vocabulary, speakers, and timing.
Named examples span free and paid. YouTube auto-captions are free but uneven. Zoom live transcription is included on paid plans. Otter.ai offers a free tier and paid plans for meetings. CART, provided by human captioners through services such as VITAC, costs more and suits legal, medical, and academic settings.
The NICHD's assistive device types overview spans mobility aids, hearing aids, cognitive aids, and adapted daily-living tools. It also covers computer-access software such as screen readers and voice recognition, plus closed captioning.
Captioning sits alongside hearing aids and other listening supports, not in place of them. Captions should not be treated as covering every hearing or communication need.
Check speaker names, punctuation, non-speech sounds, delay, language, saved transcripts, and whether captions remain visible above controls. For a live medical, legal, educational, or work exchange, evaluate the accuracy and service appropriate to the stakes. Live captioning depends on a stable connection, so the home network maintenance and security checklist covers router updates and coverage that keep captions flowing.
A fair comparison test
Use a five-part scorecard for the same three tasks:
- CompletionCan the user finish independently?
- AccuracyHow many corrections or missed items occur?
- EffortWhat happens to fatigue or discomfort over time?
- PrivacyDoes the method expose speech, content, or recordings?
- ContinuityCan the user recover after restart, update, or disconnection?
Record required workarounds. A tool that works only with one browser, file type, or helper nearby has a narrower fit than the demonstration suggests.
Common questions
Does a screen reader read every item visible on screen?
Not necessarily. It depends on how the interface exposes labels, roles, states, and reading order.
Is magnification the same as increasing text size?
No. Magnification enlarges a view, while text sizing may cause an app or page to reflow its layout.
Can voice control replace a keyboard everywhere?
Not always. Secure screens, noisy spaces, unsupported apps, privacy needs, and correction tasks may require another input method.
Are automatic captions enough for every situation?
No. Their accuracy and suitability vary with audio and context. High-stakes communication may need a different service.







