Reading with a screen reader

When launching Thorium, if a screen reader is detected, the relative functions are activated.

Note that screen readers use a separate buffer to access the HTML content which results in the visual scroll / page offset to be out of sync. In other words, Thorium is generally not aware of the user's current reading location inside the screen reader buffer, unless the screen reader user triggers standard web API events (such as mouse click anywhere in the text, or keyboard tab into focusable HTML elements).

Accessibility features addressed to assistive technologies are strongly affected by the quality of the book files.

Keystrokes common to NVDA and JAWS

(contribution from Prashant Verma, DAISY Consortium)

For reading text, navigating in tables, navigating in the browser window, the insertion key can be replaced by the modifier key of the screen reader (e.g. the Caps Lock key).

The screen reader must be in navigation or virtual cursor mode.

Action	Command
Next title	`H`
Next table	`T`
Next chart	`G`
Next list	`L`
Next form field	`F`
List of elements	`INSERT` + `F7` (works regardless of cursor mode)

Use the above keys with the Shift key to move to the previous item.

Text

Action	Command
Previous character	`←`
Next character	`→`
Previous word	`Ctrl` + `←`
Say Next word	`Ctrl` + `→`
Previous line	`↑`
Next line	`↓`
Previous sentence	`Alt` + `↑` (Jaws only)
Next sentence	`Alt` + `↓` (Jaws only)
Read all	`INSERT` + `↓`
Announce Font	`INSERT` + `F`

Tables

Action	Command
Cell to the right	`Alt` + `Ctrl` + `→`
Cell to the left	`Alt` + `Ctrl` + `←`
Cell below	`Alt` + `Ctrl` + `↓`
Cell above	`Alt` + `Ctrl` + `↑`

Annotating & bookmarking

Based on the feedback we have received so far, it looks like it is very difficult for screen reader users to select an actual range of characters. Such assistive technology allows users to select text inside the screen reader’s virtual buffer, so the application isn’t aware that the user is reading at a particular location and is selecting text there.

Consequently, screen reader users prefer to use the bookmarking feature which anchors a simple, optionally-labelled bookmark to a coarse reading location (there is an icon button in the top toolbar, but the CTRL + B keyboard shortcut can be used too). Thorium actually does pick-up the mouse cursor click on a single text character, as this helps focus a bookmark on a more precise HTML location.

In Thorium, currently the bookmark label editor supports multiple lines of text input, which makes it possible to annotate coarse document locations. This can therefore be perceived as a replacement for the actual annotator, but there is no import/export of bookmarks data.

More feedback is welcome thru our discussion tracker there:Message to Thorium screen reader users… #2867.

Using bookmarks with NVDA

Windows screen readers use a virtual buffer for reading. To place a bookmark where you are currently reading press the key combination NVDA + Enter, and NVDA will say “activate “. This synchronizes the virtual buffer with the current real location so Thorium can deduce the object on which the cursor is located. Then use Ctrl + B to place a bookmark. A sound notification will inform you of the number of bookmarks in the current publication.

Reading with a screen reader