Skip to content

Retroreader 2026.10.16

Pre-release
Pre-release

Choose a tag to compare

@SeanRandall SeanRandall released this 16 Oct 14:46
· 11 commits to main since this release

This is the first date-numbered Retroreader release and a substantial change
to how speech engines and firmware images are packaged.

Highlights

  • Use the chosen speech engine consistently across the reader, menus, file
    transfer and NVDA Remote, rather than falling back to OpenEVV in network
    modes.
  • Provide one complete firmware image for each engine and voice. The web
    installer and downloadable files now use the exact same images.
  • Add OpenEVV pause modes: always pause, pause at sentence endings, or disable
    automatic pauses. Eloquence fans will want to set this to always for a faster reading experience, which sounds counterintuitive.
  • Improve long-sentence splitting so speech is not unnecessarily broken in
    the middle of a sentence. Examples of this were rare, but it broke my flow.
  • Add an accessible browser file manager alongside WebDAV file transfer, with
    browsing, upload, download, rename, folder creation and deletion. Opening the
    reader's address in a web browser forwards to the file manager automatically.
  • Expand the website documentation, installation guidance, speech samples and
    measured speaking-rate information.

Identity and storage changes

  • Rename product-specific identifiers from EVV to Retroreader or RR. OpenEVV
    remains the name of one available speech engine.
  • Change device hostnames from EVVxxxx to RRxxxx.
  • Change the hidden SD-card settings folder from .evv to .rr.
  • Supply a new SD-card template for the revised layout. Earlier alpha layouts
    are not supported by this release.

Speech engines

Complete images are prepared for:

  • OpenEVV (Eloquence), US and British English;
  • SVOX Pico, US and British English;
  • DECtalk DTC-01;
  • DECtalk 4.3;
  • OpenBST (Keynote Gold), 1998 English;
  • Monologue XEN11K8; and
  • TruVoice US English.

Packaging

  • Introduce a dated firmware manifest as the single catalogue used by release
    builds, direct downloads and the browser installer.
  • Replace separate bootloader, reader and network-service downloads with one
    complete .bin file per engine and voice.
  • Rebuild the standalone Windows installer for the complete images.
  • Add SHA-256 checksums for downloadable firmware.

Experimental addition

  • Add DoubleTalk Hybrid, based on the early-1990s RC Systems DoubleTalk PC. It retains the authentic firmware front end for pronunciation and prosody while using a native ESP32 acoustic renderer. Rate boost, pitch, voice, tone, articulation and expression are available in Speech settings.