Retroreader changelog

Retroreader remains alpha software. Releases may change settings, storage layouts and behaviour without preserving compatibility with earlier builds.

2026.10.17 — pre-release

• Add the British Dolphin Apollo II speech synthesizer, reconstructed from the original hardware firmware and rendered natively on the ESP32-S3.
• Provide Apollo controls for rate, voice volume, pitch, its male and non-male speaker tables, and expression. The original Apollo deliberately reduces expression at faster rates.
• Stream Apollo speech as it is synthesized for responsive, interruptible menus and reading, while yielding regularly to the reader interface.
• Fix a crash when moving through engine-specific voice settings.
• Add Apollo to the web installer, downloadable firmware, voice samples and source-provenance information.
• Add the 1988 Speech Plus Prose 2000 v3.4.1 US English synthesizer, with rate, voice volume, pitch and three voice-character controls.
• Add Prose 2000 to the web installer, downloadable firmware, voice samples and source-provenance information.

2026.10.16 — pre-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.

An additional experimental image adds DoubleTalk Hybrid, based on the early-1990s RC Systems DoubleTalk PC. It retains the original firmware front end for dictionary processing, pronunciation and prosody while replacing the sample-by-sample acoustic work with a native ESP32 renderer. Its speech menu includes rate, pitch, voice, tone, articulation, expression and a working faster-than-stock rate boost.

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.

0.5 alpha — 13 October 2026

• Added TruVoice US English to the public installer and downloads.
• Added accessible browser-based SD-card file management.
• Expanded the voice-sample page with approximate speaking rates.
• Expanded website and SD-card documentation.

0.4 alpha — 9 October 2026

• Renamed the project from EVV Reader to Retroreader.
• Added a modular speech-engine interface and public engine selection.
• Added Monologue XEN11K8 with rate, pitch, voice-volume, dictionary and pause controls.
• Added DECtalk DTC-01, OpenBST 1998 English and SVOX Pico builds alongside OpenEVV.
• Added British English OpenEVV and Pico builds.
• Launched the Retroreader website, cross-platform web installer and speech sample pages.

0.3 alpha — 6 October 2026

• Added folder-aware library browsing, recent files and configurable sorting.
• Added plain-text, MP3, M4A and M4B support.
• Added improved EPUB indexing, SD-card caching and support for much larger books.
• Expanded navigation, status and reader settings.
• Improved recording reliability and keypad responsiveness.
• Added conditional File Transfer, NVDA Remote and RemSound modes, shown only when their configuration files are present.
• Added the browser-based Web Serial installer.
• Included clean Wi-Fi, NVDA Remote and RemSound configuration examples in the SD-card template.

0.2 alpha — 1 October 2026

• Added the full SD-card template to the downloadable package.
• Added example configuration files and the reader guide.
• Added my tactile physical description of the Freenove kit to the guide.

0.1 alpha — 1 October 2026

• First public alpha for the Freenove Media Kit for ESP32-S3, 1.14-inch model.
• Included a standalone Windows installer and complete US English, British English and French-voice OpenEVV firmware images.
• The French voice retained the English user interface.