Plasmo Voice Legacy Forge 0.5.0

Full Changelog: legacy-0.4.0...legacy-0.5.0

Plasmo Voice Legacy Forge 0.4.0

Full Changelog: 2.1.17-forge-1.7.10-r3...legacy-0.4.0

Plasmo Voice 2.1.17 for Forge 1.7.10 - R3

Third revision of the manual Plasmo Voice 2.1.17 backport for Minecraft 1.7.10 / Forge, primarily targeting GT New Horizons + lwjgl3ify / LWJGL3.

This release introduces automatic Voice Leveling and built-in update notifications.

Warning

This is still a community-maintained testing backport.

Please test this release carefully on your own setup, especially Voice Leveling with different microphones, players, distances and audio configurations.

If you encounter pumping, clipping, unexpected volume changes, playback issues or update-check problems, please report them with logs and reproduction steps where possible.

Note

This release remains based on upstream Plasmo Voice 2.1.17.

r3 is the revision of the Forge 1.7.10 backport and does not change the Plasmo Voice network protocol.

Voice Leveling

Added optional automatic per-player voice leveling.

The client now can automatically reduce excessively loud microphones and moderately boost quiet microphones so different players sound closer to the same perceived volume.

The new section is available in:

V → Volume → Voice Leveling

Settings:

• Voice Leveling - On / Off
• Target Level - default -18 dB

Voice Leveling is disabled by default.

The processing is performed independently for each player and happens before positional attenuation, so normal proximity behavior is preserved.

This means:

• distant players still sound quieter;
• Sound Occlusion still works;
• Directional Sources still work;
• HRTF / positional audio is unaffected;
• manual per-player volume remains authoritative.

Internal safeguards include:

• smooth attack/release;
• silence/noise gating;
• limited maximum boost/reduction;
• clipping protection;
• coherent mono/stereo processing.

Runtime testing reduced a roughly 21 dB difference between loud and quiet test speakers to less than 1 dB while preserving distance attenuation.

Update Notifications

Added a lightweight update checker specifically for the Forge 1.7.10 backport.

The checker:

• checks published releases from this repository;
• supports backport revisions such as r1, r2, r3, etc.;
• correctly compares numeric revisions such as r9 and r10;
• includes published pre-releases;
• ignores drafts and unrelated tags;
• runs asynchronously;
• performs only one check per client session / dedicated server startup;
• never downloads or installs updates automatically.

When a newer release is available, clients receive a concise chat notification with a clickable GitHub release link.

Dedicated servers log the available update once during startup.

The checker can be disabled with:

check_for_updates=false

No telemetry is collected. The request does not send player UUIDs, names, server addresses, audio devices or voice settings.

Verification

Automated verification:

• 162 tests passed
• 0 failed
• 0 skipped
• verifyShadedJar passed
• full production build passed
• dedicated-server classloading verified

Voice Leveling was runtime-tested with two clients for:

• loud and quiet speakers;
• transitions between different input levels;
• silence/background noise;
• manual player volume;
• distance attenuation;
• Sound Occlusion;
• Directional Sources;
• HRTF;
• disconnect/reconnect;
• OpenAL playback stability.

Update Notifications were tested with:

• live GitHub release data;
• disabled update checking;
• simulated network timeout;
• controlled newer-version notification;
• clickable release links;
• reconnect behavior;
• dedicated-server classloading.

Compatibility

• Minecraft 1.7.10
• Forge 10.13.4.1614
• GT New Horizons
• lwjgl3ify / LWJGL3
• OpenAL Soft
• based on upstream Plasmo Voice 2.1.17

Release files

• plasmovoice-forge-1.7.10-2.1.17-r3.jar
• SHA256SUMS.txt

Please verify the downloaded JAR against SHA256SUMS.txt.

Plasmo Voice 2.1.17 for Forge 1.7.10 - R2

Second revision of the manual Plasmo Voice 2.1.17 backport for Minecraft 1.7.10 / Forge, primarily targeting GT New Horizons + lwjgl3ify / LWJGL3 environments.

This revision contains a small GUI compatibility pass focused on the Players Volume screen.

Warning

This is still a manually maintained Forge 1.7.10 backport and should be tested carefully before wider deployment.

Please verify the release on your own setup, especially with:

• custom 64x64 skins;
• skins using the second head/hat layer;
• different GTNH client configurations;
• multiplayer servers;
• Players Volume filtering/search;
• different GUI scales and resolutions.

If you notice broken avatars, incorrect skin layers, GUI rendering issues, search-field problems or other regressions, please report them with screenshots and client logs where possible.

Note

This release keeps protocol/upstream compatibility at Plasmo Voice 2.1.17.

r2 is only the revision of the Forge 1.7.10 backport.

Changes

Fixed modern GTNH player skin rendering

Player avatars in the Players Volume screen now correctly support modern GTNH skin layouts.

Previously the head renderer assumed a legacy 64x32 skin texture.

GTNH's SimpleSkinBackport converts player skins to the modern 64x64 layout, so the old renderer sampled incorrect texture regions and could show a completely wrong-looking face.

The renderer now:

• detects the actual loaded texture height;
• correctly supports 64x64 modern skins;
• preserves legacy 64x32 compatibility;
• renders the base face layer;
• renders the transparent second head/hat layer;
• supports HD skin texture sizes;
• reuses Minecraft/GTNH's existing cached skin texture;
• does not introduce any custom skin downloader or additional skin cache.

The same head rendering helper is also used by the in-game voice HUD, so modern skin rendering is corrected there as well.

Fixed Players Volume search placeholder

Fixed Search... remaining visible after the search field received focus.

The search field now behaves like a normal textbox:

• empty + unfocused -> Search... is visible;
• empty + focused -> placeholder is hidden;
• typed text -> only the actual query is shown;
• clearing the field while focused -> field remains visually empty;
• clicking outside -> placeholder appears again.

The placeholder is purely visual and is never stored as the actual search query.

The behavior was aligned with the GTNH NEI search field without introducing an NEI runtime dependency.

Verification

Runtime-tested on the GTNH/lwjgl3ify client with:

• SimpleSkinBackport 1.0.3;
• modern 64x64 skins;
• transparent second head layers;
• custom synthetic skin patterns;
• SimpleSkinBackport default skins;
• Players Volume filtering;
• focus/unfocus behavior;
• surrounding GUI controls after avatar rendering.

The modern player head matched the in-world rendered head, including transparency and the second layer.

Automated verification:

• 126 tests passed
• 0 failed
• 0 skipped
• verifyShadedJar passed
• full production build passed
• git diff --check clean

Compatibility

• Minecraft 1.7.10
• Forge 10.13.4.1614
• GT New Horizons
• lwjgl3ify / LWJGL3
• SimpleSkinBackport
• based on upstream Plasmo Voice 2.1.17

Release files

• plasmovoice-forge-1.7.10-2.1.17-r2.jar
• SHA256SUMS.txt

Please verify the downloaded JAR against SHA256SUMS.txt.

Plasmo Voice 2.1.17 for Forge 1.7.10 - R1

Manual backport of Plasmo Voice 2.1.17 for Minecraft 1.7.10 / Forge.

Primarily designed and tested for GT New Horizons environments using lwjgl3ify / LWJGL3 / OpenAL Soft.

Warning

This is still a manually maintained backport and should be treated as a testing release.

Core voice functionality has been extensively tested, but additional testing on different hardware, audio devices, networks, launchers and server environments is highly recommended.

If you encounter voice dropouts, unexpected reconnects, broken microphone activation or other audio issues, please report them with both client and server logs.

Note

This revision contains major playback, voice activation, jitter buffer, UDP reliability and diagnostic improvements based on direct comparison with upstream Plasmo Voice 2.1.17.

For detailed diagnostics, enable:

debug.enabled=true

in the Plasmo Voice client/server configuration.

Debug logging does not dump microphone PCM, encryption keys, UDP secrets or encoded voice payloads.

Major fixes

OpenAL playback

Fixed a major difference between the legacy playback state machine and upstream Plasmo Voice.

Previously, after a short OpenAL underrun, the client could incorrectly treat newly queued audio as an empty stream and repeatedly reset the source.

Typical symptoms:

• voice works normally after joining;
• after approximately 10-15 seconds voice becomes almost silent;
• only occasional fragments remain audible;
• UDP stays connected;
• voice packets continue arriving and decoding;
• reconnecting temporarily restores normal playback.

The playback state machine now follows upstream 2.1.17 behavior much more closely.

The old failure was reproduced during runtime testing and no longer reproduced with the corrected implementation under the same stress conditions.

Voice activation

Voice Activation now checks the raw microphone capture, matching upstream behavior.

Previously, the Forge 1.7.10 port checked the signal after gain, downmix and RNNoise processing.

This caused activation thresholds to behave very differently from upstream.

For example, a threshold around:

-30 dB

could require approximately:

-53 dB to -56 dB

in the old legacy implementation.

No artificial dB offset has been introduced.

UDP reliability

Improved client and server UDP error handling.

Changes include:

• recoverable UDP datagram errors no longer unnecessarily stop the client worker;
• a failed keepalive send to one player no longer stops the entire UDP server;
• stale capture/playback state is cleaned up after hard UDP timeout;
• reconnect lifecycle is more robust;
• ConfigPacket handling no longer depends on an unnecessary race with asynchronous UDP socket initialization.

Upstream-compatible timeout behavior is preserved:

• soft timeout: 7 seconds
• hard timeout: 30 seconds

Jitter buffer

The legacy jitter buffer now follows upstream static/adaptive behavior more closely.

Changes include:

• removed the artificial 50-frame queue limit;
• corrected packet ordering behavior;
• corrected stale packet handling;
• corrected low-delay FIFO behavior;
• aligned source/jitter lifecycle with upstream.

Diagnostics

Permanent client/server debug diagnostics are now available.

Diagnostics include:

• UDP state and traffic;
• keepalive silence and recovery;
• worker health;
• raw and processed microphone dB levels;
• voice activation state;
• OpenAL source state;
• queued and processed OpenAL buffers;
• submitted and consumed audio frames;
• stream restart/re-prime information;
• source gain and distance;
• OpenAL errors.

Debug mode is disabled by default.

Verification

This revision was tested with:

• dedicated Forge 1.7.10 server;
• two simultaneous clients;
• GT New Horizons;
• lwjgl3ify;
• LWJGL3 / OpenAL Soft;
• Push-To-Talk;
• Voice Activation;
• prolonged continuous voice transmission;
• forced OpenAL underruns;
• artificial playback clock drift;
• UDP timeout and reconnect;
• normal disconnect/reconnect;
• dedicated-server classloading checks.

Automated verification:

• 124 tests passed
• 0 failed
• 0 skipped
• verifyShadedJar passed
• production build passed

Warning

Please test this revision before deploying it broadly.

Useful test cases:

• continuous and intermittent speaking for several minutes;
• Push-To-Talk;
• Voice Activation around your normal threshold;
• microphone reconnect/device change;
• player disconnect/reconnect;
• server reconnect;
• proximity range;
• multiple players speaking;
• different microphones/headsets;
• different NAT/network environments.

If voice disappears, becomes intermittent or reconnects unexpectedly, enable debug logging and attach both client and server logs to the report.

Known scope differences

This is a core voice backport, not a complete port of every modern Plasmo Voice subsystem.

Currently intentionally incomplete or out of scope:

• full addon API parity;
• addon-only source types;
• About/Addons UI;
• runtime Crowdin integration;
• update checker;
• full modern permission API;
• some modern platform-specific integrations.

Core proximity voice functionality remains the primary target.

Compatibility

• Minecraft 1.7.10
• Forge 10.13.4.1614
• GT New Horizons / lwjgl3ify environments
• LWJGL3 / OpenAL Soft
• based on upstream Plasmo Voice 2.1.17

Release files

• plasmovoice-forge-1.7.10-2.1.17-r1.jar
• SHA256SUMS.txt

Always verify the downloaded JAR against SHA256SUMS.txt.

Plasmo Voice 2.1.17 for Forge 1.7.10

Note

This is an unofficial Forge 1.7.10 backport of Plasmo Voice 2.1.17 and is not an official Plasmo Voice release.

Warning

There are currently critical issues with UDP connections that may cause instability or connection failures.
A patch addressing these problems will be released soon.

Plasmo Voice 2.1.17 for Forge 1.7.10

A backport of Plasmo Voice 2.1.17 for Minecraft 1.7.10, focused on preserving the original user experience and core voice chat functionality as closely as practical on the legacy Forge stack.

The backport is designed primarily for modernized 1.7.10 environments such as GT New Horizons using lwjgl3ify.

Features

• Proximity voice chat
• Push-to-Talk
• Voice Activation
• Configurable activation threshold
• Microphone and output device selection
• Microphone volume
• Master voice volume
• Per-player volume and mute controls
• Noise Suppression using RNNoise
• Sound Occlusion
• Directional Sources
• Configurable directional source angle
• HRTF support
• Stereo microphone capture
• OpenAL microphone capture
• Java Sound microphone fallback
• Adaptive jitter buffer
• Packet-loss concealment
• Native Opus codec
• Concentus codec fallback
• Positional voice playback
• Voice distance controls
• Voice HUD and player indicators
• Distance visualization
• Original-style Plasmo Voice settings interface
• Server-side mute persistence
• Voice configuration reload
• Dedicated server support
• Singleplayer and LAN support

Networking

The backport includes the Plasmo Voice 2.1.17 protocol implementation with:

• UDP voice transport
• RSA handshake
• AES encrypted UDP payloads
• UDP authentication
• Keepalive and reconnect handling
• Client version validation
• Player state synchronization

Commands

• /vlist
• /vrc
• /vmute
• /vunmute
• /vmutelist
• /vreload

The plasmovoice: command namespace aliases are also supported.

Requirements

• Minecraft 1.7.10
• Forge 10.13.4.1614
• Java 25
• lwjgl3ify-compatible environment

This build was developed and tested primarily for GT New Horizons-style 1.7.10 environments.

Differences from upstream

The goal of this backport is core user-facing parity with Plasmo Voice 2.1.17.

The following upstream functionality is intentionally not included:

• Addon API
• Addon-only static, entity and direct voice sources
• Addons tab
• About tab
• Runtime Crowdin translation downloading
• Vanish integration
• Full upstream permission API

Some internal implementation details are adapted for Minecraft 1.7.10 and Forge, including configuration format, threading, device handling and audio lifecycle.

Verification

The release artifact was built by GitHub Actions on Windows using Java 25.

Verification includes:

• 94/94 automated tests passed
• verifyShadedJar passed
• Dedicated server startup and shutdown
• Two-client voice transmission
• OpenAL microphone capture
• Java Sound microphone fallback
• RNNoise
• HRTF
• Sound Occlusion
• Directional Sources
• Adaptive Jitter Buffer
• Reconnection
• Device switching
• Dedicated server client-classloading safety

Release artifact

plasmovoice-forge-1.7.10-2.1.17.jar

SHA-256:

1ba09aa6913010ae1e93a8a3092d09e91d030cf8db0fd6df77babef17ed0e2df

A SHA256SUMS.txt file is included with the release assets.

Credits

Plasmo Voice is developed by the Plasmo Voice team.

This repository contains an unofficial Forge 1.7.10 backport based on Plasmo Voice 2.1.17.