Skip to content

CHANGELOG.md

Latest commit

 

History

History
86 lines (67 loc) · 2.66 KB

CHANGELOG.md

File metadata and controls

86 lines (67 loc) · 2.66 KB

Changelog

[0.1.0] - Initial Release

Basic functionality.

[0.2.0]

Merged text and NarrationOptions into a single object parameter

Previous:

narrator.speak(text, options);

Now you pass a unified object, e.g.:

narrator.speak({ text, ...options });
narrator.resolve({ text, ...options });

Sound Settings removed from NarrationOptions use NarrationPlaybackOptions instead

Sound settings are now configured directly on the narrator instance and can be passed to the narrator methods. The Narration now only contains the metadata and the audio string. They were not configured properly before.

const narrator = createMockNarrator();

// set narrator default playback options
narrator.setDefaultPlaybackOptions({
    gain: -5.2,
    playbackRate: 1.2,
});
    
// or provide them per narration
narrator.speak({
    text: "Hello, world!",
    }, 
    {
        gain: -5.2, // Override default volume for this narration
    }
);

// also available when using start()
narrator.start(narration,
    {
        gain: -5.2, // Override default volume for this narration
    }
);

Added support for Minimax TTS Provider

Thanks to NeverOccurs for the contribution! See PR here.

Introduced DedupedProvider

A wrapper provider that deduplicates identical text requests to avoid generating the same audio multiple times when the same text is requested more than once.

Migrated to motion-canvas-cache package

The AudioCache and ViteCachePlugin have been replaced with the generic motion-canvas-cache package. This provides a cleaner separation of concerns:

Breaking Changes:

  • Vite Plugin: Import motionCanvasCachePlugin directly from motion-canvas-cache instead of the old motionCanvasNarratorPlugin
  • AudioUtils: The following methods have been moved to CacheUtils from motion-canvas-cache:
    • AudioUtils.generateAudioId() → CacheUtils.generateCacheKey()
    • AudioUtils.streamToArrayBuffer() → CacheUtils.streamToArrayBuffer()
    • AudioUtils.blobToDataUrl() → CacheUtils.blobToDataUrl()
  • AudioUtils now only contains audio-specific methods: getAudioDuration() and its helpers

Migration Guide:

// Before
import { motionCanvasNarratorPlugin } from 'motion-canvas-narrator/vite-plugin';

// After
import { motionCanvasCachePlugin } from 'motion-canvas-cache/vite-plugin';

The cache directory now defaults to motion-canvas-cache instead of narrator-cache.

Addition of Narrator.resolveAll method

This allows resolving multiple narrations in parallel, e.g.:

const narrations = yield narrator.resolveAll(["A", "B", "C"]);

[0.3.0] WIP