Repository navigation
Releases
v0.1.7
Added
- Native PNG and WebP cel playback with clip-local timing, hold, loop, pingpong, and bounded image caching.
- Camera-relative pan per depth plane and optional filmic root exposure, highlight roll-off, and shadow toe.
- The Morning Shift, an original layered industrial dawn example with illustrated assets, animated rain and fog, and a 12 FPS character sequence.
- The Morning Shift — Into the Shadows, a 15-second three-shot short with an original 12-cel
walking cycle, world-space character lighting, foreground occlusion, and animated wet-pavement
reflections. The example includes the new illustrations, cels, and their drawing source. AETHER_PROFILE=1now reports fitted-image hits, misses, evictions, packed entries, and
raster and cache byte counts. It also times SVG parsing and rasterization, image fitting,
packing, blitting (including alpha blending and unpacking), and frame finishing.AETHER_IMAGE_PACK=offselects the raw fitted-image cache for profiling and output comparisons.
Changed
- Recut Into the Shadows with independently drawn side-elevation and lamp-yard plates,
constant-scale tracking, heel and toe pivots, and synchronized sole shadows. Refined
the street-space lighting masks, reflections, haze, and rain for the 1080p revision. - Into the Shadows hides both shot changes: a motion-blurred foreground column carries a
linearwipe-leftat 5 s and a lamp-lit fog veil covers a 0.14 s crossfade at 10 s.
A regression test renders both windows of the example and checks that the wipe edge
stays under the column, that the crossfade happens under the fog, and that cold,
forked, and reverse-order renders match byte for byte. The 4K master was validated
frame by frame (seeBENCHMARK.md). - Extended the cinematic validation with a 15-frame 4K benchmark, an animated-zoom probe,
native MP4 cadence checks, and byte-identical comparisons with the original continuity PNGs. - Under byte pressure, the fitted-image cache stores compressible SVG rasters in lossless LZ4
row bands within the existing 64 MiB budget. Painting unpacks only the bands it needs into a
reusable 256 KiB buffer. Cache keys, bitmap and sequence storage, and rendered pixels stay the same. - On the native macOS cinematic benchmark, rendering three 4K PNGs took 6.21 s, down from
15.30 s. CPU time fell 59.7% and peak RSS fell 31.1%. The isolated 4K still used 33.7% less
peak RSS and took 0.12 s longer. The 1080p run was effectively unchanged.
Benchmark details - Added regression coverage for packed row bands, clipped alpha blits, and partial bands.
The three 4K continuity PNGs matched byte-for-byte, and a cache-pressure fixture produced
identical segmented exports at 1, 2, 4, and 64 configured threads.
v0.1.5
Added
-
Typography:
font-kerning: normalapplies the font's pair kerning (GPOSkernpair adjustments or
the legacykerntable; on the embedded Roboto it matches Chromium's shaping on every tested pair),
including across the letters ofdata-hf-split="chars".font-style: italic|obliqueuses an
installed italic face or slants the upright one by 14 degrees around the baseline, as browsers do
(generic families always slant the embedded Roboto, so they match on every machine);
<em>and<i>are italic, as elements and as inline runs. The defaultfont-kerning: autodoes
not kern, so existing layouts are unchanged. Seeexamples/typography.html. -
Multi-line kinetic text:
data-hf-splitaccepts<br>; each line becomes a row that follows the
element'stext-align, and the stagger continues across lines. -
Render progress:
renderredraws a progress line with throughput and ETA on an interactive stderr;
--progress jsonwrites JSON lines (done,total,fraction,fps,elapsedMs,etaMs) and
--progress offdisables it.aether_rendersends MCPnotifications/progresswhen the request
carries_meta.progressToken. The library exposesProgressObserver,ProgressUpdate,
encode_native_with_progress, andrender_part_with_progress. Progress never changes the output. -
Distributed rendering:
render --output part.aseg --segments A..Bencodes a range of one-second
segments into a part file, andrender --output video.mp4 --parts a.aseg,...assembles parts (any
order, checked for full coverage and matching settings) with the composition's audio, producing the MP4
a single render writes.inspect --fps Nprints the segment count. The library exposesrender_part,
assemble_parts, andsegment_count.render --frames 0,45,90with{frame}in the PNG name writes
several stills from one process.encode.rsis nowencode/(segments.rs,parts.rs). -
Inline rich text:
<span>,<strong>,<b>, and<code>inside an element with its own text become
styled runs (color, family, weight, letter spacing, fill, stroke, gradient) that wrap and align together.
-webkit-text-stroke(stroked from the font outlines),-webkit-text-fill-color, and
background-clip: textgradient text. Plain text renders byte-identically. Seeexamples/rich-text.html. -
data-hf-drawon inline<svg>: strokes draw themselves along their measured length, staggered in
document order.data-hf-motion-path: any element travels along SVG path data (arcs included), with
optional-rotate="auto". A newvectormodule measures paths. Seeexamples/vector-motion.html. -
clip-pathwithinset(),circle(),ellipse(), andpolygon()(px or %), drawn with exact
antialiased coverage and animated throughclipPathindata-hf-keyframes;mask-imagewith a linear
or radial gradient. Both apply after the element'sfilter. Seeexamples/clip-mask.html. -
Scene transitions
dip-to-black,dip-to-white,flash,blur,whip-left|right|up|down,glitch,
andrgb-split, compiled to filter tracks and new pixel-effect tracks drawn through the effect layer.
data-hf-keyframesacceptsglitch(0 to 1, a deterministic slice displacement that changes 24 times a
second) andrgbSplit(px). Seeexamples/light-transitions.html. -
Audio effects on
<audio>and the sound of<video>:data-hf-volume-ramp(eased gain keyframes in
volumeordb),data-hf-duck(sidechain ducking under trigger clips with look-ahead attack, hold,
release, amount, and threshold), anddata-hf-eq(up to 16 RBJ biquad bands or thevoice,music-bed,
telephone,warm, andbrightpresets).check/aether_checkreportaudio.clips[].effectswith the
EQ response, ducked regions, and automation keys. Clips without effects mix byte-identically. See
examples/audio-effects.html. -
Variable speed on
<video>and<audio>:data-hf-speed(constant rate, 0.05 to 16) and
data-hf-speed-ramp(JSON keyframes withat,speed0 to 16, andeasing) for slow motion, fast
forward, freezes, and speed ramps. Source time is the integral of the rate; a video's sound follows its
picture. Audio is varispeed through an antialiased windowed-sinc reader.checkreportsspeed,
mediaEnd(videos), andmediaSeconds(audio clips). Seeexamples/speed.html. -
--quality draft|standard|high|max(CLI),"quality"inaether.json, andqualityinaether_render:
H.264 quality presets with fixed quantizers 30, 26, 22, and 18 (CRF 28, 25, 21, and 18 with
--encoder ffmpeg).standardis the default and byte-identical to earlier exports. -
filtercolor functions (brightness,contrast,saturate,grayscale,sepia,invert,
opacity,hue-rotate) anddrop-shadow(), after an optionalblur(). On the root, color functions
grade every finished frame. -
mix-blend-modewith the separable modes (multiply,screen,overlay,darken,lighten,
color-dodge,color-burn,hard-light,soft-light,difference,exclusion) andplus-lighter. -
data-motion-bluron the root: deterministic shutter motion blur withdata-motion-blur-shutter
(degrees) anddata-motion-blur-samples(automatic from the fastest motion when omitted). -
Animated filters:
blur,brightness,contrast,saturate,grayscale,sepia,hueRotate, and
invertindata-hf-keyframes, andfilterin CSS@keyframes. An animated blur uses a continuous
exact-variance kernel, so it eases without steps. -
checkreportseffects(motion blur settings, root grade, color filters, blend modes, and animated
filters). -
examples/effects.htmlandexamples/filter-motion.html. -
Composition variables: typed declarations in
data-composition-variables(string,number,color,
boolean,enum,image,font), bound withvar(--id)in CSS,data-var-text,data-var-src, and
data-var-if. Values come from--variablesor--variables-file(render,preview,check,
bench,inspect),"variables"inaether.json, or thevariablesargument ofaether_check,
aether_write,aether_snapshot, andaether_render. Unknown ids, wrong types, and out-of-range values
are errors.checkreportsvariables(values, defaults, uses, unused ids, removed elements). See
examples/variables.html.
Changed
- Determinism: MP4 output is now byte-identical for every
--threadsvalue and every distributed
split, including compositions with cached gradients and soft shadows (premium-promo.htmlused to
differ between--threads 1and4). An uncached draw of a whole-pixel gradient or soft shadow now
computes the same fixed-point texels a cached tile holds, so first draws and snapshots of those
layers can differ from 0.1.0 by one level per channel (two in a few pixels inside filtered layers). - Faster drawing with identical pixels: branch-free per-row kernels (
render/paint/kernels.rs),
compiled for both the baseline target and AVX2 and selected at run time, for two-stop gradients,
radial gradient parameters, cached gradient tiles, shadow masks, and the vignette and grain pass;
directly drawn radial gradients skip the transparent area outside their last stop. At 1080p on one
thread:premium-promo.html−13%,glass.html−15%, drifting glows −9% to −20%, grain −54%,
vignette −19% (median of three runs; seeBENCHMARK.md).AETHER_SIMD=offforces the baseline
kernels, for comparison. <video>sources with B-frames are rejected when the composition is parsed, with a re-encode
command, instead of failing during rendering: the bundled OpenH264 decoder cannot decode them (the
contract used to say B-frames were supported).- The encoder no longer pretends to use a
24..32quantizer range: with rate control off OpenH264 always
encoded at quantizer 26, which is now the explicitstandarddefault, so default exports are unchanged. - AAC encoding is about 12 times faster (60 s of stereo audio: 16.7 s down to 1.4 s on one core) with
byte-identical output: the vendoredoxideav-aaccomputes its forward MDCT with an FFT. - Gradient, shadow-mask, glyph, and small translucent fills avoid soft-float
roundand libmhypotcalls in their
per-pixel loops, cutting drawing instructions by about 11% on the promo examples with identical pixels.
v0.1.0
First stable release.
Added
- npm distribution as
@aether-motion/cli(launcher plus per-platform binary packages, no install scripts), published from GitHub Actions with trusted publishing; seedocs/releasing.md. When started through npm,initwrites annpx-based MCP entry if the binary is not onPATH. init,preview,add,agent setup|status, anddoctorcommands;renderandchecknow work without
arguments inside a project (aether.json).initinstalls the bundled skills and MCP config for coding agents.- Native renderer for a documented subset of HyperFrames HTML: PNG frames and H.264 MP4 output without a
browser or Node.js (OpenH264 built in, FFmpeg optional with--encoder ffmpeg). - Layout and paint: flex layout, margins, text wrapping,
border,border-radius, linear and radial
gradients,box-shadow,text-shadow,overflow: hidden,transform,filter: blur(),
backdrop-filter: blur(), vignette and film grain. - Text uses an embedded Roboto (Apache 2.0) for the generic families (
sans-serif,serif,
monospace) and as the final fallback, so text renders identically on every machine. Named families
still resolve against system fonts. - Motion:
data-hf-keyframes,data-hf-camera,data-hf-split,data-hf-reveal,data-hf-counter,
data-hf-transition, fades, and CSS@keyframescompiled to native tracks. - Audio.
<audio src>clips withdata-start,data-duration,data-volume,data-hf-fade-in,
data-hf-fade-out,data-hf-trim-start, anddata-hf-loop. Sources can be WAV, FLAC, Ogg Vorbis, or MP3
(mono or stereo, any sample rate). Clips are mixed onto one 48 kHz stereo timeline with whole-sample
timing, encoded to AAC-LC with the pure-Rustoxideav-aac, and muxed as a second MP4 track whose edit list
removes the encoder delay, so audio starts exactly on the first video frame. Audio needs the native
encoder;--encoder ffmpegreturns an explicit error for compositions that use it. vendor/oxideav-aac:oxideav-aac0.1.7 with a table-driven forward MDCT (about 5x faster AAC encoding).- Ready-made blocks and themes (
block list,block <name>). - CLI (
check,inspect,render,bench,block) and an MCP server over stdio (aether_guide,
aether_write,aether_block,aether_check,aether_snapshot,aether_render). The server speaks MCP
protocol2026-07-28(withserver/discover, per-request metadata, and modern result fields) and accepts
legacy clients that begin withinitialize. --audio-bitrate(CLI) andaudioBitrate(aether_render) in kbit/s;checkreports the audio clips,
the peak level, and clipped samples;inspectlists audio clips.- 1080p and 4K export presets, multi-threaded rendering with a low-priority default.
- Tile cache, reduced-resolution wide blurs, and identical-frame reuse; see
BENCHMARK.md. examples/audio.html.- Static SVG. Local
<img src="file.svg">and inline<svg>are rasterized deterministically withresvg
at the target size (sharp at any--scale;object-fit, fades, keyframes, and transitions apply). Shapes,
paths, strokes, gradients, clip paths, masks, and filters render;<text>, SMIL animation,
<foreignObject>,<script>, event handlers, and external references are rejected or ignored.check
validates SVG files and inline markup.examples/svg.html. - Clearer errors for unsupported scripts and SVG content: the
fixhint points to the nativedata-hf-*
motion and CSS@keyframesequivalents.