TurboLERC is one of eight drop-in codecs in the Bitruvius Imagery & 3D Suite — keep your wire format, swap the backend.
Explore the SuiteSame wire format. Much faster.
A SIMD-optimized LERC2 codec, byte-for-byte compatible with the LERC reference. Drop it in where you already use LERC and get 2–4× faster encode, 1.4–1.9× faster decode, with the exact same compressed output.
Performance
Speed at parity.
Compression ratio matches the LERC reference to 100%. Throughput is where turbolerc moves the needle. All numbers are medians across the matched corpus.
median across all domains
peak 0.0× on multispectral
median across all domains
peak 0.0×
compression match to reference LERC
decodes legacy v1 (CntZImage) too
Prove It
Race it yourself. Right here.
TurboLERC and the Esri LERC reference decode the same bytes in your browser — our WebAssembly against theirs, clocked live. Pick a raster type and hit START.
Cross-Domain
Faster across the board.
The same lossless data, run through the LERC reference and through TurboLERC, on the same hardware. Median speedup per domain across every sample we tested.
Solid bar = median across the matched corpus.Faded bar = 90th percentile (clipped to chart bounds when extreme outliers stretch beyond).
Solid bar = median across the matched corpus.Faded bar = 90th percentile (clipped to chart bounds when extreme outliers stretch beyond).
Codec Landscape
100% compression parity.
Materially faster.
Same lossless data, ten codecs, on the same hardware. TurboLERC and LERC produce byte-identical output — same compression bar — and TurboLERC dominates the speed bars. The generic compressors (Zstd, LZ4, Brotli, Deflate) hit a different point on the tradeoff.
Solid bar = median across the matched corpus.
TurboLERC and LERC produce identical compressed bytes. Same ratio — drop-in replacement.
Solid bar = median across the matched corpus.
Solid bar = median across the matched corpus.
LZ4 wins on raw speed but loses on compression. Brotli and Deflate hit similar ratios to LERC but encode an order of magnitude slower. TurboLERC keeps LERC's compression while moving both encode and decode bars meaningfully to the right.
Why It's Faster
Engineered.
Not configured.
The LERC2 wire format dates from 2010. The reference implementation reflects the constraints of that era — single-issue scalar code, branch-heavy inner loops, generic fallbacks where a few well-chosen specializations would have made an order-of-magnitude difference.
TurboLERC keeps the wire format and rebuilds the codec around how modern CPUs actually execute work. Vectorized inner loops, branchless hot paths, specialization for the cases that matter, and encoder heuristics that skip work the reference can't avoid. The specific techniques are proprietary and ship only as compiled binaries through the Bitruvius Developer Hub.
The bars above are what comes out of that work. We'd rather show the result than the recipe.
Compatibility
Drops in. Doesn't break a thing.
Wire format
Emits LERC v6 byte-identical to the reference. Decodes both v2 and legacy v1 (CntZImage) transparently — anything any LERC-aware client can parse, TurboLERC can too.
Data types
i8, u8, i16, u16, i32, u32, f32, f64. Single-band, multi-band, nDepth (multiple values per pixel), and masked variants — every LERC API surface is covered.
Pipelines
Slot TurboLERC into Cloud Optimized GeoTIFF, ArcGIS pipelines, GDAL drivers, or your own LERC-aware tooling. Existing readers don't know the difference.
Platforms
Native everywhere.
Even in the browser.
AVX2 on x86_64, NEON on Apple Silicon and Graviton, WASM SIMD128 in the browser running at 60–107% of native throughput, and a portable scalar fallback for any CPU. SVE2 scalable-vector dispatch lands once benchmarks complete in Q3.
Same payload, two backends — elevation (DEM, F32): AVX2 510 MB/s decode, NEON 480 MB/s decode.
Platform matrix
All backends produce bit-exact identical output. Auto-detection at runtime; portable scalar fallback for any CPU.
Access
Free decoder. Paid encoder.
Decoder
Free
Open the Developer Hub, pick your platform, ship with the TurboLERC decoder for free. Decodes any LERC v1 or v2 payload at 1.5× the reference's speed.
Encoder + Decoder
Commercial license
Production encoding via the Developer Hub. Same pricing tiers as RIPT — seats, transactions, OEM redistribution. Compiled binaries, no source.
LERC.
Just faster.
Existing pipelines, existing data, existing tooling. Replace the codec, keep everything else.
Esri, ArcGIS, and LERC are trademarks of Environmental Systems Research Institute, Inc., used here solely to identify the products and formats being compared. No affiliation, sponsorship, or endorsement is implied. LERC is published by Esri under the Apache 2.0 license; the live demo runs Esri's unmodified reference decoder under that license. TurboLERC is an independent, clean-room implementation of the openly published LERC format — compatibility claims, including byte-for-byte output, describe interoperability with that format as verified by our test suites. Performance comparisons reflect our own measurements under the stated methodology; results vary by workload and hardware. Full third-party notices: Attributions.