Performance vs zxing-js

[bisubus]

I was excited to learn that WASM alternative to zxing-js exists. Was the performance measured against zxing-js for the supported codes? I'm curious if WASM extra footprint pays off

[anntnzrb]

No official head-to-head benchmark published, but here's what we know:

The underlying engines are different:

• zxing-js is a JS port of the original Java ZXing
• zxing-wasm wraps zxing-cpp, which has "many improvements in terms of runtime and detection performance" compared to the original Java library

So you're comparing two different codebases, not just JS vs WASM.

WASM vs JS tradeoffs:

• WASM generally wins on CPU-intensive tasks (image processing, decoding) — runs closer to native speed
• JS has less overhead for small ops and avoids WASM instantiation cost
• The ~919KB reader-only footprint is mostly one-time load cost

Practical take:
For barcode scanning, the bottleneck is usually decoding algorithm complexity. zxing-cpp's optimized C++ compiled to WASM should outperform a JS port of the older Java code — but real-world perf depends on your use case (barcode types, image quality, batch size).

If you need hard numbers, you'd have to benchmark both with your actual images. Would be a great contribution if someone ran one!