Sinisistar 2 -v0.2.0.4- -nennai 5- -

Before diving into the new update, make sure your system meets the minimum requirements to ensure smooth gameplay.

| Node | Category | Functionality | Performance | |------|----------|---------------|-------------| | GranularSynth | Audio | Real‑time granular synthesis with envelope‑shaping and pitch‑modulation. | 0.8 ms per 512‑sample block (single‑core). | | SpectralGate | Audio | Frequency‑domain gating using adaptive thresholding. | GPU‑offload optional via Vulkan‑Compute. | | LFO‑Matrix | Modulation | 16‑channel LFO matrix with inter‑LFO routing, sync to BPM or external clock. | Negligible CPU (< 0.1 ms). | | Video‑FFT | Visual | Real‑time FFT of video frames, outputs spectrum texture for downstream shaders. | 60 fps at 1080p on RTX 3080. | | OSC‑Bridge | IO | Bi‑directional OSC with per‑node address mapping; supports /bundle and typed arguments. | No measurable overhead. | | MIDI‑PolySynth | Audio | Polyphonic subtractive synth with 32‑voice voice allocation and per‑voice filter. | 1.2 ms per 256‑sample block (CPU). |

| Task | SiNiSistar 2 + Nennai 5 (LOC) | EdgePulse (LOC) | Avg. Time (hrs) | |------|-------------------------------|-----------------|-----------------| | Add MobileNet‑V2 model | 42 | 78 | 1.5 | | Create custom DSP‑NN hybrid node | 28 | N/A (requires external glue) | 1.0 | SiNiSistar 2 -v0.2.0.4- -Nennai 5-

The concise plugin API reduces development effort considerably.

| Module | Function | Implementation Highlights | |--------|----------|----------------------------| | NN Runtime | Tiny inference engine for ONNX‑Lite and custom binary formats. | Fixed‑point quantisation (8‑bit), operator fusion, ARM‑NEON / RISC‑V vector intrinsics | | Model Loader | Parses and validates model graphs at start‑up. | Supports static graph optimisation (constant folding, dead‑node removal) | | Accelerated Ops | Optimised kernels for convolution, depth‑wise separable conv, and fully‑connected layers. | Winograd algorithm for 3×3 conv, cache‑aware tiling | | Pipeline Integration | Provides NNNode that can be inserted into SiNiSistar pipelines like any other DSP node. | Automatic tensor shape propagation, zero‑copy buffers between DSP and NN layers | | Edge‑AI Utilities | On‑device model management, versioning, and OTA update handling. | Secure hash verification, rollback support | Before diving into the new update, make sure

Nennai 5 is deliberately lightweight (≈ 200 KB binary) and deterministic, making it suitable for hard‑real‑time constraints (≤ 5 ms worst‑case latency for typical inference tasks).

| Scenario | SiNiSistar 2 + Nennai 5 (ms) | SignalForge 1.3 + External AI (ms) | Δ (%) | |----------|------------------------------|-----------------------------------|-------| | Audio Denoising → Keyword Spotting | 18.2 ± 0.5 | 31.7 ± 0.8 | –42.5 | | IMU/LIDAR Fusion → Obstacle Classification | 27.9 ± 0.6 | 46.5 ± 1.0 | –40.0 | | Module | Function | Implementation Highlights |

The integrated pipeline eliminates inter‑process communication overhead and benefits from shared memory pools.

| Roadmap Item | Expected Release | Description | |--------------|------------------|-------------| | v0.3.0 – Distributed Pipeline | Q4 2026 | Enable pipelines that span multiple devices via gRPC, with automatic load‑balancing. | | GPU Offload Layer | Q2 2027 | Add Vulkan‑Compute backend for kernels that benefit from massive parallelism (e.g., large‑scale FFT). | | Dynamic Quantisation | Q3 2027 | Runtime selection of quantisation parameters based on input statistics, improving accuracy for variable‑precision models. | | Formal Real‑Time Guarantees | Q1 2028 | Integration of a deterministic scheduler (based on Rate‑Monotonic Analysis) and certification‑grade timing analysis tools. | | Python Bindings | Q2 2026 | High‑level API for rapid prototyping, while preserving zero‑copy through Pybind11 buffers. |

Released as a development preview for supporters, v0.2.0.4 is not just a bug-fix patch. It is a content drop that reframes the mid-game experience. Here is what the version number entails: