---
title: Frequency-Adaptive Non-uniform Compression (FANC)
url: https://www.emergentmind.com/topics/frequency-adaptive-non-uniform-compression-fanc
type: topic
---

# Frequency-Adaptive Non-uniform Compression (FANC)

Frequency-Adaptive Non-uniform Compression (FANC) encompasses a set of algorithmic strategies and neural network modules that adaptively allocate coding resources across frequency domains, spatial regions, or temporal positions within high-dimensional data. FANC leverages explicit or learned decompositions—typically in the Fourier, DCT, or wavelet bases—and provides non-uniform, content-adaptive bit allocation that maximizes efficiency for sources exhibiting spectral or spatial inhomogeneity. By exploiting both the statistical and perceptual variability across frequency components, FANC achieves substantial improvements in rate-distortion trade-offs, architectural efficiency, and computational/memory savings across diverse domains including image compression, speech enhancement, large language model (LLM) KV-cache compression, and computational wave physics.

## 1. Core Principles and Theoretical Foundations

FANC is unified by three primary algorithmic principles:

1. **Explicit Frequency Decomposition:** 
   FANC systems utilize spectral transforms (DCT, DWT, FFT) or learned proxies (e.g., frequency masks from error-variance maps) to decompose data into interpretable frequency bands or subbands [2511.20151][2301.01290][2112.06417].

2. **Adaptive Bit Allocation:** 
   Allocation of coding or modeling resources is performed non-uniformly across frequency components, spatial regions, or time—driven by learnable weights, ablation-derived masks, or statistical heuristics. Content-adaptive strategies are central, enabling robust handling of local signal complexity and information density [2511.20151][2507.20030][2602.00568].

3. **Band- or Mask-driven Encoding/Decoding:**
   Bits are spent on frequencies/regions according to importance, as inferred from learned error maps, DCT-magnitude statistics, message-passing marginals (LDGM codes), or information-theoretic ablation studies [1004.2628][2507.20030][2112.06417].

These principles enable FANC frameworks to efficiently represent both low- and high-frequency content, realizing performance unattainable by uniform, non-adaptive coding.

## 2. Algorithmic Realizations Across Domains

### a) Learned Image Compression

In neural image codecs, FANC is instantiated by modules such as the Adaptive Frequency Decomposition (AFD) [2112.06417], which predicts an error-variance map $\sigma(x)$ per pixel and channel:

\[
(\hat{x},\;\sigma) = \mathrm{AFD}(x,z)
\]
\[
m_L(i) =
\begin{cases}
1,& \sigma(i)\le\tau \\
0,& \sigma(i)>\tau
\end{cases}
\]

Where $m_L$ and $m_H = 1 - m_L$ define low/high-frequency masks; $\tau$ is an image- and channel-adaptive threshold. Compression proceeds in a coarse-to-fine pipeline, first encoding low-frequency pixels with a dedicated network, then encoding high-frequency pixels conditioned on the decoded low-frequency content [2112.06417].

Advanced architectures such as HCFSSNet further embed FANC by local DCT decomposition and Adaptive Frequency Modulation Modules (AFMM), learning per-frequency weights $W_f(u,v)$ via CNNs, modulating latent representations before entropy coding. This facilitates direct learning of optimal bit allocation schemes:

\[
\widetilde X(u,v) = W_f(u,v) \cdot X(u,v)
\]

[2511.20151] demonstrates bidirectional fusion of spatial (state-space/VONSS) and frequency (AFMM/DCT) cues, enhancing both long-range and fine-detail compression.

### b) Speech and Audio Spectrograms

FANC encoders in speech enhancement, as in DVPD [2602.00568], partition the spectrogram along frequency bands, applying non-uniform compression:

- Low ([0, $F_\ell$], e.g., 0–2 kHz): no downsampling, preserving harmonics.
- Mid ($F_\ell < f \leq F_m$, e.g., 2–4 kHz): moderate compression (stride=2, medium dilation).
- High ($F_m < f \leq F$, e.g., >4 kHz): aggressive compression (stride=4, large dilation).

The mapping is:
\[
C(f,t) = B_\ell(f)\cdot \mathrm{Conv}_{3,1,1}(S) + B_m(f)\cdot \mathrm{Conv}_{3,3,2}(S) + B_h(f)\cdot \mathrm{Conv}_{3,5,4}(S)
\]

Such band-splitting aligns with psychoacoustic relevance and empirical sparsity, yielding both efficiency and preservation of perceptual quality [2602.00568].

### c) LLM KV Cache Compression

In KV cache applications, FAEDKV [2507.20030] employs an Infinite-Window Discrete Fourier Transform (IWDFT) to transform the cache into the frequency domain. Layer-specific spectral band selection is performed via ablation studies:

\[
K^{f,t+1,\ell}[k]=W_k\left(\frac{N-1}{N}K^{f,t,\ell}[k]+\frac{1}{N}\,k_{t-R}^\ell\right)
\]

Masking is applied to prune uninformative bins (mask $M^{(\ell)}(k)$ derived from ablation scores), achieving up to 4× prefill speedups and maintaining accuracy at up to 10× compression [2507.20030].

### d) Wave Physics Inverse Problems

For full-waveform inversion, FANC combines frequency-adaptive grid selection with cascaded lossy compression (temporal/spatial downsampling, thresholding in spatial, wavelet, or wave-atom domains). Discretization is frequency-band scheduled, with each band modeled at spatial/temporal resolutions just sufficient for its spectral content, and error-controlled compression ensuring the optimization remains stable [2111.04700].

## 3. Rate-Distortion, Training, and Optimization

The FANC training regime typically minimizes a rate-distortion Lagrangian:

\[
\mathcal{L} = R + \lambda D
\]

with $R$ the estimated code length (cross-entropy or entropy bottlenecks), $D$ a distortion function (MSE, MS-SSIM), and $\lambda$ the rate-distortion tradeoff parameter. For coarse-to-fine or multi-band FANC, separate mask-weighted reconstruction and rate terms are assigned per frequency or region [2112.06417][2301.01290].

For LDGM-based lossy FANC, a belief-propagation plus decimation encoder achieves near-theoretic efficiency on non-uniform Bernoulli sources by carefully matching quantizer parameters and graph degrees to the target bias and distortion [1004.2628].

Optimization for complexity is domain-driven: convolutional FANC reduces parameters/MACs in speech models [2602.00568]; FAEDKV achieves $O(N\,\log N)$ prefill and $O(r\,N\,\log N)$ per-step reconstruction [2507.20030]; FWI compression yields 30% runtime reduction and $10^3$–$10^4\times$ memory savings with $<0.2\%$ error penalty [2111.04700].

## 4. Empirical Performance and Ablation Evidence

Extensive experiments across domains document the efficiency of FANC:

- **Lossless image compression**: State-of-the-art on DIV2K, CLIC.p/m, with FANC outperforming prior learned and hand-engineered codecs by 1–7% in bpp. Dense ablations show losses of up to 11% in high-frequency bpp when the coarse-to-fine scheme is ablated [2112.06417].
- **Learned lossy image compression with scalability**: Overhead for quality-scalable FANC is just 1–2%, much lower than classical approaches (15–25%), and supports region-of-interest (ROI) enhancement at minimal incremental cost [2301.01290].
- **KV cache compression in LLMs**: FAEDKV surpasses token eviction and learned vector projection approaches by up to 22% on LongBench, and maintains flat retrieval accuracy across position on Needle-In-A-Haystack, unlike convolutional approaches [2507.20030].
- **Speech enhancement**: FANC enables extreme architectural efficiency (1.9M params, 10.2G MACs) with PESQ improvements over reference models at under half the parameter count [2602.00568]. Ablations confirm 0.06–0.07 absolute metric drops when FANC is disabled.
- **Full-waveform inversion**: Consistent ~30% runtime reduction and $10^3$–$10^4\times$ memory savings; compressed gradients yield $\langle \theta \rangle < 40^\circ$ and admit stable optimization [2111.04700].

## 5. Practical Deployment and Extensibility

FANC’s modularity allows instantiations in:

- Neural codecs (images, video, hyperspectral): via learned frequency masks or transform-domain residuals [2112.06417][2511.20151].
- LLM memory systems: frequency mask adaptation, ablation-driven retention, and finite/online masking [2507.20030].
- Physics solvers: frequency-adaptive gridding and flexible sparse-domain thresholding [2111.04700].
- Scalable/ROI-enabled inference: partitioning enhancements and supporting selective refinement [2301.01290].
- Speech/audio: band-partitioned convolutional encoders exploiting information density [2602.00568].

Potential extensions include joint spatial-temporal FANC for video, learned adapters for online frequency mask tuning, direct application to attention map compression, or integration with quantization [2507.20030][2112.06417].

Limitations include the need to transmit additional mask/threshold data (e.g., $\tau$ per subimage), and in some cases issues with differentiable rate control due to hard masking [2112.06417].

## 6. Comparative Overview and Representative Methods

| Domain           | Key FANC Mechanism                 | Experimental Gains                                               |
|------------------|------------------------------------|-----------------------------------------------------------------|
| Image Compression| Error-variance maps, DCT/AFMM      | 1–7% bpp savings, 11% high-freq bpp cut, 18–25% BD-rate on SOTA [2112.06417][2511.20151] |
| Speech Enhancement| Tri-band convolutional partition   | 40–60% MACs reduction, 0.06–0.07 metric gain                   [2602.00568]            |
| KV-Cache (LLM)   | IWDFT, frequency ablation masks    | 4× speedup, 22% accuracy gain over token eviction, flat retrieval position profile [2507.20030] |
| Wave Physics     | Frequency-adaptive gridding, wavelet/atom thresholding | 30% runtime cut, 3–4 orders memory win, $<0.2\%$ final error increase   [2111.04700]            |

FANC approaches thus provide domain-specific, quantitatively validated improvements over uniform coding baselines, with ablations attributing the bulk of performance to adaptive, fine-grained frequency allocation.

## 7. Future Directions

Ongoing research aims to generalize FANC by:

- Extending adaptive masking to spatio-temporal domains for video and dynamic data [2112.06417].
- Coupling non-uniform frequency allocation with quantization and vector-coded mask signaling [2112.06417][2507.20030].
- Online or jointly trained frequency mask modules for LLMs and neural feature compressors [2507.20030].
- Applying FANC concepts to self-attention map, hidden state, or intermediate feature compression well beyond input and output data representations [2507.20030][2511.20151].

A plausible implication is that FANC-based frameworks may become the unifying abstraction for efficient resource allocation in high-dimensional, structured, and evolving signal representations across diverse computational domains.

Source: https://www.emergentmind.com/topics/frequency-adaptive-non-uniform-compression-fanc