AI-generated digital cinema mastering room with a reference display and projection screen showing detailed coastal imagery

JPEG2000 FPGA IP Core


JPEG 2000 (ISO/IEC 15444) is an international wavelet-based image compression standard supporting lossy and lossless coding, with a focus on image quality and compression efficiency.

BergLogic JPEG2000 IP Core is an RTL (Verilog) codec for FPGA platforms. Both Standard and Low-Latency variants process images and video, with configuration and customization guided by format interoperability, latency and the target platform.

For Applications Where Image Detail Matters

JPEG2000 suits applications that value image detail, transport bandwidth and storage efficiency. Formats and compression settings are selected for the product requirements.

AI-generated empty digital cinema auditorium with mountain and lake imagery on the projection screen

Digital Cinema

High-quality image production, storage and playback.

AI-generated medical image review workstation and scanning equipment without patient information

Medical Imaging

Image compression, storage and transmission for medical imaging workflows.

AI-generated generic Earth observation satellite above detailed land and ocean imagery

Satellite Imagery

Processing and transmission of high-resolution remote sensing images.

Why Choose JPEG2000?

JPEG2000 is worth evaluating when an application needs both reduced data volume and high image fidelity. At equal compression ratios it can typically deliver higher quality than conventional JPEG and lightweight intra-frame codecs such as JPEG XS, especially on detail-rich natural images. Results depend on content and encoder settings.

Medical and satellite imaging often place a high value on fine texture and tonal gradation. JPEG2000 is more complex and computationally demanding, requiring more FPGA logic and memory. When fidelity is the priority, that additional implementation effort can be justified by the resulting image quality.

Preview of test_pic1: autumn foliage against a blue sky, with dense leaf and branch textures and fine color detail
Test image preview: test_pic1 (autumn foliage)
PSNR curves at compression ratios from 3 to 25 for autumn foliage, comparing JPEG2000 9/7 and 5/3 with JPEG XS MAIN and HIGH configurations
PSNR at equal compression ratios: test_pic1 (autumn foliage)

For the autumn foliage test image, JPEG2000 9/7 and 5/3 achieve higher PSNR than JPEG XS MAIN/HIGH from 3:1 to 25:1. JPEG2000 uses five wavelet decomposition levels. See the full comparison for other test images, settings and exceptions.

Compare JPEG XS and JPEG2000: Selection Guide and Measured PSNR ↗

Core JPEG2000 Compression Characteristics

High Image Quality and Fine Detail

Emphasizes fine textures, edges, color and tonal gradation for fidelity-focused applications. Actual quality depends on image content and the encoding operating point.

Full-Frame Rate-Distortion Optimization

Typical JPEG2000 encoders use post-compression rate-distortion optimization (PCRD-opt). Rate-distortion slopes at codeblock (CB) truncation points—the distortion reduction per additional bit—guide bit allocation across the full frame to reduce overall distortion at the target bitrate.

Lossy and Mathematically Lossless

Lossy coding trades image quality against data volume. Mathematically lossless coding reconstructs every pixel exactly for applications requiring complete image information.

Wavelet Coding, Less Blocking

Unlike the small-block DCT transform used in conventional JPEG, JPEG2000 uses wavelet transforms to help reduce familiar blocking artifacts at high compression, producing more natural texture and tonal transitions.

BergLogic IP Advantages

  • Lean Resources, Lower Deployment Cost Datapaths are optimized around JPEG2000’s computational needs to reduce unnecessary logic, memory traffic and power overhead while meeting quality and processing targets.
  • Multiple Configurations, Custom Performance The baseline design targets 4K60, with multiple configurations available. Throughput, parallelism and resources can be tailored to resolution, frame rate, bit depth and the target platform, with application-specific optimization within the selected standard or variant requirements.
  • Proven in Commercial Deployments Fully validated and deployed in volume on AMD Xilinx Kintex-7, UltraScale and UltraScale+ platforms, with a stable commercial track record.
  • Multiple FPGA Platforms Deployable on AMD Xilinx, Altera, Pango and other FPGAs, with adaptation to the target device’s resources, interfaces and performance requirements.
  • Responsive Engineering Support Direct engineering contact supports fast responses for integration and debugging, including issue investigation, platform adaptation, application optimization and technical consultation.

JPEG2000 IP: Standard and Low-Latency Variants

Both variants process images and video. Their strengths differ in application fit, formats and latency requirements.

Standard: Formats and Interoperability

Follows JPEG2000 Part 1 Amd 1 with lossy 9/7 and mathematically lossless 5/3 wavelet coding, plus .jp2 and .j2c formats. Particularly suited to individual-image processing, archiving and standard-format workflows.

Two-way encode/decode interoperability with the official reference software eases integration into existing imaging workflows.

Low-Latency: Real-Time Pipeline Optimization

Uses lossy 9/7 wavelet coding with a focus on low-latency video. Truncation points can be predicted and calculated after a few image lines arrive, reducing processing wait time.

A BergLogic-proprietary container and the corresponding encoder/decoder pair reduce container overhead and enable fast decoding.

The Low-Latency variant can keep end-to-end latency below 8.5ms, with selected optimized configurations reaching below 2ms. This measures HDMI input to HDMI output, including JPEG2000 encoding, network transport and decoding. Actual latency depends on video format, system configuration and network conditions.

Technical Specifications

Compression ratio, wavelet levels, quantization steps and codeblock sizes are configurable. The table describes the baseline 4K60 design; other performance configurations and custom specifications depend on the application, selected variant and target FPGA platform.

Standard Low-Latency
Baseline Design Performance 4K@60fps 4K@60fps
Latency (end-to-end) 1 frame + 2ms 8.5ms
Color space RGB / YUV444 / YUV422 / Grayscale (PGM) RGB / YUV444 / YUV422
Bit depth 8 / 10 / 12 8 / 10 / 12
Wavelet (DWT) type 9/7 wavelet (lossy), 5/3 wavelet (lossless) 9/7 wavelet (lossy)
Wavelet (DWT) levels 3 to 6 3 to 6
Quantization step Configurable Configurable
Quality layers 1 1
Lossy compression ratio 1× to 25×, dynamically configurable 1× to 25×, dynamically configurable
Encoder input format AXI-Stream AXI-Stream
Encoder output format .jp2, .j2c BergLogic proprietary format
Decoder input format .jp2, .j2c BergLogic proprietary format
Decoder output format AXI-Stream AXI-Stream
Codeblock (CB) size 32×32, 64×64, 64×4 64×4, 16×16, 32×8
Codestream order CPRL PCRL

The actual compression ratio in mathematically lossless mode depends on image content.

Demonstration Platform and Integration

A hardware demonstration platform based on AMD Xilinx KU3P FPGA is available to observe bitrate, image quality and latency in practice. Combined with HDMI/SDI I/O, VPSS and network transport modules, the codec can support distributed video systems and other imaging products.

BergLogic JPEG 2000 IP
JPEG2000 codec demo environment

Contact us for configuration advice, customization and integration support for your target FPGA, video format and latency requirements.

Discuss Your Integration ↗
© Copyright - Berglogic, Limited      BergLogic: Building 9, Floor 7, Shenzhen Bay Tech Eco-Park, Shenzhen            Phone: +86 755-26928530 粤ICP备2021019798号