Chips&Media

Chips&Media, Inc. is a South Korean semiconductor intellectual-property company whose technology is integrated into system-on-chip designs rather than sold primarily as finished consumer hardware. Its principal business is the development and licensing of reusable semiconductor IP cores for video encoding and decoding, image signal processing, and computer vision. These building blocks allow chip designers to add multimedia and vision capabilities to products such as mobile devices, digital cameras, home multimedia equipment, video-conferencing systems, surveillance products, automotive systems, and other embedded or edge-computing devices. The company has developed video technology since 2003. Its codec portfolio has covered a broad succession of industry and regional standards, including MPEG-2, MPEG-4, H.263, H.264/AVC, VC-1, RealVideo, AVS, MVC, VP8, AVS2, HEVC/H.265, and VP9. The company’s IP has been licensed by semiconductor and platform businesses including Freescale, VIA, Realtek, Novatek Innofidei, and Telechips. This licensing model positions Chips&Media as an enabling technology supplier: its customers incorporate the cores into their own chips, while Chips&Media focuses on architecture, implementation, standard support, optimization, and related technical delivery. Its video portfolio includes the WAVE, CODA, and BODA families. WAVE is designed for multi-standard ultra-high-definition video processing, supporting 4K video at up to 60 frames per second and combining encoding and decoding support for standards such as HEVC and H.264 with decoding support for additional formats. CODA is a more established multi-channel video engine aimed at applications requiring simultaneous processing of multiple channels or formats, generally at resolutions up to 2K. BODA focuses on video decoding across a broad range of legacy and contemporary formats, including Full HD VP8 decoding at up to 60 frames per second. Chips&Media expanded beyond codec IP during the second half of the 2010s. It announced image-processing products in 2016 and 2017, and reported a first image-processing license agreement in March 2018. Its image-signal-processing portfolio includes configurable pipelines such as CARPO, LEDA, and METIS, with different intended image resolutions. These can be complemented by specialized functions including multi-exposure high-dynamic-range processing through NIX, motion-adaptive three-dimensional noise reduction through HYDRA, and wide-angle lens-distortion correction through KERBEROS. The company also added object-detection IP in 2018. Its c.WAVE100 computer-vision core is intended to identify objects such as vehicles, people, and cyclists in real time from 4K video input at up to 30 frames per second. The design uses a dedicated hardwired architecture for convolutional-neural-network processing, with an emphasis on reducing memory-bandwidth requirements. That architecture is suited to computationally demanding edge applications, including automotive and surveillance systems. Chips&Media’s positioning is therefore based on specialized, licensable semiconductor technology for multimedia and visual intelligence. Its growth path, as described in available reference material, has moved from broad video-codec support toward a wider portfolio covering image quality, camera processing, and machine-vision workloads. Publicly available material does not establish the company’s founder, headquarters, ownership structure, current executives, financial performance, or detailed corporate history beyond its technology-development timeline.

Chips&Media is a South Korean semiconductor IP company focused on technology that can be embedded in system-on-chip products. Available reference material dates the beginning of its video-technology development to 2003. Rather than manufacturing a complete range of branded consumer electronics, the company develops licensable hardware designs and related technology for semiconductor customers. Its initial and most established area was video codec processing. Over time, the company built support for a wide set of video standards, ranging from MPEG-2 and MPEG-4 to H.264/AVC, VC-1, VP8, HEVC/H.265, and VP9, alongside other formats such as H.263, RealVideo, AVS, AVS2, and MVC. This breadth enabled its cores to serve different generations of multimedia equipment and different market requirements. Reference material identifies licensees or users of its technology including Freescale, VIA, Realtek, Novatek Innofidei, and Telechips. The video portfolio developed into several product families. BODA provided broad-format decoding capabilities for resolutions from D1 through Full HD, including a hardware implementation capable of decoding Full HD VP8 streams at 60 frames per second. CODA was positioned as a multi-standard video engine supporting simultaneous channels or formats, with applications in mobile handsets, digital camcorders, home multimedia devices, conferencing, and surveillance. WAVE represented a higher-resolution direction, targeting ultra-high-definition 4K processing at up to 60 frames per second. Its published capabilities included HEVC and H.264 encoding and decoding, together with decoding for VP9, AV1, and other formats described in the product material. During 2016 and 2017, Chips&Media expanded into image processing. The company introduced image signal processing products intended to provide configurable processing pipelines at different resolution levels. CARPO, LEDA, and METIS were differentiated by their intended image-resolution ranges, while additional modules addressed particular camera-processing tasks. NIX supported multi-exposure HDR through fusion of two exposures. HYDRA used multiple image frames and motion-adaptive processing for three-dimensional noise reduction. KERBEROS addressed distortion caused by wide-angle lenses. The company reportedly signed its first image-processing license agreement in March 2018, marking a commercial step in the extension of its licensing business beyond video codecs. In the same year, it added object-detection IP. The c.WAVE100 design was described as a computer-vision core able to detect objects such as cars, people, and cyclists from 4K video input at up to 30 frames per second. Its use of a dedicated hardwired convolutional-neural-network architecture was intended to limit memory-bandwidth requirements, making it relevant to edge devices in automotive and surveillance settings. The available material does not provide a reliable account of the company’s founder, headquarters, ownership, financial results, executive leadership, or subsequent corporate transactions. Its documented strategic direction is nevertheless clear: Chips&Media evolved from a video-codec IP specialist into a broader supplier of licensable technologies for video, camera imaging, and embedded visual intelligence.

• 2018

  First reported image-processing license

  [1]Chips&Mediaen.wikipedia.org

  The company reportedly signed its first image-processing license agreement in March.

• 2018

  Object-detection IP introduced

  [1]Chips&Mediaen.wikipedia.org

  Chips&Media added object-detection IP to its product lineup.

• 2017

  Image-processing portfolio expanded

  [1]Chips&Mediaen.wikipedia.org

  The company continued announcing image-processing technology and products.

• 2016

  Image-processing products announced

  [1]Chips&Mediaen.wikipedia.org

  Chips&Media announced image-processing products, extending its scope beyond video codec cores.

• 2003

  Video-technology development begins

  [1]Chips&Mediaen.wikipedia.org

  Reference material identifies 2003 as the beginning of Chips&Media’s video-technology development.

A specialist licensor of reusable semiconductor IP for video processing, imaging, and edge computer vision.

• Video codec IP

  WAVE

  WAVE is a multi-standard video encoder and decoder family aimed at ultra-high-definition applications. It is designed for 4K processing at up to 60 frames per second. The published capability set includes HEVC/H.265 and H.264/AVC encoding and decoding, with decoding support for additional formats such as VP9 and AV1. It is positioned for SoC designs requiring high-resolution video capability.

• Video codec IP

  CODA

  CODA is a multi-standard video engine supporting encoding and decoding, including simultaneous processing of multiple channels or formats. It is intended for applications such as mobile handsets, digital camcorders, home multimedia equipment, video conferencing, and surveillance. Compared with WAVE, the reference material describes CODA as a lower-resolution family, supporting up to 2K at as much as 60 frames per second and not supporting HEVC encoding and decoding.

• Video decoder IP

  BODA

  BODA is a video-decoding family supporting formats including H.264/AVC, VC-1, MVC, VP8, H.263, MPEG-4, MPEG-2, RealVideo, and AVS. Its stated operating range extends from D1 through Full HD. A notable capability described for the family is fully hardware-based VP8 decoding for Full HD streams at up to 60 frames per second.

• Image signal processing IP

  CARPO, LEDA, and METIS

  CARPO, LEDA, and METIS are configurable image-signal-processing pipelines differentiated by intended image resolution. CARPO is described for resolutions up to 2 megapixels, LEDA for up to 5 megapixels, and METIS for resolutions above 5 megapixels. The pipelines can be combined with specialized modules such as HDR processing, noise reduction, and lens-distortion correction.

• HDR image-processing IP

  NIX

  NIX is a multi-exposure high-dynamic-range image-processing design. It fuses two exposures using a line-interleaved method, allowing it to serve as an additional processing function alongside Chips&Media’s basic image-processing pipelines.

• Image noise-reduction IP

  HYDRA

  HYDRA is a three-dimensional noise-reduction system that uses information from multiple image frames to suppress image noise. Its motion-adaptive approach is intended to reduce artifacts when scene content or camera subjects move between frames.

• Lens-distortion correction IP

  KERBEROS

  KERBEROS is a lens-distortion correction method designed to compensate for the geometric distortion introduced by wide-angle lenses. It can be used as a complementary function within an image signal processing pipeline.

• Computer-vision IP · 2018

  c.WAVE100

  c.WAVE100 is an object-detection IP core for computer-vision applications. It is described as capable of processing 4K video input at up to 30 frames per second and detecting objects such as cars, people, and cyclists. The design uses a fully hardwired, dedicated convolutional-neural-network architecture intended to reduce memory-bandwidth requirements. Its target applications include automotive and surveillance edge devices.

• 2018

  Add object-detection IP to the product lineupProduct launch

  Background

  The company’s image-processing expansion created a basis for addressing broader computer-vision workloads.

  Action taken

  Chips&Media introduced object-detection IP, including the c.WAVE100 core for real-time 4K edge-video analysis.

  Aftermath

  The product direction extended the company’s relevance to automotive and surveillance applications, although detailed commercial results are not provided.

  Source

• 2016

  Expand from video codecs into image processingStrategy

  Background

  Chips&Media had established a portfolio of video codec IP and began broadening its technology scope into image signal processing.

  Action taken

  The company announced image-processing products in 2016 and continued the expansion in 2017.

  Aftermath

  The expansion was followed by a reported first image-processing license agreement in March 2018.

  Source

• 2026.01.05

  Chips&Media and Visionary.ai launch AI-based full Image Signal Processor[2]Chips&Media与Visionary.ai推出基于AI的全图像信号处理器design-reuse.com

  Chips&Media and Visionary.ai announced the launch of the world's first AI-based full Image Signal Processor (ISP), a software-defined imaging pipeline that replaces traditional hardware-based pipelines with neural networks and can be updated over the air. The technology will debut at CES 2026 and be available for OEM evaluation shortly thereafter.

Read more

1. 1Chips&Mediaen.wikipedia.org
2. 2Chips&Media与Visionary.ai推出基于AI的全图像信号处理器design-reuse.com