TL;DR
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
Nibnib is an experimental image format designed for very fast image decompression, particularly on RP2350-class microcontrollers. Its decoder uses a 256-entry palette of color deltas to produce two RGB565 pixels from each encoded byte, while its current encoder is much slower and image quality may improve with further work.
Nibnib, an experimental image format described in a GitHub project report, is designed to decompress images at high speed on RP2350-class microcontrollers. Its decoder processes two RGB565 pixels from each compressed input byte using a palette of color differences, targeting a tight per-pixel processing budget for device displays.
The project began as an experiment in image quality and performance for DVI output on an RP2350. The report says devices in this class often use low-bit-depth images and palette lookups, with an estimated budget of about six processing cycles per pixel. Nibnib is designed to fit that constraint by keeping the decoding operation short and predictable.
Each encoded byte selects one of 256 palette entries. Rather than representing a complete color, the selected entry represents a delta added to the previous pixel, which is duplicated to form a pair. The decoder produces two RGB565 pixels from that operation, giving the data stream a nominal rate of four bits per pixel after the initial pixel values. The report says this delta-based approach is intended to represent smooth image regions better than a conventional palette of absolute colors.
The project includes an RP2350 decoder implementation in the image demo for the vic-2350 repository. The format stores width and height, a 256-entry palette, and image rows; each row starts with two RGB565 pixels and continues with palette-index bytes. Its author describes the format as deliberately simple and says it is not intended for archival use.
Fast Decoding for Microcontroller Displays
Nibnib explores a tradeoff relevant to devices with limited processing capacity: spending more effort when preparing an image so that a small device can render it with little work. If the decoder performs as intended on its target hardware, this could make it useful for display applications where per-pixel processing time matters more than fast encoding or compatibility with established image tools.
The distinction between compression and decompression speed is central. The report says the current encoder is orders of magnitude slower than the decoder, reflecting the search required to choose palette entries and minimize visual error. That may be acceptable for assets prepared in advance, but it limits uses where images need to be encoded quickly or changed frequently. Nibnib is a project experiment, not evidence of a generally adopted image standard or a measured performance advantage across devices.
Products Worth Considering
DJI Mic 3 Wireless Microphone System (2 Transmitters, 1 Receiver, Charging Case)
DJI Osmo Mobile 8 Gimbal with Mic Mini Transmitter, Infinity Black
Logitech MX Brio Ultra HD 4K Webcam, Graphite
As an affiliate, we earn on qualifying purchases.
A Format Built Around Pixel Pairs
Nibnib defines the image format primarily through its deterministic decoding process; encoding is a separate search for data that reconstructs an acceptable approximation of the original. The first two pixels in each row are stored directly, then the decoder reads palette indices and applies their deltas to the preceding pixel value. The row stride is rounded to a multiple of four bytes.
For its initial encoder, the project uses a palette combining a cube of small delta values with randomly selected values drawn using weighted Gaussian distributions. The report says the small-delta entries help avoid overflow-related artifacts near black and full scale. The encoder tests palette choices and distributes resulting error using Floyd–Steinberg error diffusion, adapting that process to select each codeword for a two-pixel pair rather than a single pixel.
The author identifies limitations in the current approach. The palette is not selected adaptively for each source image, and the greedy search can leave artifacts trailing to the right of sharp edges. The report proposes palette optimization and deeper searches as possible improvements, while noting that they would make encoding slower without changing the existing decoder.
“There’s a budget of approximately 6 cycles per pixel.”
— Nibnib project report
Image Quality and Performance Still Need Measures
The supplied report does not give benchmark results, a measured decompression rate, or comparative tests against other formats on the RP2350. Its estimate of approximately six cycles per pixel is presented as the device-class budget motivating the design, not as a reported benchmark for Nibnib.
Image quality is also described as an area for improvement. The report calls the encoder basic and says that better palette selection and a deeper search could improve results, but provides no quantitative quality scores or broad image-set evaluation in the supplied material. Support beyond the cited demo, interoperability with other software, and whether the format will be developed into a stable release are not specified.
Possible Encoder Improvements
The project report identifies adaptive palette selection and more extensive searches for lower-error codewords as possible next steps. It also suggests changing the encoding scan direction to reduce artifacts near sharp edges, while keeping decoding left to right. These approaches could raise encoding cost; no schedule or commitment to implement them is stated.
Further evaluation would clarify how the format performs in practice: measured decoding speed on the target hardware, image quality across different content, and comparisons with alternative formats under the same constraints. The supplied report does not announce a release date or a planned milestone, so the project’s next development steps remain open.
Key Questions
What is Nibnib?
Nibnib is an experimental image format designed to make decompression simple and fast, with a focus on RP2350-class microcontrollers.
How does its decoder produce two pixels from one byte?
The byte indexes a 256-entry palette. The selected color delta is added to a duplicated previous pixel, producing a pair of RGB565 pixels.
Is Nibnib faster than other image formats?
The report describes the design goal and its decoding method, but the supplied material includes no comparative benchmark results establishing that it is faster than other formats on tested hardware.
Is Nibnib intended for archival images?
No. The project report says the format is not intended for archival and is more likely to be used alongside software that includes its decoder.
What could improve the format?
The report points to image-specific palette selection, deeper encoder searches, and changes to the encoding scan direction as potential improvements. It gives no timetable for those changes.
Source: github.com
