A drop-in compression engine for cameras and sensors. Up to 6:1 at full acquisition rate — no buffering, no signal loss, no pipeline changes.
— Why compression matters now
With Jetraw Core embedded at the camera level, that data is compressed before it ever hits your network — live, at acquisition rate.
TRL 9-qualified for space. Jetraw Core triples satellite image capacity per orbit without changing hardware, downlink, or ground station.
Your customers shouldn’t have to choose between image quality and throughput. Jetraw Core-enabled cameras deliver all three — same sensor, same interface, same pipeline.
— The compression your AI actually needs
Every mainstream codec optimises for human perception. Jetraw Core optimises for measurement accuracy — the difference between data your AI can trust and data it can’t.
JPEG 2000 (visually lossless)
Standard lossless
Jetraw Core
Compression ratio
4:1–20:1 (lossy)
1.7:1–2:1
5:1–10:1
Quality target
Human perception
Mathematical exactness
Metrological accuracy
“Visually lossless means a human couldn’t tell. We give you proof.”
Jetraw Core understands the noise and signal characteristics of your sensor, and uses that knowledge to introduce only bounded deviations that behave like the natural noise already present in any image. You can validate data quality after compression, generate synthetic training data, and present an auditable quality record to safety certification bodies. JPEG 2000 cannot do any of these things, the standard was never designed for measurement.
Quality impact: a 1.2 dB SNR reduction, comparable to the difference between ISO 100 and ISO 130. Quantified, reproducible, auditable.
— The physics-based difference
Dotphoton calibrates a noise model for your sensor — read noise, shot noise, dark current, gain. One-time. 1–2 days with your team.
Signal is what your AI needs. Noise is statistically predictable. At acquisition time, the noise model separates them — in real-time, at line rate.
Jetraw treats signal and noise separately. The signal, the information your processing pipeline actually uses, is preserved exactly. The noise, which contains no useful information, is compressed aggressively. The result is a file as small as a lossy codec's, but with the integrity of a lossless one.
Stored in your existing RAW format. No new format, no pipeline changes. Free decompressor — no licence on the receive side.
— Two paths to integration
Camera manufacturers · Satellite OEMs · Sensor-level or before output interface
Integrates into your FPGA as an AXI4-stream pipeline stage. Compresses at line rate, ~60 clock cycles latency. Verified on AMD/Xilinx and Altera.
Throughput
Up to 6.4 Gpx/s (32px parallel)
Latency
~60 clock cycles
FPGA resources
5000 LUT (1px) -> 120,000 LUT (32px)
Interface
AXI4-stream
Pixel formats
16-bit, configurable
Sensor support
Mono, Bayer
Platforms
AMD/Xilinx, Altera
Compression ratio
Typically 6:1
System integrators · Camera manufacturers · AI/ML pipeline integration
Codec callable directly from your acquisition application. Multi threading delivers 6.2 GB/s on modern hardware. C, C++, C#, Java, Python.
Throughput
6.2 GB/s (i9 14900K, Windows)
Threading
Multi-thread support
Languages
C, C++, C#, Java, Python
OS
Linux, Windows, macOS
CPU
x86-64, arm64
Pixel formats
RAW pixel buffer or TIFF files
Compression ratio
Typically 6:1
— Real numbers from production deployments
8 Mpx camera · 36 fps · one test vehicle · full day of operation
Metric
Without Jetraw
With Jetraw Core
Daily acquisition volume
7 TB
1.2 TB
Annual volume per vehicle
1.7 PB
0.3 PB
Upload time over 10 Gbps
1.5 hours
0.2 hours
Annual cloud transfer time
16 days
3 days
Storage cost per year
€706,700
€117,800
“It is extremely appealing for PCO to obtain a compression method that exploits the individual image capture chain of each camera model.”
Gerhard Holst
Head of Science and Research, Excelitas PCO
— How we work together
We analyse your optical system, data volumes, and target application. Quantified benefits estimate — compression ratio, throughput, cost savings. No commitment.
Your team ships a camera or sensor. Our engineers build the noise model. You receive calibration files and a technical integration brief.
You receive an R&D licence — full access, no restrictions. Integrate the FPGA IP Core or software library into your test pipeline. Dotphoton engineers are available for support. You run your own benchmarks.
When satisfied, you take a per-device production licence. Pricing scales with volume. Free decompressor for end users — no friction on the receive side.
Dedicated support from Dotphoton engineers throughout your deployment. Long-term maintenance and updates included.
— FAQ
Is Jetraw lossless or lossy?
Neither — and both. We call it “better-than-lossless.” Noise is compressed lossily according to the sensor’s statistical noise model — information-theoretically safe. Signal is compressed losslessly. The underlying measurement is preserved exactly, while file size approaches a lossy codec.
Do our end users need a licence to decompress?
No. The Jetraw decompressor is free for Windows, Linux, and macOS. Your customers can decompress without any licence or cost.
Does Jetraw require a new file format?
No. Jetraw is format-agnostic. It works within your existing RAW format. Your downstream pipeline, customer software, and archival systems do not need to change.
What sensors does Jetraw Core support?
CMOS, sCMOS, and CCD sensors in mono, Bayer, and CFA configurations. 16-bit image data. Any resolution — Jetraw Core is configurable to your specific sensor dimensions.
What FPGAs are supported?
Xilinx is the primary supported platform. Intel FPGAs are available on request.
How long does sensor calibration take?
Typically 1–2 days. Calibration can be done with hardware shipped to Dotphoton, or remotely with raw image samples captured to a specification we provide.
Is Jetraw Core certified for safety-critical use?
Yes. Integrated into safety-critical automotive ADAS and space imaging systems. Contact us for documentation relevant to your certification process.
What’s the difference between Jetraw Core and Jetraw Platform?
Jetraw Core is for manufacturers — it embeds compression into your device or acquisition system. Jetraw Platform is for end users — data management, storage, and workflow tools for labs and imaging teams. The two are complementary.
We use cookies to enhance your browsing experience and analyse site traffic. By clicking "Accept all" you consent to our use of cookies. Manage preferences