Samplify Emerges with Good Enough Compression

Samplify Systems, a two-year-old chip startup coming out of stealth mode today, has the potential to create faster data compression and simplified gadgetry for everything from wireless communications to ultrasound machines — as long as users don’t mind losing a little data. The startup offers a mixed-signal chip that translates analog signals, such as sound waves, into digital signals comprised of zeros and ones; but it can compress data too. This results is less circuitry on the chip and the ability to move data in a smaller format.
The startup is also cost-efficient — it built its first chip with a meager $5.6 million budget — but the most noteworthy aspect of Samplify’s technology is the way it thinks about compression. Compression is a black or white issue; it’s either lossless, meaning all the data is retained, or it’s lossy, meaning some of the less important data is dumped in the name of file size.
Samplify wants to inject some grey into that issue with a compression algorithm that enables the chip to lose a predetermined amount of data. This allows for fixed error levels that offer a middle ground between lossless and lossy compression. An EDN article from June calls this “good enough” compression and speculates that wireless base stations and medical imaging could benefit from the technology. Since we’re moving further into the digital age with everything from our phone calls to our medical records, Samplify is obviously hoping this is an idea whose time has come.