2 ms·

Seems like 'limit' is a misnomer here.

• magicalhippo 3y ago
  It's a limit similar to how the resolution of a N-bit analog to digital converter (ADC) is limited to 1/2^N. You can break that limit by adding noise, known as dithering[1], to get even better resolution. Like with the squeezed light in LIGO there is no free lunch, but the idea in both cases is to pay for the lunch with something you have plenty of or otherwise don't care that much about, rather than your limited amount of money. With dithering you have to filter out the noise, but if your ADC can sample faster than you require then you just oversample and average. For LIGO they care very much about the phase accuracy but not so much the amplitude, so they pay by having worse amplitude certainty[2]. [1]: https://www.allaboutcircuits.com/technical-articles/what-is-dithering-using-noise-dithering-for-eliminating-the-quantization-distortion/ https://www.allaboutcircuits.com/technical-articles/what-is-... [2]: https://www.ligo.caltech.edu/video/ligo20231023v2 https://www.ligo.caltech.edu/video/ligo20231023v2
  • queuebert 3y ago
    I would argue that dithering is fake resolution, so you're not exceeding the limit in the information theoretic sense. Perhaps a better analogy would be compressed sensing on a sparse, structured signal where you can seemingly violate Nyquist but only because you have additional information about the sampled signal.