14:55 – quantization in everyday life
“Quantization happens everywhere in life. We use quantisation without knowing it. And we can earn a lot of money on this research,” says professor István Kollár, a guest researcher from Budapest University of Technology and Economics sponsored by Ericsson, in his live broadcasted talk on the ACCESS Distinguished lecture series
– When I ask ‘what time is it?’ then we usually say ‘it is 11:24’. This is already quantized. If you use a digital watch, (which I hate, so I would rather use an analogue one), then you see time quantized, since you have numbers instead of analog quantities. 11:24 is not exact,” says István Kollár, but it is good enough for daily life.
The same is true for the 100 grams of ham that we buy in the store, which is in reality not exactly 100 grams. Another example is Easter Sunday which is the Sunday following the full moon after the March Equinox – this not quite true, but approximately – and this is nothing else than quantization of the date first by 28, then by 7 days.
– All the same we think that a computer is exactly precise. But it’s not, he says.
Quantization represents all we need to understand
We state quantities in as many digits as are required to keep the necessary information, Kollár explains. The seconds or other small fractions are often not very important for our understanding in everyday life. Quantization means that in nature we represent a continuous quantity by finite-digit numbers.
So why do we need to quantize?
– is part of life. So we have to understand it, Kollár replies thoughtfully.
István Kollár is Professor of Electrical Engineering at the Department of Measurement and Information Systems, Budapest University of Technology and Economics, Budapest, Hungary. In 2009 Ericsson is sponsoring a guest research grant at KTH ACCESS Linnaeus Centre.
For PhD student David Zacharias, Professor Kollár’s lecture was a light bulb moment.
– When I was learning sampling and quantization as an undergrad I always wondered whether it could by applied something analogous to the sampling theorem in quantization, since it seemed in some sense to be just sampling along a different axis. The talk caught my attention by showing that this intuition had some grounds and there was indeed a well-established theory for it, says David Zacharias at the Signal Processing Lab.
This year, Professor Kollár first visited Sweden during the Erasmus IP Summer School on Distributed Measurement Systems, June 22 - July 3, 2009. During his second visit in September, he gave a talk within the ACCESS Distinguished Lecture Series entitled “Quantization and round-off: abstract science or problems of everyday life?” The visit was hosted by the Signal Processing Lab.
Professor Kollár emphasizes the applicability and economical return on this research:
– During digital sound recording, it is important not to have background noise that changes with time. This can be achieved by using triangular dither. This is pleasant to the ear - and you can earn a lot of money with more pleasant recording...
