Skip to main content
U.S. flag

An official website of the United States government

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Why are we concerned with large covariance matrices?



Paul D. Hale, Andrew M. Dienstfrey, Chih-Ming Wang


In order to promote discussion, we outline the two different paradigms for propagating uncertainties from the various waveform measurement impairments to the waveform parameters. The first method propagates uncertainty from the impairments directly to a specific parameter without explicit reference to the waveform. The second propagates uncertainty from the impairments to the waveform and then, using hte chain rule of derivatives, propagates uncertainty from the waveform to the parameter. The later, more general, method can generate a covariance matrix for the waveform that is order NxN. For N large, the problem becomes numerically intractable and statistical inference based on classical arguments can fail for M


covariance, metrology, uncertainty, waveform


Hale, P. , Dienstfrey, A. and Wang, C. (2010), Why are we concerned with large covariance matrices?, NIST, [online],, internal:/nist,gov (Accessed May 21, 2024)


If you have any questions about this publication or are having problems accessing it, please contact

Created June 1, 2010, Updated April 12, 2019