From: Joerg Hansmann (info_at_jhansmann.de)
Date: 2002-02-27 11:50:59
Hi,
----- Original Message -----
From: Dave <dfisher_at_pophost.com>
To: <buildcheapeeg_at_yahoogroups.com>
Sent: Wednesday, February 27, 2002 2:20 AM
Subject: Re: [buildcheapeeg] EEG FFT Display help
...
BTW: There is a way to make the display interpolation appear in the data by resampling
to e.g twice the original sample rate:
Edit->Convert Sample Type
> However, for things such as
> FFT analysis, is the interpolated analog data being used, or the actual sampled
> data?
I presume the FFT runs on the real data.
After rethinking the interpolation a bit, it seems to me that it is
mathematically correct.
For the FFT there seems to be a way to make some sort of interpolation:
In options->settings->spectral
is the option to increase the resolution to get a more accurate frequency display
at the cost of a less accurate time-resolution.
(That is a characteristic for all sorts of time-frequency analysis)
By reducing the window width (e.g. dividing by 2 for each doubling of the FFT resolution
) the time resolution gets more accurate seemingly without loosing frequency
resolution. However I am not sure if this can really be true, because it would
mean, that both frequency _and_ time resolution could be increased at the same time.
IMO this could be a sort of interpolation, but I am not sure.
Regards,
Joerg
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:39 BST