[FieldTrip] Hann taper vs. single Slepian taper

[Robert Oostenveld]

To complement Ingrids good explanation: the shape of the Slepian  
tapers depend on how many you compute, i.e. if you compute one it will  
have a (slightly) different shape that the first one if you compute  
four. The first Slepian taper never goes exactly to zero at the edges,  
whereas the Hanning (or Hann, as it is also called) does. The  
consequence is that a single Slepian leaks more than a Hanning. The  
reason for using a Hanning taper is typically that you don't want to  
spectrally smooth (usually because the time window is already on the  
short side), so the leakage of a single Slepian is in that case not  
desired.

Robert




On 18 Dec 2010, at 1:13, Ingrid Nieuwenhuis wrote:

> Dear David,
>
> The Slepian tapers are sets of multiple (orthogonal) tapers, that  
> only combined accomplish this "optimally maximized frequency  
> resolution while minimizing leakage from distant frequencies". If  
> you use only a single taper, the Slepian taper is suboptimal  
> compared to the Hanning taper. I've attached some teaching slides I  
> have with some background on multitapers.
>
> Hope this helps,
> Cheers,
> Ingrid
>
>
> On 12/16/2010 6:01 AM, David Groppe wrote:
>> Hi Fieldtripians,
>>    Could anyone explain to me when, if ever, it would be better to
>> perform a DFT with a Hann taper vs. a single Slepian taper?  It was  
>> my
>> understanding that Slepian tapers in general optimally maximized
>> frequency resolution while minimizing leakage from distant
>> frequencies.  However, a colleague of mine says he was told at a
>> workshop at the Donders Institute that a Hann taper was better than a
>> single Slepian taper (if you are only going to use one taper).
>>        thanks for your help,
>>             -David
>>
>
> -- 
> Ingrid Nieuwenhuis PhD
> Sleep and Neuroimaging Laboratory
> Department of Psychology
> University of California, Berkeley
> California 94720-1650
>
> <multitapers.pdf>_______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip