Change search
ReferencesLink to record
Permanent link

Direct link
Pole-zero modelling of speech for use in nasality based speaker recognition.
KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
KTH, School of Engineering Sciences (SCI), Mathematics (Dept.).
2012 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

In this bachelor thesis we performed a comparison between

different methods of fitting pole-zero filters to data, and

their usefulness for nasality-based speaker recognition in

particular. It is believed that nasality has low intra-speaker

variation and high inter-speaker variation, and that polezero

filters are good at capturing the nasal characteristics.

We describe to the extent possible, the theory underpinning

the various methods, and then compare them in various

ways. We used simulated speech data, to see to what

extent the methods provided good estimates of the true filters,

when noise had been introduced afterwards. Another

way we compared the methods was to determine how well

they perform with respect to classifying real speech data,

when various features had been extracted with help from

the computed filters.

Our results were that the all-pole method was superior to

the other methods that we considered, at speaker recognition

on nasal phonemes, contradicting our hypothesis.

Abstract [sv]

I detta kandidatexamensarbete unders¨okte vi olika metoder

f¨or att skatta pole-zero filter utifr°an givna data och metodernas

anv¨andbarhet i nasalitetbaserad talarigenk¨anning.

Vi f¨ormodar att nasalitet har l°ag variation ¨over en och samma

talare, samtidigt som den varierar markant fr°an talare

till talare. Vi f¨ormodar vidare att pole-zero filter ¨ar l¨ampade

f¨or att beskriva nasalitet. Vi beskriver i m¨ojligaste m°an teorin

bakom de olika metoderna och j¨amf¨or sedan dessa p°a

olika s¨att. Vi anv¨ander bland annat simulerade talsignaler

med p°alagt brus f¨or att unders¨oka metodernas robusthet.

Vi j¨amf¨or sedan metoderna genom att se hur v¨al de klassifierar

talare utifr°an olika features som vi ber¨aknat fr°an


V°ara resultat var att metoden f¨or att skatta all-pole filter

var ¨overl¨agsen p°a att klassifiera talare utifr°an nasala fonem,

och mots¨ager allts°a v°ar ursprungliga hypotes.

Place, publisher, year, edition, pages
2012. , 71 p.
National Category
Engineering and Technology
URN: urn:nbn:se:kth:diva-118352OAI: diva2:605852
Available from: 2013-02-15 Created: 2013-02-15 Last updated: 2013-02-15Bibliographically approved

Open Access in DiVA

Rickard Norlander, Máté Szekér kandidatex(1386 kB)414 downloads
File information
File name FULLTEXT02.pdfFile size 1386 kBChecksum SHA-512
Type fulltextMimetype application/pdf

By organisation
Mathematics (Dept.)
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 414 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 134 hits
ReferencesLink to record
Permanent link

Direct link