SpeechWiki - Recent changes [en]

Timeliner

Xzhuang2 — Sun, 22 Nov 2009 14:11:27 GMT

table

← Older revision		Revision as of 14:11, 22 November 2009
Line 1:		Line 1:
-
	== ruby codes ==		== ruby codes ==

Line 28:		Line 27:

	== feature generating (to be detailed) ==		== feature generating (to be detailed) ==
-	~~3a. filter bank~~

-	3b. ~~posterior output~~ (~~ANN~~, ~~HMM~~(~~DBN~~) )	+	/workspace/ifp-32-2/hasegawa/xzhuang2/AED2009/tmp/forCamille
		+
		+	A. To make a feature file:
		+	HCopy -C $ConfigFile AIT_20061020_AmarkIII_1.ch4.wav $ResultFile
		+
		+	B. To view a feature file:
		+	HList -h $ResultFile
		+
		+	{\| {{prettytable}}
		+	\|+ HCopy-made features in htk-format
		+	! ConfigFile
		+	! ResultFile
		+	! Description
		+	\|-
		+	\| HCopy_FB.cfg
		+	\| AIT_20061020_AmarkIII_1.ch4.feaFB
		+	\| narrow-band filterbank (window 25ms, step 10ms), 78-dim
		+	\|-
		+	\| HCopy_MFCC.cfg
		+	\| AIT_20061020_AmarkIII_1.ch4.feaMFCC
		+	\| MFCC (window 25ms, step 10ms), 78-dim
		+	\|-
		+	\| HCopy_FB_w6ms_o2ms.cfg
		+	\| AIT_20061020_AmarkIII_1.ch4.feaFB_w6ms_o2ms
		+	\| wide-band filterbank (window 6ms, step 2ms), 78-dim
		+	\|}
		+

-	3x. ~~general feature writer function~~	+	Note:
		+	AIT_20061020_AmarkIII_1.ch4.wav wavform
		+	AIT_20061020_AmarkIII_1.ch4.annhtk 92 dims (14 dim decorrelated event-classifier neural network output + 78 dim filter bank parameters)

-	~~== todo ==~~	+	1) For each 78 dim filter bank parameters (or MFCCs), the first 26
-	~~feature writer function; doc; verify htk feature read in ruby~~	+	dims are original, the second and third 26 dims are first-order and
		+	second-order regression coefficients derived from the first 26 dims
		+	respectively.

Sst09

Posenhuang — Thu, 19 Nov 2009 22:10:45 GMT

← Older revision		Revision as of 22:10, 19 November 2009
Line 43:		Line 43:
	\| 11/19		\| 11/19
	\| Po-Sen Huang		\| Po-Sen Huang
-	\| Speech Enhancement Employing Laplacian-Gaussian Mixture	+	\|[[Media:Speech_Enhancement.pdf‎\|Speech Enhancement Employing Laplacian-Gaussian Mixture]] ([[Media:speech_distribution.pdf‎\|slide]])
-	\| [Media:~~Speech distribution~~]	+
	\|-		\|-
	\| 11/26		\| 11/26

File:Speech distribution.pdf

Posenhuang — Thu, 19 Nov 2009 22:10:32 GMT

uploaded "[[File:Speech distribution.pdf]]"

New page

Sst09

Posenhuang — Thu, 19 Nov 2009 21:59:50 GMT

← Older revision		Revision as of 21:59, 19 November 2009
(One intermediate revision not shown)
Line 44:		Line 44:
	\| Po-Sen Huang		\| Po-Sen Huang
	\| Speech Enhancement Employing Laplacian-Gaussian Mixture		\| Speech Enhancement Employing Laplacian-Gaussian Mixture
		+	\| [Media:Speech distribution]
	\|-		\|-
	\| 11/26		\| 11/26

Landmarks09F

Mark Hasegawa-Johnson — Tue, 17 Nov 2009 18:59:11 GMT

Landmark-Based and Prosody-Dependent Speech Recognition, 2009-2010:

← Older revision		Revision as of 18:59, 17 November 2009
(One intermediate revision not shown)
Line 1:		Line 1:
	==Landmark-Based and Prosody-Dependent Speech Recognition, 2009-2010==		==Landmark-Based and Prosody-Dependent Speech Recognition, 2009-2010==
		+
		+	; Tuesday December 8, 2009, 12:30-2:00, 2169 BI
		+	: Tim Mahrt
		+	: Automatic P-score and B-score labeling using HMMs
		+
		+	; Tuesday December 1, 2009, 12:30-2:00, 2169 BI
		+	: Yoonsook Mo
		+	: Speaker-dependent vs. speaker-independent models of prosody
		+	: Boundary detection with vs without pause
		+
		+	; Tuesday November 11, 2009, 12:30-2:00, 2169 BI
		+	: Jui-Ting Huang and Po-Sen Huang
		+	: Variable-parameter HMM indexed by P-score (prominence score)

	; Tuesday October 20, 2009, 12:30-2:00, 2169 BI		; Tuesday October 20, 2009, 12:30-2:00, 2169 BI

User:Techychic

Techychic — Tue, 17 Nov 2009 09:33:49 GMT

Created page with 'Hi everyone, I have been creating interesting posts that you could read and make some discussions but right now I am still preparing a lot of things to settle things perfectly. D...'

New page

Hi everyone,
I have been creating interesting posts that you could read and make some discussions but right now I am still preparing a lot of things to settle things perfectly. Don't worry; I'll be keeping in touch. Anyway, if you are searching for love, try this [http://www.date.com/es/ Latino en linea].It's an online free matchmaking dating site, Search millions of local personals to find your true love and soulmates.

User:Techychic

Techychic — Tue, 17 Nov 2009 09:32:26 GMT

New user account

New page