Bit-rate: wideband & narrowband (#ARIJIT BISWAS# )


Subject: Bit-rate: wideband & narrowband
From:    #ARIJIT BISWAS#  <arijit17@xxxxxxxx>
Date:    Mon, 26 Jun 2006 18:28:17 +0800

This is a multi-part message in MIME format. ------_=_NextPart_001_01C6990B.69E7F292 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Dear List: =20 This e-mail may not be relevant for this group, but anyway.... =20 I was reading the book, "Speech Coding and Synthesis" by Kleijn & = Paliwal. =20 In page 445, from the table we see that for scalar quantization of the = reflection coefficients, it costs 4 bits per parameter for 1 dB = distortion. =20 Similarly, in page 447, from the table we see that for LAR scalar = quantization, it costs 3.2 bits per parameter for 1 dB distortion.=20 =20 The above numbers are for narrowband speech coding. Could anyone please = let me know what happens to the above numbers (4 bits and 3.2 bits) for = wideband LP based speech coding? It would be great if you could please = suggest some references as well.=20 =20 In case, there are no relevant wideband speech coding paper that = addresses the numbers for the reflection coefficient and LAR = representation, you may also let me know the numbers for the LSF = representation for the narrowband and wideband scenarios.=20 =20 Thanks a lot in advance. =20 Best Regards, ~Arijit ------_=_NextPart_001_01C6990B.69E7F292 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Diso-8859-1">=0A= <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">=0A= <HTML><HEAD>=0A= =0A= <META content=3D"MSHTML 6.00.2900.2912" name=3DGENERATOR>=0A= <STYLE></STYLE>=0A= </HEAD>=0A= <BODY bgColor=3D#ffffff>=0A= <DIV id=3DidOWAReplyText20021 dir=3Dltr>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">Dear =0A= List:</SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt"></SPAN>&nbsp;</P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">This =0A= e-mail may not be&nbsp;relevant for this group, but anyway....</SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">&nbsp;<?xml:namespace =0A= prefix =3D o ns =3D "urn:schemas-microsoft-com:office:office" =0A= /><o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">I was =0A= reading the book, &#8220;Speech Coding and Synthesis&#8221; by Kleijn = &amp; =0A= Paliwal.<o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">&nbsp;<o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">In =0A= page 445, from the table we see that for scalar quantization of the =0A= <B>reflection coefficients,</B> it costs <B>4 bits per parameter for 1 = dB =0A= distortion.</B><o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">&nbsp;<o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">Similarly, =0A= in page 447, from the table we see that for <B>LAR</B> scalar = quantization, it =0A= costs <B>3.2 bits per parameter for 1 dB distortion. = <o:p></o:p></B></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">&nbsp;<o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">The =0A= above numbers are for narrowband speech coding. Could anyone please let = me know =0A= what happens to the above numbers (4 bits and 3.2 bits) for <B>wideband = LP based =0A= speech coding</B>? It would be great if you could please suggest some = references =0A= as well. <o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">&nbsp;<o:p></o:p></SPAN></P>=0A= <P class=3DMsoBodyText style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; mso-bidi-font-size: 12.0pt">In case, there are = no =0A= relevant wideband speech coding paper that addresses the numbers for the =0A= reflection coefficient and LAR representation, you may also let me know = the =0A= numbers for the LSF representation for the narrowband and wideband = scenarios. =0A= <o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">&nbsp;<o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">Thanks =0A= a lot in advance.<o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">&nbsp;<o:p></o:p></SPAN></P>=0A= <P class=3DMsoNormal style=3D"MARGIN: 0cm 0cm 0pt"><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-bidi-font-size: = 12.0pt">Best =0A= Regards,<o:p></o:p></SPAN></P><SPAN lang=3DEN-US =0A= style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Verdana; mso-ansi-language: = EN-US; mso-fareast-font-family: 'Times New Roman'; mso-fareast-language: = EN-US; mso-bidi-language: AR-SA; mso-bidi-font-size: 12.0pt; = mso-bidi-font-family: 'Times New = Roman'">~Arijit</SPAN></DIV></BODY></HTML>=0A= ------_=_NextPart_001_01C6990B.69E7F292--


This message came from the mail archive
http://www.auditory.org/postings/2006/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University