[Csnd] vocoder and the physics of speech

[Csnd] vocoder and the physics of speech

Date	2013-12-31 20:29
From	Machina
Subject	[Csnd] vocoder and the physics of speech
Attachments	None None

Date	2013-12-31 21:40
From	Richard van Bemmelen
Subject	Re: [Csnd] vocoder and the physics of speech
	Well, there is of course the vowgen udo: http://www.csounds.com/udo/displayOpcode.php?opcode_id=76 This can be used for creating vowels. For consonants, like s, p, t, etc, you could use filtered noise and/or short impulse signals. Richard 2013/12/31 Machina <machinadeoro@gmail.com> hi guys. im experimenting with vocoder and having lots of fun. im trying to find a way to make the human speech as clear as possible thru a vocoder. so im trying to find the right EQ/Compression/ FFT size (Seems important) to get the clearest result possible. if anyone has comments as to what kind of signal would be best to use i would also like to hear that. and… i noticed ableton vocoder has this cute "unvoiced" knob that lets you hear unvoiced syllables like sss/sshhhh/ppp/kkk. anyone has any thoughts to how it could be implemented?

Date	2013-12-31 22:53
From	peiman khosravi
Subject	Re: [Csnd] vocoder and the physics of speech
	Might be worth taking a look LPC too. It gives you an output for periodicity which lets you detect unvoiced speech. Unfortunately it's non-real-time only. Manual/lpanal.html Manual/lpread.html P www.peimankhosravi.co.uk \|\| RSS Feed \|\| Concert News On 31 December 2013 21:40, Richard van Bemmelen <zappfinger@gmail.com> wrote: Well, there is of course the vowgen udo: http://www.csounds.com/udo/displayOpcode.php?opcode_id=76 This can be used for creating vowels. For consonants, like s, p, t, etc, you could use filtered noise and/or short impulse signals. Richard 2013/12/31 Machina <machinadeoro@gmail.com> hi guys. im experimenting with vocoder and having lots of fun. im trying to find a way to make the human speech as clear as possible thru a vocoder. so im trying to find the right EQ/Compression/ FFT size (Seems important) to get the clearest result possible. if anyone has comments as to what kind of signal would be best to use i would also like to hear that. and… i noticed ableton vocoder has this cute "unvoiced" knob that lets you hear unvoiced syllables like sss/sshhhh/ppp/kkk. anyone has any thoughts to how it could be implemented?

Date	2013-12-31 23:32
From	Andres Cabrera
Subject	Re: [Csnd] vocoder and the physics of speech
	Hi, If I remember correctly, you can get whispered speech from a vocoder by randomizing the phases of the FFT bins for every window. I might be wrong though... (and I'm not sure whether that is possible in Csound) Cheers, Andrés On Tue, Dec 31, 2013 at 12:29 PM, Machina <machinadeoro@gmail.com> wrote: hi guys. im experimenting with vocoder and having lots of fun. im trying to find a way to make the human speech as clear as possible thru a vocoder. so im trying to find the right EQ/Compression/ FFT size (Seems important) to get the clearest result possible. if anyone has comments as to what kind of signal would be best to use i would also like to hear that. and… i noticed ableton vocoder has this cute "unvoiced" knob that lets you hear unvoiced syllables like sss/sshhhh/ppp/kkk. anyone has any thoughts to how it could be implemented?

Date	2014-01-01 21:27
From	Oeyvind Brandtsegg
Subject	Re: [Csnd] vocoder and the physics of speech
	For traditional vocoding (it sounds as if that's what you're after?), you could try pvsvoc, or even pvsmorph for more flexibility. A multiband compressor will probably aid in getting the sibliants more clearly through, and as for the 'unvoiced' variant, I guess someone here (Victor? Richard Dobson? ++?) has knowledge of how to get rid of the fundamental of the speech signal. best Oeyvind 2013/12/31 peiman khosravi : > Might be worth taking a look LPC too. It gives you an output for periodicity > which lets you detect unvoiced speech. Unfortunately it's non-real-time > only. > > Manual/lpanal.html > Manual/lpread.html > > P > > > > > > > www.peimankhosravi.co.uk \|\| RSS Feed \|\| Concert News > > > On 31 December 2013 21:40, Richard van Bemmelen > wrote: >> >> Well, there is of course the vowgen udo: >> http://www.csounds.com/udo/displayOpcode.php?opcode_id=76 >> >> This can be used for creating vowels. For consonants, like s, p, t, etc, >> you could use filtered noise and/or short impulse signals. >> >> Richard >> >> >> 2013/12/31 Machina >>> >>> hi guys. >>> >>> im experimenting with vocoder and having lots of fun. >>> im trying to find a way to make the human speech as clear as possible >>> thru a vocoder. >>> so im trying to find the right EQ/Compression/ FFT size (Seems important) >>> to get the clearest result possible. >>> >>> if anyone has comments as to what kind of signal would be best to use i >>> would also like to hear that. >>> >>> and… >>> i noticed ableton vocoder has this cute "unvoiced" knob that lets you >>> hear unvoiced syllables like sss/sshhhh/ppp/kkk. anyone has any thoughts to >>> how it could be implemented? >> >> > -- Oeyvind Brandtsegg Professor of Music Technology NTNU 7491 Trondheim Norway Cell: +47 92 203 205 http://flyndresang.no/ http://www.partikkelaudio.com/ http://soundcloud.com/brandtsegg http://soundcloud.com/t-emp