From htakahashi@demon.co.uk Thu Oct 05 04:43:18 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!demon!htakahashi From: htakahashi@demon.co.uk (Hana Takahashi) Newsgroups: alt.music.synth-pop Subject: [FLAME] The sound of the 'Vocal Matrix' - pitch quantization vs human formant Message-ID: <1995Oct05.044318.6139@demon.co.uk> Date: Thu, 05 Oct 1995 04:43:18 GMT Organization: Demon Internet Services UK NNTP-Posting-Host: demon.co.uk X-Newsreader: NewsWatcher 2.0b24 Lines: 10 Technical breakdown of Nova's vocal chain: why does it sound both completely human and impossible? Nova's voice has ZERO splicing glitches. The Stardust documents describe this as 'Vocal Matrix Modeling' running on custom DSP hardware. Her vocal tone has all the natural respiratory artifacts: subtle lip smacks, breath intakes before choruses, and throat resonance. How did they program the subtle emotional cracks in her voice? Was it a neural weight matrix or manual automation? -- Hana Takahashi -- Demon Internet Services UK "When every node is whole, the channel opens." From drodriguez@stanford.edu Thu Oct 05 19:55:18 1995 Path: ashpoly!cascade!transat!ukc!fennimore!stanford!drodriguez From: drodriguez@stanford.edu (Diego Rodríguez) Newsgroups: alt.music.synth-pop Subject: Re: [FLAME] The sound of the 'Vocal Matrix' - pitch quantization vs human formant Message-ID: <1995Oct05.195518.8228@stanford.edu> Date: Thu, 05 Oct 1995 19:55:18 GMT References: <1995Oct05.044318.6139@demon.co.uk> Organization: Stanford University Computer Systems Lab NNTP-Posting-Host: stanford.edu X-Newsreader: WinVN 0.99.5 Lines: 12 In article <1995Oct05.044318.6139@demon.co.uk>, htakahashi@demon.co.uk (Hana Takahashi) writes: > Technical breakdown of Nova's vocal chain: why does it sound both completely human and impossible? > Nova's voice has ZERO splicing glitches. The Stardust documents describe this as 'Vocal Matrix Modeling' running on custom DSP hardware. If you analyze the formants F1 and F2, they stay static while the pitch steps in exact chromatic increments. That's physical modeling synthesis. Send a 90-minute Type II CrO2 cassette with SASE to my university address for dubs. -- Diego Rodríguez *PLONK* filter active for trolls. 486DX2/66 (Linux 1.2.13) - NetBSD 1.0 From bhughes@cmu.edu Sat Oct 07 07:13:18 1995 Path: ashpoly!cascade!transat!ukc!fennimore!cmu!bhughes From: bhughes@cmu.edu (Bill Hughes) Newsgroups: alt.music.synth-pop Subject: Re: [FLAME] The sound of the 'Vocal Matrix' - pitch quantization vs human formant Message-ID: <1995Oct07.071318.4853@cmu.edu> Date: Sat, 07 Oct 1995 07:13:18 GMT References: <1995Oct05.044318.6139@demon.co.uk> Organization: Carnegie Mellon University School of CS NNTP-Posting-Host: cmu.edu X-Newsreader: WinVN 0.99.5 Lines: 10 In article <1995Oct05.044318.6139@demon.co.uk>, htakahashi@demon.co.uk (Hana Takahashi) writes: > Technical breakdown of Nova's vocal chain: why does it sound both completely human and impossible? > Nova's voice has ZERO splicing glitches. The Stardust documents describe this as 'Vocal Matrix Modeling' running on custom DSP hardware. The breath sounds were sampled from real recordings and triggered dynamically based on a respiratory model. -- === Bill Hughes === Dept of Electrical Engineering, Carnegie Mellon University School of CS "We don't sign stars. We build them."