Computers can modify pitch and formants independently of speech timing

Evidence

Computers can detect both pitch and formants; we have seen the tools for this in action (Fourier analysis, sonograms) many times by now. What is perhaps not as obvious is that computers can independently change pitch and formant while keeping the rate of speaking—the timing of words—unchanged. This is far from the old chipmunk effect of speeding up playback. … Changing formants is perhaps easiest to understand: once detected, they can be artificially reduced by filters. The key is to make a sonogram transformation, which as we know is a kind of mixed time and frequency analysis. Artificial reduction of the formants is combined with artificial creation of others to create new formants. Then, the sonogram is transformed back to the original time domain. — Eric J. Heller, Why You Hear What You Hear- An…, 1720

Categorization

• XREF: Independent pitch/formant manipulation (PSOLA-type algorithms) connects to how voice transformers work and to hearing research on perceptual dissociation of these cues.
• SEED: Article angle: how voice-transformation technology works and why formant manipulation (not just pitch) is what makes a male voice sound convincingly female and vice versa.