Monthly Archives: February 2016

Praat Scripting

I’m still not having success with formant estimation. I’ve tried yet another method of LPC analysis, this time using code from Numerical¬† Recipes in FORTRAN 77. I suspect the problem lies in my root finding code. In the meantime, I’m … Continue reading

Posted in Uncategorized | Tagged , , | Leave a comment

Debugging the Toolchain

One of the ways that the voice signal can be analyzed is via LPC (Linear Predictive Coding),¬† which essentially creates a digital filter on n coefficients. The original purpose of LPC was to compress speech. Speech is segmented into frames … Continue reading

Posted in Uncategorized | Tagged , , | Leave a comment

Guessing at Transitions

My current goal is to see if I can come up with an HMM-like approach that gives me most of the benefits of HMM without actually having to implement a full HMM system. Training for monophones is fairly (in theory) … Continue reading

Posted in Uncategorized | Tagged , | Leave a comment

Questions about Coarticulation and HMMs

One important factor in getting better synthesis is not just to capture the sustained sound of the phonemes, but the transitions – the coarticulations – that happens between the phonemes. For example, to create a more realistic results from UTAU, … Continue reading

Posted in Uncategorized | Tagged , , | Leave a comment

Looking at Hidden Markov Models (HMMs)

While I think the synthesis from synSinger is intelligible, it’s certainly not very natural. The technology behind synSinger is quite old-school: rule-based formant synthesis. The rest of the world has since moved on to using HMMs – Hidden Markov Models. … Continue reading

Posted in Uncategorized | Tagged , , | Leave a comment