These are beautiful artistically...<p>...but I don't have the slightest clue what they <i>mean</i>, and I've certainly dabbled in FFT and spectrogram and wavelet work, on top of a lot IPA vowel work, but I'm missing the <i>why</i> behind the formulas given and I'm missing how these plots are supposed to relate to frequencies visually.<p>A spectrogram of someone pronouncing vowels is extremely straightforward. Recognizing patterns of formants in spectrograms is quite simple.<p>So what is this trying to reveal that spectrograms don't? Besides that, what are the axes? Why are these circular or presumably polar? Why are they spiky? Why the particular blue/red bandpass filter? And what does autocorrelation have to do with vowels?<p>I'm not sure I've ever found myself so mystified by something I feel like I should have the background to understand quite easily.<p>If they're just supposed to be works of art then that's cool. But the title "visual morphology of vowels" seems like the plots are intended to reveal some kind of link between frequencies and the shape of the mouth maybe? But the example images aren't even labeled by which vowel they represent so I'm just baffled.
These are certainly very pretty, but I can’t help but feel that the usual visualisations are much more information-dense and helpful [0] [1]. Spectrograms and vowel diagrams let me instantly see the formants of a vowel, allowing me to compare them to surrounding sounds and understand how they’re made in the mouth. By contrast, I can’t quite understand how to identify any resonant frequencies in this visualisation; the bands which are so obvious in spectrograms are difficult if not entirely impossible to find here. It would be very interesting to see if this visualisation could be modified to make this more prominent.<p>(They really are very pretty, though!)<p>[0] <a href="https://en.wikipedia.org/wiki/Vowel_diagram" rel="nofollow">https://en.wikipedia.org/wiki/Vowel_diagram</a><p>[1] <a href="https://en.wikipedia.org/wiki/Spectrogram" rel="nofollow">https://en.wikipedia.org/wiki/Spectrogram</a>
Pink Trombone (only "safe for work" if you use headphones):<p><a href="https://dood.al/pinktrombone/" rel="nofollow">https://dood.al/pinktrombone/</a><p>How to Break Pink Trombone:<p><a href="https://www.youtube.com/watch?v=djUxAqss4KY" rel="nofollow">https://www.youtube.com/watch?v=djUxAqss4KY</a><p>Pink Trombone takes on "Take On Me"<p><a href="https://www.youtube.com/watch?v=5UaRqfOpt90" rel="nofollow">https://www.youtube.com/watch?v=5UaRqfOpt90</a>
I've been casually researching visualisations of sound for 2 years now, and recently came across another interesting discovery, and wanted to share it.<p>It slightly improves the way ACF images are presented, but this small improvement makes a big difference. It works best on "small sounds" that last 1-2 sec, such as vowels or sample recordings of flute, violin and so on. The sound is analysed with FFT with the sliding window of 1/4 sec that advances by 1/500 sec at a time until it covers the entire waveform. After computing FFT spectrum for each frame, a basic bandpass filter is applied to separate high and low frequencies. The result is fed to the inverse FFT, thus computing ACF, and presented in polar coordinates using a basic red-blue color scheme. The effect is that low frequencies appear red and high frequencies appear blue.<p>To my surprise, this basic method reveals a large variety of distinctive, yet visually appealing, shapes for vowel sounds.
This is great. I like the presentation of the sounds. I would like to see which vowel sound each autocorrelation function represents. I don't see labels for any of them.<p>Excellent work though.