According to Alice Wellborn, a dyslexia expert of some note: Many years ago, researchers believed that dyslexia was a visual perceptual problem - that it was based in how a person saw letters and words. Now we know for sure, through brain imaging studies, that dyslexia is a problem in the language system of the brain, not the visual system.
Dyslexia is the result of a significant weakness in the phonological processing system, or how a person's brain understands and can use the sound-based reading "code". A dyslexic reader has difficulty cracking that code.
Does that mean that dyslexics have trouble with the spoken word? The Wikipedia page lists such issues as associated conditions:
https://en.wikipedia.org/wiki/Dyslexia#Associated_conditions
``Many people with dyslexia have auditory processing problems ...''
If someone is using the wah-wah pedal with an electric guitar to change the timbre of a note, can you tell?
Is it really about sound? Or is it just that the "phonetic analyzer" isn't distinguishing the phonemes?
If you can consciously hear the acoustic difference in timbre, how can you then not use that to tell the words apart (even if slowly/inefficiently)? As in "I just heard something that may be pin or pen; now I have to concentrate on what I heard. Hmm, the vowel sounded like a band filter tuned toward a higher frequency so it must be pin, rather than pen."
It's plausible that we can manipulate the raw input signal in order to recreate a similar semantic handicap later in a normally functioning pipeline, such as slowed reading due to not being able to resolve the order of letters in a word.
The demo was inspired by a particular dyslexic's remark that the letters seem to "jump around" for her.
This could be a "piece of the puzzle"; of course we can't naively believe that this reveals everything that dyslexia is about, in all its manifestations.