Anyone else having a problem with the Speaker in Hiragana Listening Practice?

I do...it's very hard to understand it's pronounciations  ふ , ご ,and  ほ​. Are just some of the ones that sound a little off.
For exaple: When the Hiragana is お it sounds as if it's saying ''guo'' or something among the lines of that.
Or am I the only idiot with hearing problems?

I think the hiragana listening samples are a bit hard to understand because they are spoken at speaking pace which makes it a bit hard to distinguish the sounds.  They are also computer generated sounds, so they don't sound as perfect and clear as if a human had said them unfortunately.

Also some of the Japanese pronunciations are different to how you would expect in English.  For example the ふ (fu) is a bit different to how you pronounce it in english, its a bit hard to describe but the sound is made while exhaling when saying it in Japanese, while the English you tend to purse your lips a bit more when you say it.

When I listen to them, they do as you say, sound a little bit off, but I think that is mostly because they are computer generated.

Anyway, interested to hear what everyone else thinks of them.

Most of the samples are ok, but there are some trouble spots.  What surprised me, is that there is not a soundboard (like the class of cellphone apps) that lets the user play any character they wish (and thus compare and learn the computer generated voice). Currently the site only lets a user play characters from just a set.

Example, computer generated voice for え seems to have an bit of a leading 'n' sound making me auditorily confuse it for ね (which does sound correct). Or as seems though my head phones, to be closer to 'ne' than the more 'ea' sound that I'm use to.  And since they're in different 'sets' directly comparing the computer generated voice for え & ね is unnecessarily difficult.

To effectively use the listening practice, a user has to learn the computer generated voice samples used by the site.

The other thing that I thought might help to the computer generated voice, would be a text input to computer generated voice samples.  Where a user could type "kakakananana" to hear three か directly followed by three な in the computer generated voice samples. Yes it will likely sound as bad as computer generated voices in Phone Menu systems, but it would help a user who's trying to differentiate the sounds. (I'm not personally having any issue with か & な, its just an example )