Random Review wrote:The problem is that, unless they are specifically trained voice actors, the pronunciation of a native is also adversely affected beyond a certain speed and that will apply even if you slow the audio down with Audacity or similar software; only that in order to be understood your pronunciation has to deteriorate in more or less the same way as that of the the natives does (and in the absence of an unbelievably good ear or feedback and guidance from a native or advanced speaker, it won't). I definitely found the speed of Glossika Mandarin a struggle (I only managed level 1).
As far as I can see, the only way to deal with this is to get some kind of feedback. The easy but expensive way would be to record yourself and have a teacher check it and give you feedback. The affordable but uncomfortable way would be to follow Mike Campbell's recommended method fully and include a stage where you record yourself and compare your voice with the files. I guess the recommendation is there for a reason. I know he's right about this, but I couldn't bring myself to do that. Initially it would be excruciating hearing my clumsy attempts.
At the other extreme, most courses will make you sound unnaturally "overpronounced" (to coin a new word) in your speech.
Unfashionable though it may be to say it, I thought the Mandarin FSI course gets it spot on. It's fast enough to sound kind of natural (in that strangely military way that all FSI courses seem to have), but slow enough to be easy to copy. The first few modules they also point out things like tones becoming neutral that other courses don't (and this omission is one of the most confusing things about being a beginner in Mandarin IMO- and it's unnecessary confusion at that!).
A speech recognition software like Dragon can give instant feedback: If the text produced by Dragon matches the transcription of what you said, you probably pronounced it correctly, and a human judge will probably agree. (More studies need to be done on this).
I think the transcription software technology has significantly improved and can be a tool in learning the correct pronunciation. I hope more users can try it or polyglots or linguistics or educators, prepare a study with college kids to check if it effective to improve pronunciation.
Speech to Text software will probably work with most European languages. I don't think it will work with Chinese characters or Arabic scripts, but I may be wrong.
When I mimic the German Glossika phrases, the Dragon software does not like it at all and gives me a different transcription. I get better accuracy by mimicking expressions from other sources like Assimil. I think the speed rate of the phrases is the problem for a German beginner.