I've tried auralog, myet, instant immersion, an old copy of Syracuse software, and eyespeak looking for software that does a really good job of analysing when students say a word incorrectly. Some of these (not auralog or instant immersion) break the speech down into phonemes and show when a students says each phoneme correctly or not.
Most of them do a pretty good job when a student says a word correctly.
The problem I'm having is finding software that does a good job catching errors when they occur. Most of the software I've looked at will give a high or perfect score if the student says "ice cream" instead of "by train" or if the student says "horse" or "curse" instead of "course."
Does anyone know of software that can really pinpoint the correct/incorrect aspects of pronunciation?
any really accurate pronunciation software?
Moderators: Dimitris, maneki neko2, Lorikeet, Enrico Palazzo, superpeach, cecil2, Mr. Kalgukshi2
Try Pronunciation Patterns.
Hi, try Pronunciation Patterns at www.PronunciationPatterns.com/download.html which actually breaks sound files anyway you want to break them so that students can follow not just the whole sound but also each phoneme.
works?
Sorry to resurrect this thread--but I'm really interested in how these software based pronunciation programs work.
does anybody here have a technical understanding of what's going on? Does anybody just have hands on experience with them? Are they effective?
Probably more cost effective than having a real live teaching...that is assuming that they actually work and aren't just a gimmick.
does anybody here have a technical understanding of what's going on? Does anybody just have hands on experience with them? Are they effective?
Probably more cost effective than having a real live teaching...that is assuming that they actually work and aren't just a gimmick.
-
longshikong
- Posts: 88
- Joined: Mon Oct 26, 2009 1:49 am
Six years later, it seems they're not much better because the software compares your pronunciation not to a range of native speakers but to a single recorded sample of a single indiviidual. Even that person would struggle to get 100%. Furthermore, such software programmers neglect the fact that pronunciation is not fixed but varies according to speed and context.
Ignore the results data and go by what you hear.
Ignore the results data and go by what you hear.