Talk:Acoustic model

Untitled

This article is very focused in speech recognition, ignoring acoustic models in noise analysis (see the article 'Noise map'). It would be interesting to extend the scope of it. Cheers! — Preceding unsigned comment added by 146.83.216.144 (talk) 21:20, 9 October 2009‎ (UTC)

References

No references and has various errors. Needs 50% rewrite at least. History2007 (talk) 13:16, 7 February 2011 (UTC)

I agree. E.g. the desktop case suggests that the bandwidth limitation for speech recognition is in computer hardware. Yet speech is a human-to-human communication channel and the human ear is bandwidth limited to 20 Khz at best, 15 Khz in practice. Humans can even understand speech over the phone, which is bandwidth-limited to 4 Khz. — Preceding unsigned comment added by 37.153.240.6 (talk) 12:26, 9 September 2014 (UTC)

Proposed Deletion

This article has little value, even within the context of speech recognition, where the distinctions between Acoustic Model and Language Model are conventions with little encyclopedic value. The primary portion of this article is already well covered in the Speech Recognition article. Worse, the sections on various channels are dated, inaccurate, and largely irrelevant to the problems of acoustic modelling. --PaigePhault (talk) 16:55, 26 May 2013 (UTC)

I've removed the PROD tag, as this subject seems easily sourceable, from e.g. the sources found in this search. I agree it could do with a lot of cleaning up, but clean-up isn't a reason for deletion. I don't think it really looks like a merge candidate either, as it easily passes Wikipedia's notability guidelines. Perhaps you could reduce it down to a stub article and ensure that what is there is accurate and backed up by citations? And feel free to ask me on my talk page if you have any questions. Best regards — Mr. Stradivarius ^{♪ talk ♪} 13:03, 27 May 2013 (UTC)