Accent improvement: more natural-sounding tones

[Tamu]

Baron -

 

How much time do you spend work on your accent on a daily/weekly basis?

It's hard to separate a lot of pronunciation-only work from work on other areas. I spend a huge amount of time recently talking to many, many people I meet in cafes and restaurants, record their speech and mimic it later... so it's pronunciation work, but it's also hours of conversation practice, listening, word choice, vocab, etc. And I talk to lots of people about their impressions of accents, speech, language usage... which is obviously pronunciation-related, but also is just as much about culture and society. Plus, it's all in Chinese... so it all is mixed up.

 

But if I try to limit it to just pure accent improvement work on my own like analyzing speech patterns and singing along with the tone contours, as I wrote above, I'd guess I've put in somewhere from 150-200 hours in the last 2 months on that alone. I was a disaster at first and had to do 3-4 hours a day purely on tone perception and production, all the things I described in the post. Now I'm doing 1-3 hours a day on shadowing.

 

Meng Lelan -

I really don't know about your situation, maybe you've already tried this? But the first thing I thought on reading about your problems with rising tone is that if you can do a steady low and a steady high tone, why not say a fast low tone followed by a high tone... they end up blurring into a rising tone.

 

I played around now with the word "liáo" and got it to work for me:

 

- I recorded a nicely-shaped 2nd tone "liáo".

- I recorded myself saying "lǐ" ("li" in 3rd tone) and "āo" ("ao" in 1st tone). Using the visual feedback from Praat graph, I practiced saying lǐ -- āo quicker and quicker.

- I found I was able to get a pretty close replica of liáo. There was a small gap between the lǐ and the āo, but it's pretty small. It's still a flat-3rd-tone followed by a 1st tone, but at normal speech speed it sounds very clearly like "liáo".

 

I recorded myself saying "liáotiān" this way, dividing it into "lǐ - āo - tiān". It comes out sounding pretty clear. And no asphyxiation!

 

I really don't know if this could be helpful in any way, but thought to at least mention it.

[abcdefg]

@manjusri33 -- I think your membership is too new to be able to send and receive private messages. I'm out of town right now, but will try to send you contact information in a couple days.

 

[Meng Lelan]

@Tamu, thanks, I am going to try that. Something else too, when I get too focused and too concentrated on producing the right tones,  I start turning falling tones into rising tones and the rising tones into falling tones. 

[imron]

See here for previous discussion on speakgoodchinese.

[Tamu]

Meng Lelan - How did it go for you in trying to create rising tones through a quick low-then-high tone?

 

Btw... I was thinking about what you wrote. I'd definitely recommend Praat over SpeakGoodChinese (SGC) for what you're doing. SGC is pretty limited even in normal situations, and not enough for more than a basic word. Praat has a lot more options and might take a bit of fiddling with the settings, but the flexibility and power are really useful.

[Meng Lelan]

I am going to try Praat over spring break. I did try your suggestion and it did to some extent work, but what worked most was that I made myself stop worrying about whether I can match "normal" speech 100% of the time because when I do that I get really tense and can't even think about what I was going to say. Even in English my speech isn't normal 100% of the time anyway. 

[Kev]

Never say never.

I'm worried you may have internalised bad habits.

Have you tried minimal pairs with native speakers?

Would be great to have some youtube vids of you speaking Chinese and of using the software.

My minimal pair rave =

http://kmccready.wordpress.com/2013/10/09/learning-a-second-language/

[makochan]

 

 

So I used the computer to measure my vocal range in normal English speech and the vocal range of Chinese speakers who have more or less my voice register. Based on that, I plotted out where in my register a normal 1st tone should be, where a normal 4th tone must start, where 2nd tone starts and ends, the bottom of 3rd tone. Then I did voice training by using a computer piano to drill myself to sing that tone.

 

Tamu, don't know if your able to answer this, or maybe someone else.  But I'm really curious about this topic about vocal range.  Let's say that your normal voice register is around middle C to one octave above (C4 to C5).  Where would each of the 4 tones be roughly located? (i.e. around what musical note is first tone?)

 

And is this roughly consistent across different vocal ranges for different people?  I've heard one of my Chinese teachers say that if your a male and have mostly female teachers, you have to be careful or you might get into the habit of speaking Chinese in a (uncomfortable) unnaturally high register...and I've never seen any textbook address this type of issue.

[Flickserve]

I will be trying praat soon but this has been a very interesting thread.

[Flickserve]

I managed to play around with praat just now.

 

Suffice to say, it's a phenomenal piece of software.

 

I had a recording of a tutor saying 恐怕. I can't get it right in the first 15 attempts! The 恐 is always flat for me when I say it but not when the tutor says it. For the tutor, praat clearly shows a dip with a slight upstroke which I can't detect with my ear.

 

I tried another easier(!) sequence - 回头见. Mine is now much more similar to the tutor's shape but each word that I speak is 0.05seconds longer in duration compared to the tutor's version. Not sure if that makes a difference to how natural the accent sounds.

 

It is a bit difficult to compare as the tutor has a female voice and I have a deeper voice. I will have to prepare some male voice recordings to really compare better.