Post a sample of your pronunciation here!

[mokoyo]

I'm surprised that you picked up the slight Cantonese/Hokkien. My mom's side of the family speaks Hokkien and I spent much of my early childhood in Hong Kong. I'm not translating directly, but sometimes those things bleed into my Mandarin.

I probably should have chosen a different selection as I listened to it before posting and heard the errors, but had a "Bah" moment. I'm currently "studying" this and erred in places.

Just out of curiousity, what are you comparing people's pronunciation to. Your own, other people's?

[陳德聰]

Hm? Not quite sure I understand your question. I usually just comment on whether people are matching the "standard" pronunciation of whatever the pinyin is for the various words they're saying or not.

The j~z~zh, q~c~ch, x~s~sh blurring thing is not a big issue for comprehension, but I noticed you do it a fair bit.

[LearneЯ]

Hey 陈德聪

It's probably the first time ever I am trying to speak Chinese and it has been the worst aspect out of reading, writing (typing), listening and speaking. The excerpt is taken from 画皮 of 汉语风 series. My tones are obviously messed up, but I would appreciate if you point out basic mistakes (e.g. sound substitution, speed, intonation etc.) as well. Other people are also more than welcome to critique my recording!

I have a cold and recorded it with my phone, so apologies for the bad sound quality. The text is quite simple, but I will put in under spoiler tag to see if you are able to understand me without accompanying text.

画皮是一个很久很久以前的故事，应该有几百年了吧。几百年前中国是什么样你能想得出吗？那时候没有汽车、火车，也没有电话、电影、电视、电脑什么的……那时候很多中国人不知道除了中国以外，还有美国、英国，更没有听说过什么是外国“留学生”……虽然现在的好多事情，好多东西那时候都没有，但是那时候有妖怪。对，妖怪。你相信不相信？

Mandarin Recording.mp3

[陳德聰]

It's your first time ever to speak Chinese? Look at you, priming with an opening like that to make me lower my expectations!

I listened without the text and I understood the whole thing. You don't have any obvious consistent pronunciation problems other than tone issues. Intonation and overall sentence flow seems okay, I only had three really stand-out moments:

1) "hěn jiǔ hěn jiǔ y[ ì ]q[ing] de gùshi"

2) "nà shíhòu méiyǒu qìch[e]"

3) "su[ í ]r[à]n"

At first I didn't really notice (1) because when I hear 很久很久 I just naturally assume 以前 is going to come after it, I guess.

For (2), before you said “hu[ó]ch[e]", I thought you were saying there were no reporters (记者 jìzhě) at that time. Even though you clearly said "qì", I think the way you made the word stressed-unstressed made it harder to get that it was a car until I had the context of the train after it.

Finally, I thought all was going well until (3) hit me in the face like a Russian accent.

Minor things like "nǐ-neng xiàngdechū ma?" were not huge barriers to understanding, but the ending of "ní xiàngxin-bu-xiàngxin?" was certainly amusing.

These look like they happened because of what looks to be a tendency to change whatever word you're stressing into a 4th tone coupled with a tendency to drop the tone on the second syllable of a word. Even though you asked for non-tone-related feedback, it seems to me your biggest (and it doesn't seem to be that colossal an issue really) is your tones.

How to go about fixing it? This is just a guess, but I think if you focused on your tendency to overstress the first syllable of keywords it might naturally make these issues go away.

[LearneЯ]

Thanks a lot!

Haha, yes, I would admit that I said "first time" as an escape-route from any inevitable embarrassment . But yeah, that is probably the first time of actually speaking, well, reading something aloud. I have yet to hold a conversation with natives or non-natives that goes beyond one sentence, but hopefully someday I will overcome my fear. I am in the same situation with most Chinese students who know English, but don't (or can't) speak it.

You are correct about the stress patterns and how I emphasise on the first syllable(s). I was trying hard to not sound like Microsoft Sam there, so I let some conscious and unconscious "slurs" in there: like yi3 qia(n)2，sui1 ra(n)2 where I didn't fully release the final n and 听说过 among others. I don't know about the 相信不相信 part, I tried it a couple of times and every time I somehow messed it up: xiang1 xin, xiang4 xin, xiang4 xin1.... you name it! I could never, for the life of me, could pronounce what was in my mind, guess I finally got the first one right and went into xiang4 xin in 不相信 part.

I am surprised that you were only perplexed by 虽然 pronunciation after 汽车. I found many amusing parts in there myself - every 那时 sounded off and 中国人不知道 was hilarious! 不知道 had the same problem, stressed 不 too much and left 知道 toneless, much more prominent in my last recording though!

I asked for non-tone related feedback because I thought (and still think) there are a few substitutions, but tones are obviously ingrained in words and just as important. I just listened to the recording again and my x is becoming more and more palatalised in the recording, almost reaching English sh [ʃ] in 相信不相信. Maybe I should do what I read somewhere a long long time ago before I even picked Chinese up seriously: pronounce x as s with tongue only touching your bottom teeth. I observed that phoneme later on, but it was not that widespread or apparent - most people only had this in specific words.

Thanks once again for taking your time and writing this thoughtful post, I will definitely work on them! If you some extra time someday, please have a look at my thread (linked in the first post) as well.

[陳德聰]

Er, I think you would be better off on working on the things that are barriers to understanding first, rather than working on your x~sh issue since [x] and [sh] are in complementary distribution and are inconsequential in comparison to your other issues. I'm sure you know tones are phonemic too, considering your understanding of linguistics?

If you want to focus specifically on your pronunciation (rather than your tone), you spend too long on the in your jiǔ, making it sound unnatural like [tɕi:u] when it should be closer to [tɕjəʊ]. Also your -ian endings are super pre-nasalized.

I'm a bit confused about how you appraise your own pronunciation and based on what criteria? Your palatalized x was pretty much a non-issue. Your basic ability to make native-ish consonant sounds is at the point where if your tones were correct people would be able to say "wow you have pretty standard Chinese". Your -iu diphthong needs work and your -an, -ian too. Aside from that, tones. It is actually impossible for you to improve your speaking if you don't improve your tones.

[LearneЯ]

Thanks again, that's where I was actually confused.

I did not know whether tones were the primary issue to focus on or the actual sounds -- in my current situation -- which lead me to post a sample here. Thanks for clarifying it, I will get back on track with my tones with the help of Audacity and praat. Praat, especially is an amazing piece of software for voice analysis. I will start with exaggerated tones in Microsoft Sam mode and then slowly build up my speed from there. I guess 慢速中文Slow Chinese is also quite good for shadowing purposes.

Glad to know I am not as bad as I thought and am able to produce native-ish sounds with mostly correct tones!

EDIT: Chinese breeze CD also has 慢速 folder in there, I should check it out as well.

Edited July 9, 2013 at 02:09 AM by LearneЯ

[陳德聰]

Preach! I use Praat to teach people pronunciation all the time, ;)

[tysond]

Interested to use Imron's method to improve pronunciation, so here's a recording at the start of the process.

Assimil 94 - July 25.mp3

I used the Assimil post from some time back (我今天早上起床的时候觉得有点儿冷。。。）

http://www.chinese-f..._40#comment-165946

[陳德聰]

Here is what that sounded like to me in pinyin (mistakes bolded):

Wǒ jīntiān zǎoshāng qǐchuáng de shíhòu juēdē yǒu diǎnr lěng.

Qíchér wàibiánr tiānqì hěn hǎo.

Wō hùrán fǎxiàn zuótiān wànshang chuānghu měiyóu guān hǎo.

There aren't enough words in this passage to get a really good idea of things you have a tendency to do, but the most jarring mistakes were the ones where you pronounced a first tone where it didn't belong. At this speed, it was still comprehensible. The sh -> ch may have just been my earphones, but I listened to it twice because without reading the text it wasn't instantly apparent to me what you were saying.

--

Totally unrelated, has anyone ever told you you sound like Philip DeFranco from YouTube?

[tysond]

Thanks, very helpful, thanks for taking the time.

Producing reliable tones is clearly an issue (as is self detection of the mistakes). Not sure if the sh/ch is a common error - I'll post a longer sample of text at a later stage.

Nobody has ever said I sound like Philip DeFranco - had to look him up. I learned a lot.

[Tamu]

Hi.

 

I'm trying to improve my pronunciation.

 

I'm studying on my own. I'd like to make my speech more natural. I talk with native Chinese speakers, but it's hard to get honest evaluations of how bad I am and detailed assessments of what to work on.

 

I'd appreciate any feedback on everything you think I could do to improve my accent, from very specific and detailed to general suggestions.

 

I used the short text that others on this thread have been using.

 

http://vocaroo.com/i/s0OQk7XHBWz4

 

Thanks.

[陳德聰]

Is there any way you can get a more clear recording?

 

I listened before scrolling up and checking what the actual text was, and the last bit was hard for me to understand. I think this is more recording than anything, because only the tone was coming through and I couldn't quite hear the consonants.

 

What it should be:

Wǒ jīntiān zǎoshàng qǐchuáng de shíhòu juéde yǒu yìdiǎn lěng. Qíshí, wàimiàn tiānqì hěn hǎo. Wǒ hūrán fāxiàn, zuótiàn wǎnshàng chuānghu méi guān hǎo.
 

What I heard:

Wǒ jīntiān zǎoshàng, qǐchuáng de shí~(t)hòu, juéde yǒu yìdiǎn lěng. Jíshí, wàimiàn jīngjì ... ... wǒ hūrán fāxiàn, zuótiān ... fàn, ránhù méi guānfǎ.

I think that was a combination of recording plus pronunciation. What I would guess is that the q~j issue is in your pronunciation and not the recording. However, the recording is probably responsible for tiān in tiānqì sounding like jīng. For the most part, your tones were good. But I think a better quality recording could reveal more.

[Tamu]

Hi 陳德聰.

 

Thanks for your feedback.

 

I'm sorry the recording wasn't clear. After I saw your message, I tried recording today in a few different places, but I realized that nowhere I have access to is a good recording environment; everywhere has strong background noise and/or reverb. Plus, my computer is a pretty poor recording instrument to begin with. And there are other recording quality issues as well. So it all makes it hard to get a very high quality, clean recording.

 

I've actually been using my computer quite a lot recently for recording. I've been working a lot on pronunciation, everyday recording people I come across (people in cafes and restaurants, mostly) and then later recording myself trying to mimic them. The recordings aren't great quality, but good enough that I'm able to use the computer to graph out and analyze the tones. I've found seeing 3rd tones in a sentence can be hit-or-miss because it's too low (for men, especially) and often creaky, so my system can't see it, but most other tones are pretty clear even on my low-qual set-up.

 

Here's a recording I made today. I tried to get it as good as possible, but it's not great, kinda sounds like a robot underwater lol. It does sound better on my system than yesterday's, and I just checked that it's at least clean enough to analyze the tones using the computer, but who knows how it will sound on your system. [add: I tried playing with settings to improve the playback quality, not sure if it made much difference though].

 

Hopefully you can at least hear a little better:

 

http://vocaroo.com/i/s0zxq0HgSCVj

 

If it's still not clear enough, that's that, I guess. But if you can hear well enough, I'd appreciate any feedback on how my tones sound. I've been studying Chinese for about 4 months, 3 of them in Taiwan, so I'm consciously mimicking Taiwanese Mandarin. I record myself on the computer mimicking what I record the native speakers saying and drill myself to try to get the graph of my speech to match theirs. I've found that when I memorize and practice thousands of times those sentences which I record native speakers saying, I can now get those sentences close to sounding kind of natural.

 

But when I make up my own sentences on the fly or read a text like this, it's still much less natural. My tones look ok on the computer and people understand (or at least seem to lol), but I'm still off. From analyzing and mimicking the speech of the people I record, I'm starting to get a feel for overall sentence intonation and how the same tone can be expressed differently, but I have a lot more work to do. For example, 1st tone is definitely a steady tone, but I'm just beginning to internalize different examples I've found of how its pitch varies quite a lot depending on the word's function in the sentence, the emotion, etc. I've been awed at the ability of native speakers to hear and say what seem to me to be miniscule variations in pitch level and contour of the same tone said ever so slightly differently.

 

And I've really grown even more massive respect for foreigners who've learned to the level of true native intonation. The more progress I make, the more I can start to really appreciate how much phenomenal work it must be to have achieved their level.

 

If the recording's good enough, I'd also be interested in issues with initials and finals. The q-j issue you mentioned is interesting. I tried to just speak normal and not change anything in this recording below; hopefully you can hear enough to tell me if it really is a problem and I need to aspirate more on the q's or if it was just the recording.

 

Sorry again for the poor quality recording studio lol and thanks again to you. Looks like you're the main responder on this pronunciation thread for a while now, so definitely want to express my appreciation 

[陳德聰]

I am very impressed with your progress and your determination and all those good things you described above Your tones and pronunciation are pretty spot on.

I think what you need to work on now is making your speaking sound more like natural speech. Right now you are saying each clump of phrases in a citational speech, which is not a problem for comprehension for the listener at all, but sounds stilted. The way you can go about doing this is taking a good look (and listen) at the phonology and prosody in everyday speech. You may have to get people to speak more naturally for your recordings, since a lot of the time this citational speech is what people will use to explain things to you.

 

The good news is you've already figured out how to handle 3rd tones, and this is just a more opaque phonological process that is specifically taught due to its importance. Other things you might want to look at are things like the tonal downstep that occurs throughout sentences and when the reference points for tones get "reset", where to put sentence stress, how to be emphatic in just the right places, etc.

This recording was better, but as you said, there is still some background noise. However, I won't re-transcribe it because I believe it will just all be correct.

[Orpheus]

Have always been fascinated with pronunciation and thought I'd post here to get some feedback.

Cheers,

O

Sample.wav

[Tamu]

Thanks for the feedback, 陳德聰. Everyone I talk to praises my speech, but they also praised me when I was at the "Ni hao" stage haha, so it's good to get objective feedback from you.

 

"Citational speech" is a good way to describe how I talk now lol. It's interesting you say that it's the form of speech people use when explaining things. A huge part of conversations I have with people is them explaining to me all the things I don't understand about Chinese culture, traditions, history, etc. So I definitely hear that form of speech more than cool, natural-sounding speech lol.

 

But it's not why my speech is stilted. Realistically, the main reason is that there's just too much going on for me to get it all right at once at this stage. I'm glad I've got my tones and pronunciation pretty spot-on now, but accurate pronunciation of each sound, control of each individual tone, overall sentence intonation contour, stress, emotion, etc, etc... there's a lot going on in even just a few sentences of speech, as well as learning to suppress every non-tonal-language instinct I have. Plus, there's all the non-pronunciation stuff: word choice, grammar, etc... not to mention thinking about what it is I actually want to say. My brain is overheating when I speak lol. Immersion is useful, but it has to be active: I'm trying to really listen, analyze, get a feel for it, and practice, practice, practice a lot more til I get it all internalized well enough to do effortlessly and smoothly. 

 

I've looked quite a bit at tonal downstep and resets. I've found these and some other things interesting while working on pronunciation recently. A lot of it I'd love to get your thoughts. This thread isn't the right place, I don't want to hijack it, so I'll start a new post in a bit to address it.

 

Thanks again!

[OneEye]

Everyone I talk to praises my speech, but they also praised me when I was at the "Ni hao" stage

 

You live in Taiwan. They do that to the freshly-arrived 阿兜仔 who hasn't learned anything other than 謝謝, and they'll also do it to the guy who's lived here 30 years, teaches conference interpretation and whose Chinese is indistinguishable from a native speaker's. It doesn't mean anything. They could just as easily be saying "it's nice out today, isn't it?"

 

You need to try to internalize the patterns and vocabulary you're learning to the point that you don't have to do all the thinking you're talking about. Say them over and over again with correct pronunciation, tones, intonation, grammar, usage, etc. (preferably imitating a recording of a native speaker speaking as naturally as possible) until you can't get them wrong. Then go out and use them with native speakers. Shop owners, waiters, members of your preferred sex at the bar, friends. That's how you fix stilted speech. You have to listen, listen, listen and mimic, mimic, mimic. Language is sound patterns, and the closer your production of those sound patterns resembles native speakers, the less accent you'll have and the easier it will be to communicate and integrate into society.

[Tamu]

Hi OneEye. Thanks so much for your thoughts. I've read and gotten benefit from many of the posts I've found of yours when I've been researching different Chinese-language issues recently, so I appreciate your taking the time to comment about my post.

 

You're very right that Taiwanese will laud every level of Chinese from "ni hao" to simultaneous-interpreter. That's what I meant by saying that they praise everyone so I was looking for the objective feedback of the forum. I posted my pronunciation here because I wanted to get objective, critical feedback from the experts of the forum, as opposed to the blanket praise and generic comments from non-linguist native-speakers.

 

Your idea to listen, imitate, internalize, and use the language is very solid. Here's a quick summary of what I've been doing to improve my speaking skills. What do you think of this? I mentioned above that I've been recording people's conversations. I've been getting around 30 hours each week of in-person one-on-one conversation for the last few months. Private tutors are a third to half of that; the rest comes from random people I meet in cafes, restaurants and from the internet. I've been recording the conversations. Afterwards, I go through the recordings, drill new vocab and constructions, and analyze their speech. I pick out interesting sentences, snip them out, and then practice mimicking each sentence hundreds to thousands of times. I want it to sound the same, so I record and listen to myself of course. But I know that my ability to pick out small variations is still not precise enough so I use the computer to analyze and try to get the graphs of my tone levels and contours to exactly match theirs. Seeing visually what's going on is incredibly helpful, because I'd otherwise just be flying blind about very subtle differences.

 

I don't have enough time to do it right and do hundreds of sentences thousands of times each day, so I just do all of them a few times each and then concentrate on a few sentences which I repeat hundreds of times every day. It's really slow work to get each little part to match exactly in pitch level, contour, volume, etc. But with practice, I'm getting a little better. I find that after practicing one sentence a few hundred times each day like this for a few weeks, I'm now able to get to a point where I can get that particular sentence to pass for native. [To test if I'm good enough, I've done this a few times: Record myself. Alter the voice so it's not obviously me. Throw that recording into a group of recordings I collected of both native speakers and of foreigners at various levels of Chinese ability saying different things. Then play a game of asking people to identify who's foreigner and who's native. It's not a perfect scientific experiment, but it seems good enough to give me confidence that I'm doing these massively-practiced sentences well.]

 

But on sentences I'm not mimicking, in other words those sentences I make up on my own on the fly, I'm far from natural-sounding, let alone native. I spent time at the beginning doing work with a virtual tuning fork to learn my different pitch levels and where all tones should be, and I used a virtual piano to teach myself to follow the pitch contours of native speakers with my vocal register, and in particular to maintain a stable 1st tone (I have horrible music ability and can't sing at all, so these basics probably took me way longer than it would for normal people). Also spent time getting each sound to a passable pronunciation: aspiration, voicing, etc. It all seemed to work. When I record myself speaking (on the fly, not pre-memorized and pre-practiced) and then analyze later, I can see that my tones are now all very clear. And it's nice to get confirmation here that my tones and pronunciation are all correct and spot-on. 

 

But as 陳德聰 says, my speech is "citational", stilted. When I analyze my recorded on-the-fly speech, all tones and pronunciation are usually good, and my speed of speech is decent.  But I'm still not at the level yet to be able to also build in overall smooth sentence intonation, flow between phrases, etc... hence, stilted. I definitely have brain overload when I'm speaking. It's still too conscious, not completely internalized yet.  Vocab and grammar are coming along, I'm trying to keep simplified and traditional at the same level (I'm signed up for hsk5 just for fun to gauge myself and push myself to not neglect reading speed in simplified), but I'm just not able to get it all together to speak on the fly in perfectly smooth sentences.

 

I'm now able to mostly understand talk shows, cheesy TV programs and some documentaries, so I've been splicing them up and adding them into the mix of sentences I vocab mine, analyze and mimic. This was the only change I've made in the last few weeks. My plan was otherwise just to keep on with this regimen and build on this base.

 

I'm not in any Chinese program; I just studied a few weeks overseas on my own, then been in Taiwan about 3 months. As I wrote above, I'll make a separate post about pronunciation-specific techniques and questions I've come across. But I wrote this here because I really appreciate your taking the time to comment on my post and I'd appreciate any insight you could give about how I can improve what I'm doing!

 

Thanks!

[OneEye]

I think if you've made it to the point that you feel ready for the HSK 5 after 3 months and change, I probably don't have anything to offer you.    Keep it up.

 

On second thought, you might try shadowing to work on your fluency. Just make sure the recording you're shadowing is of an articulate native speaker speaking in a natural manner, not in the ridiculous stilted way you hear on most language textbook recordings.