Listening to Audiobooks

[phills]

On 1/18/2022 at 6:16 PM, alantin said:

On 1/18/2022 at 3:38 PM, carlo said:

Not sure about the copyright situation, but for 三体 see here: https://www.youtube.com/watch?v=tSzhwT7Rzzo

 

Yeah, I know this one... But I passionately dislike the narration.

 

I just started listening to that one.  I like that it's on youtube so I don't have to worry about it disappearing halfway through

 

Curious @alantin why do you dislike that narration?  I haven't listened to enough different versions to developed a sense of what's good and what's bad yet. 

 

So far, I don't like ones with overly loud background music or sound effects because it makes it hard to hear the words.  A little is ok, makes the narration more interesting.  I also prefer if male / female voices have separate voice actors rather than a single person trying to mimick both, but that's not that huge a deal.  Any other peeves?

[MoonIvy]

I prefer the fully voice acted, sound effects, background music ones. But I do agree that this can be distracting if you're trying to listen to the words. Since you get to a more advance listening level, you may grow to like these. 

[alantin]

On 1/20/2022 at 7:15 AM, phills said:

Curious @alantin why do you dislike that narration?  I haven't listened to enough different versions to developed a sense of what's good and what's bad yet. 

 

So far, I don't like ones with overly loud background music or sound effects because it makes it hard to hear the words.  A little is ok, makes the narration more interesting.  I also prefer if male / female voices have separate voice actors rather than a single person trying to mimick both, but that's not that huge a deal.  Any other peeves?

 

Actually either way is fine with me. I haven't listened to many audio books with voice acting, background music or sound effects and I also am not very accustomed to them. That seems to be a lot more common in China than it is with English audio books.

 

My main beef with the youtube version was that I just couldn't reconcile the narration with the characters in my head. The rhythm is just wrong for me! Though I think if you can listen to that speed, you should be able to listen to anything, so it should be pretty good listening practice in that sense. ?

[phills]

On 1/20/2022 at 10:58 PM, alantin said:

My main beef with the youtube version was that I just couldn't reconcile the narration with the characters in my head. The rhythm is just wrong for me!

 

Yea I'm through to episode 4, and I get what you mean.  The narrator takes weird pauses in the middle of sentences, or between sentences.  And then shoots off at high speed until he reaches his next scheduled pause.  I think he's trying to be dramatic, and sync his pauses to when there's a shift in the action, but he seems to be over-doing it. 

 

I just noticed it and it's not too annoying yet, but we'll see if I get tired of it by episode 10.  It does make it more of a challenge to listen to.  I've had to glance back at the text a few times already, because I got too lost.

[phills]

@alantinAfter you sensitized me to the uneven cadence, it started bugging me more.  I found another version on youtube:

 

https://www.youtube.com/playlist?list=PLrjIk2pnoftD-qYNsDJyMYTIN19oBpTCV

 

It's got a more even cadence, and I find the accent easier to understand.  I think I can understand 10-15% more of the narration in this version  

 

I'm going with this one for a while.  I might swap back to the other one every once in a while just to see the difference.

[Publius]

Yeah, 哈哈笑's cadence is too dramatic. Many people dislike his overacting. 用力过猛 is a phrase you see a lot.

 

Incidentally, Luke Daniels, the English narrator for Three Body book 1, has the same problem. I don't have much experience with audiobooks, but I think Stephen Fry is waaaaay better.

 

Listening to audiobooks is usually the surest way for me to go zzzZZZ. So I'm very grateful to you guys for pointing me to YouTube, where I found Harry Potter in Japanese with synchronized text, and every kanji reading is marked to boot. Cheers!

 

 

[MoonIvy]

On 1/21/2022 at 9:55 AM, phills said:

It's got a more even cadence, and I find the accent easier to understand.  I think I can understand 10-15% more of the narration in this version

Before when my listening wasn't that good, I use to find deeper male voices to be easier to understand. I just listened to a bit of that version of 三体, the narrator's voice is pretty deep. Maybe deeper voices make words clearer? 

[alantin]

On 1/21/2022 at 11:55 AM, phills said:

@alantinAfter you sensitized me to the uneven cadence, it started bugging me more. 

You're welcome, lol ?

[alantin]

On 1/21/2022 at 11:55 AM, phills said:

I found another version on youtube:

 

https://www.youtube.com/playlist?list=PLrjIk2pnoftD-qYNsDJyMYTIN19oBpTCV

 

It's got a more even cadence, and I find the accent easier to understand.  I think I can understand 10-15% more of the narration in this version  

 

Wow! This is exactly the version I've been listening on Ximalaya and now on 微信听书! I just loved it immediately!
Awesome that someone has uploaded it on youtube!

It has small changes to the actual book to make it flow better in that format though, but I don't really mind it. It is great listen!

[phills]

On 1/21/2022 at 8:29 PM, MoonIvy said:

Before when my listening wasn't that good, I use to find deeper male voices to be easier to understand.

 

Deeper voice can help, like James Earl Jones used to do a lot of movie voiceover because he's got that smooth deep voice.  But mostly a more even pace, and more even volume helps understanding.  Less extremes in loud/soft, fast/slow.  Perhaps that makes it more boring for a native listener -- more like a corporate voiceover or elevator/musuem voice, with less emotional content.

 

I also listened to a robot narration for kicks.  Wheel of Time on ximalaya. 

 

https://www.ximalaya.com/sound/449872565

 

I tried it at 75% speed since it was obviously sped up.  It's not as terrible as I thought, but I think it must be torture for natives.  You can see the first comment on there:

 

"这是AI声音吗？如不是AI是真人的声音，为啥要搞成这个样子呢？买了《时光之轮》全套书，没时间看的时候好想听啊，可这AI声真是让人听不下去啊！"

 

I don't want to learn to speak like a robot, so I didn't continue after a few minutes.  But that's so even it might actually help you learn to distinguish syllables.

 

[alantin]

On 1/22/2022 at 7:24 AM, phills said:

I also listened to a robot narration for kicks.  Wheel of Time on ximalaya. 

 

https://www.ximalaya.com/sound/449872565

 

I tried it at 75% speed since it was obviously sped up.  It's not as terrible as I thought, but I think it must be torture for natives.  You can see the first comment on there:

 

"这是AI声音吗？如不是AI是真人的声音，为啥要搞成这个样子呢？买了《时光之轮》全套书，没时间看的时候好想听啊，可这AI声真是让人听不下去啊！"

 

I don't want to learn to speak like a robot, so I didn't continue after a few minutes.  But that's so even it might actually help you learn to distinguish syllables.

 

 

That's actually pretty good!

I usually hate listening to AI reading things, but after listening to that for a minute, it that seems to have quite natural pauses!

 

[alantin]

@phills, I got interested so I checked the reading speeds for a few chapters of those AI recordings.

 

It seems that the 1x speed is a tad over what they say is a normal speaking speed (250 cpm) and 0,75x is bit lower. There is interesting variation in the speed between the chapters. I'd expected the speeds to be more consistent for an AI but that's probably because of the cadence.

 

[phills]

Thanks for the research @alantin.  Looks like it's not obviously sped up.  270cpm is within the range of other audio books.  The Narnia audio books I started with were at about 270 and I felt that was pretty comfortable. 

 

And maybe the reason I thought it was "comprehensible" was because I was listening to it at a slower speed

[carlo]

On 1/21/2022 at 12:31 PM, Publius said:

where I found Harry Potter in Japanese with synchronized text, and every kanji reading is marked to boot

sorry for the slight off-topic but do you have a link for that?

 

Narrator preferences can be really personal. At first I wasn't fond of Roy Dotrice's version of the Song of Ice and Fire books but then got hooked and ended up listening to all seven volumes...

 

[Publius]

@carlo Book1 https://youtube.com/playlist?list=PL0kjvGfEjIFHAjV8S-X7Vm9HW3Mv3VVkC

 

Book2 currently at chap 11 https://youtube.com/playlist?list=PL0kjvGfEjIFHXXH2nyZZNa6dL3E3tBS9T

 

[MoonIvy]

@phills I think that AI might be the one that 微信读书 uses but a spread up version. I actually sometimes use the AI (female voice) on 微信读书 to read some of the chapters out to me while I'm busy doing chores. It's not so bad, obviously not perfect, it sometimes does make mistake. It's a pretty popular feature on 微信读书, lots of people use it. I wouldn't recommend it though for listening practice, a human recording is by far better. AI reading is only good when your Chinese is good enough so you can tell the mistakes, else you might learn the wrong prouncation for certain words.

[phills]

Now about 40 hours in.  My progress / observations:

 

1. My comprehension improvement seems to have plateau-ed a bit.  It seems hard to raise the percentage of content understood, just by pure listening.  Mine seems to be stuck at 70-80%, for the last 20 hours or so. 

 

As a result, I've decided to glance back at the text more, maybe once every 2nd or 3rd session.  The feedback from the text seems to help me improve my comprehension rate -- faster than just puzzling out sounds like a child.  But I'll need to try this out further to be sure.

 

2. However, my ability to not get lost in the flow of Chinese audio seems to keep improving, across content types.  I've tested out my ability on some other stuff:

 

a. At 25 hours in, I tested myself on sequels.  These are books I haven't read, but are sequels of books I have read. 

 

I listened to the first chapter of the Lion Witch and Wardrobe, and was able to follow it.  I estimate I understood 10% less than I did of The Magician's Nephew -- ~70% rather than ~80% of Magician's Nephew.  That's good enough to follow the plot, and if I continued to listen, I bet it'd slowly go up. 

 

This also worked on the first chapter of a Foundation sequel that I haven't read in Chinese.  In each case, knowing the general plot already and having the same narrator, same linguistic style really helps.

 

b. At 25 hours in, I also tested myself on a podcast.  I just happened to pick this one, about the Northwest of China:

 

https://storyfm.cn/episodes/e592/

 

It turned out to be a hard one.  I got utterly lost. 

 

I tried it again 40 hours in, and this time I was barely able to follow it.  But comprehension was poor, on the very borderline of understanding, around 51% or so.  I'd rate this an Incomplete -- I'll stopped after 15 mins, and will try again later.

 

c. At 33 hours in, I tested myself on an anime movie on Netflix, Nausicca of Valley of the Winds. 

 

I tried this 6 months ago, and had trouble even with subtitles, because the subtitles didn't exactly match the words, which made it very confusing and energy draining.  I'd swap between the 2, then it would end up being too fast to process.  Instead of watching it in English, I decided to wait until my Chinese got better.

 

This time, I turned off subtitles and only used the Force, and was able to watch the movie and understand what's going on.  Comprehension was around 70%, so enough to understand the plot, but not catch everything. 

 

d. 40 hours in, I tried to watch an episide of a modern TV drama without subtitles.  This actually worked, something that I couldn't do 35 hours ago

 

I then tried a historical TV drama without subtitles.  Got horribly lost.  Needed to read subtitles to understand the plot. 

 

3. At this point, I seem to be able to listen to translated English works more easily than native Chinese works. 

 

E.g. I found Narnia easier than 活着。Foundation easier than 三体。I read these at around the same time last year.  活着 surprises me -- it's narrated with a deep smooth voice, with a slow-ish even cadence, and should be easy but is not quite as easy as I would have thought.

 

It reminds me of the pattern with reading too.  When I started reading, I found translated English works easier.  But after reading more, I ended up finding native Chinese works easier, particularly mass market works like 余华 or 古龙. 

 

We'll see if this pattern holds up in listening as well.  At some point, will native Chinese works be easier to listen to than translated English works? 

 

4. The main challenge for the next 40 hours will be getting my comprehension rate up.  I have some ideas that I'll be testing out.

[Publius]

I find non-fiction audiobooks significantly easier than fictions. I just downloaded a whole bunch of English audiobooks. Foundation and The Hitchhiker's Guide I can follow no problem since I've read them more than once. Not so lucky with Dune. And Artemis Fowl lost me completely. Then I turned to A Short History of Nearly Everything and Sapiens A Brief History of Humankind. What a breeze, even though I couldn't catch everything.

[phills]

Sapiens is a good one.  I read that 5-6 yrs ago, and wouldn't mind re-reading.  I'll see if I can find a Chinese audio track for it.

[Publius]

嗯 and may the force be with you!