Hacker News new | past | comments | ask | show | jobs | submit login
YouPronounceIt: Search for a word, get videos of YouTube speakers pronouncing it (youpronounce.it)
141 points by caiobegotti on Oct 9, 2015 | hide | past | favorite | 50 comments



Nobody mentioned Forvo yet?

http://forvo.com/about/

>Today, Forvo is the largest online reference for pronunciations with a database of over 3 million words pronounced in 325 languages – all created and maintained by native speakers.

>Forvo continues to work toward its mission as a go-to reference for authentic and veritable word pronunciations through over 116 million site visits annually and over 400,000 registered users.


I've not come across Forvo before, but it's UI comes across poorly in comparison to youpronounceit - a small search box and many steps between type a word and playing it - whereas the latter has a clear interface, plays the result quickly and provides context.

Forvo may actually be more useful, but youpronounceit is the one I'll return to.


I guess the link provided doesn't represent the page in the best possible way, http://forvo.com/ meets you with big input in the middle of the page, although they don't focus it automatically.



Humourously related is the Pronunciation Manual: https://www.youtube.com/user/PronunciationManual


My first thought was whether they explicitly blacklisted Pronunciation Manual!


These videos don't actually have (fake) captions so they shouldn't pop up (for the words in the title).


This is a brilliant research tool, because it can give you fast, unbiased answers about how English is spoken by real people instead of what's prescriptively correct. Any English dictionary will tell you that coup de grâce is pronounced with an /s/ sound at the end, but searching on YouPronounceIt it's actually very hard to find an English speaker who says it this way.


I agree that it's a useful tool, but I disagree that it's unbiased.

It is biased by the people or algorithms behind the speech-to-text conversion.

For example, if you search for "Yawk" you'll see one match, to a park ranger kidding someone from "New Yawk". This transcription was done by a human, which you can tell from how the text does "yeahhh", and how later text is elided or reworded for clarity. That is, around 0:40 the speaker says ".. unusual rock outcroppings. I mean they still attract people.." but the text says ".. unusual rock outcroppings that still attract people". Therefore, I think the transcriptionist used "Yawk" to highlight that the ranger was making a joke about differences in dialect.

That said, this bias may be low.

I did not know that about coup de grâce. Thanks!


I'm not sure what this really does. I mean it's interesting but as far I can tell it's just searching transcripts or caption data for the word and then queueing that part of the video. There's no checking if the word is actually pronounced correctly so you're not actually improving your pronunciation.

Several videos were not just the wrong pronunciation, the captions were wrong so the person never said the word the system thought was the requested word.

A good example is "candidate" which many people pronounce wrongly as "cannidate" Checking the results, first one said "cannidate", second one said something that sounded like "kennedy", 3rd one captions were wrong, 4th one was "cannidate" as well.


> pronounced correctly

If a word is so frequently "mispronounced" that you're likely to stumble frequently stumble upon a "mispronunciation" using this service, then it's not a mispronunciation at all.

Language is defined by how it is actually used.

Arguably, the fact that this service might give you "mispronunciations" is actually a point in its favour. This shows you real world examples of how a word is pronounced.

> A good example is "candidate" which many people pronounce wrongly as "cannidate"

Dropping the first /d/ in "candidate" is a common feature of colloquial American English. It might be proscribed by some, but it's not incorrect.


>If a word is so frequently "mispronounced" that you're likely to stumble frequently stumble upon a "mispronunciation" using this service, then it's not a mispronunciation at all.

Start thinking of "proper" pronunciation being a dialect (specifically, "Proper English"), and the objection comes back in force.


There's no such dialect as "proper English". There are some standardised prestige dialects.


I think we're in agreement here. If you take a descriptivist version of the original feature request, it's along the lines of "please do not show me pronunciations that are from a non-prestigious dialect."


Exactly. As learner of second/third languages, I can tell you that people lie all the time about how words are pronounced! If I ask you how "candidate" is pronounced, you'll pronounce all the syllables clearly. Then I'll ask you to explain the current candidates in the US presidential race and you'll start saying "cannidate"


Oh yeah, people seem to try to "properly" pronounce a word if you ask them. In the past, if you'd asked me to enunciate a world with an unstressed vowel, like "about", I'd probably say "a" for that first vowel even though it's always a schwa in practice.


I respectfully disagree. By your argument if I can get enough people to pronounce "cat" as dog then dog is the new pronunciation of cat.

There are correct ways to say things that people just don't know and would likely be happy to be taught. I had a friend that thought "turret" was pronounced "tur net". Another friend who thought it was "punkin" no "pumpkin". Both needed to and wanted to be corrected. It's disserve to tell people "come to this website to learn correct pronunciation" but then not have it actually be correct.

As another example are foreign words and names. There's no EE sound in Tokyo, nor Kyoto. There's no "awe" sound in "Pocky". The fact that westerners will go to great lengths to say "Nic ha wha gura" instead of "Nik ah wa gra" or "Ca pu chee noh" but won't spend the minimal effort to learn a correct Asian pronunciation is effectively racist. Some will say "but I'm saying it as it's spelled" but since they aren't following the same rules "say it as it's spelled" for European words it just proves my point.


> By your argument if I can get enough people to pronounce "cat" as dog then dog is the new pronunciation of cat.

Sure. If a large enough portion of English speakers pronounce it that way, it would be a valid pronunciation.

What other measure of validity would there be? There's no committee to dictate "standard English" (nor would what they say matter if the language actually used doesn't reflect their prescriptions), and dictionaries merely describe language as it actually is.

> It's disserve to tell people "come to this website to learn correct pronunciation" but then not have it actually be correct.

But the pronunciations offered on this site will usually be correct ones. It shows you the words as pronounced by actual people.

> There's no EE sound in Tokyo, nor Kyoto. There's no "awe" sound in "Pocky".

Maybe not in their native language, but there is in English.

> The fact that westerners will go to great lengths to say "Nic ha wha gura" instead of "Nik ah wa gra" or "Ca pu chee noh" but won't spend the minimal effort to learn a correct Asian pronunciation is effectively racist.

Maybe. People probably should be respectful of other cultures' own names if they can. Though that doesn't mean pronunciations distant from the native ones are incorrect in English.

> Some will say "but I'm saying it as it's spelled" but since they aren't following the same rules "say it as it's spelled" for European words it just proves my point.

Pronunciations which don't follow source language rules are certainly not limited to East Asian or South American words. English speakers aren't familiar with foreign orthographic conventions. That's to be expected: they are English speakers, not speakers of other languages.


Your response makes absolutely no logical sense whatsoever

By your logic if I see the word "latte" and I pronounce it "lat" because there is no "eh" sound at the end of words in English then that's perfectly fine and people should just deal with it. No, that's not what happens. People get corrected and through correction we all learn to say "lah-tay".

You seem to be arguing correction has no place. That's pretty indefensible. If we didn't correct each other no one could communicate. One person would say "ki nig et" and another would say "nait". One person would say "gar baj" and other way say "jar bayg".

Even subtle variations matter. If someone says "the dig caught the bil" when then meant the "dog caught the ball" no one will understand them even though the difference in pronunciation of of "dig" vs "dog" and "ball" vs "bil" is very tiny. In the same way say "Kee-yoh-toh" to a Japanese person and they won't understand what you're talking about until you give them many more clues.

I'd also argue correct pronunciations of names is partly of matter of respect. Maybe you're an exception but I kind of assume I called your name as "Tah-Zay Tee Su chi Nit Zel" you'd probably not respond. If you did figure out I was referring to you you'd probably correct me. If I insisted "so what, that's how I read it so that's how I'm saying it" you'd probably be offended. Maybe you're special and you wouldn't mind but most people would. I wouldn't respond to "jrej" or "guh-reg" or "guh-re-guh-guh".


> By your logic if I see the word "latte" and I pronounce it "lat" because there is no "eh" sound at the end of words in English then that's perfectly fine and people should just deal with it.

You are just a single person, so it probably wouldn't be seen as correct. However, if a large number of people within some geographical area did the same thing, then it'd be a (correct) feature of that dialect.

I never said that any feature of any person's ideolect is always considered "correct". I said that features in the English spoken by "a large enough portion of English speakers" are.

Languages aren't ideal systems that native speakers attempt to conform to, languages are simply what is actually spoken.(https://en.wikipedia.org/wiki/Idiolect#Idiolect_and_language)

> You seem to be arguing correction has no place. That's pretty indefensible.

No, I'm arguing that correcting people when they're not wrong has no place.


> No, I'm arguing that correcting people when they're not wrong has no place

Except you just argued that pronouncing Kyoto and Tokyo with an "ee" sound is okay and I just pointed out the real world consequences of not correcting their pronunciation, namely that native speakers won't be able to understand you. You'll probably argue English speakers don't need to interact with Japanese speakers. I'd argue if we just corrected the English speakers pronunciation everybody wins.


Let me know once you're done correcting all English speakers on how to "properly" pronounce Latin words then. Oh, go ahead and also make them speak "correctly" all the French words the English vocabulary has too! That would be super fun.


> Dropping the /n/ in "candidate" is a common feature of colloquial American English.

Pretty sure that's true of the first /d/, not the /n/.


D'oh! You're right, my tired brain screwed up. Fixed.


A quite humorous google-talk I stumbled upon: https://www.youtube.com/watch?v=hBpetDxIEMU&t=1233


It's humorous, but actually quite interesting as well. As a non-native english speaker I've learnt a lot watching it. Thanks.


This is fun way to discover random videos. I searched "invincible", and you get array of videos on topics from technology, cancer, health care to solar panels.


This is really nice. Although it got the first example I tried wrong in the first video (I tried "awry", and it came up with one that he actually said "a riot").

Much more useful than dictionaries that have pronunciation, since you can hear it in context spoken by real people. If some people pronounce it wrong, that's ok, because it gives you a feel for how it is actually used.


This could be incredibly useful for video editing - creating one of those "word collage" clips sometimes used in documentaries.

If you could filter the results by another keyword (within just the video title for example), you could easily use it to search within the words of a specific person and create mashup videos a la "Brian Williams rap"[1].

Of course it'd require a little bit of manual trimming of clips, but it'd get you 99% of the way there - provided the content is on YouTube and captioned. Maybe there are other cheap/free solutions out there for this kind of thing, but last I checked, there were a few transcript-searchable video services targeted at media production companies, and the subscription pricing was incredibly high if I remember correctly.

[1] https://www.youtube.com/watch?v=7CYJ73pVpVc



This is a fun idea and it's pretty clever. It would be handy if you could specify (or be informed about) which accent was used.

Also, the page does a nasty flash when it renders in Chrome.



Hahah, good point and great link! However, it might be possible also to include links to pronunciations in US and UK English.

Here is a decent example: http://dictionary.cambridge.org/us/pronunciation/english/ore...

Having said that, this is not quite what I was expecting: http://dictionary.cambridge.org/us/pronunciation/english/alu...


It would be nice if you could specify the part of speech also. For example the verb and noun pronunciations of associate and attribute are different.


http://youpronounce.it/search.jsp?q=quay

First video: American mispronouncing as "kay".

Second video: American mispronouncing as "kway".

Third video: Brit correctly pronouncing as "key".

Otherwise, amazing!


Merriam-Webster lists all three of those as accepted pronunciations [0]; like many things in English, there is significant regional variation.

[0] http://www.merriam-webster.com/dictionary/quay


Is the word "Quay" even used in US English?

https://en.wikipedia.org/wiki/Quay_(disambiguation) gives quays all over the UK, Canada, New Zealand, Australia and Singapore, but none in the US -- with the exception of a couple of tiny landlocked towns in Oklahoma and New Mexico called "Quay".

Are we allowed to call a mispronunciation a mispronunciation when it's only pronounced that way by people who have probably never encountered that word before?


Care to map the phonics path from quay to "key"? "kway" seems like the most reasonable pronunciation unless the letters are just meant to be strokes in a pictogram.


That's because what tantalor left out is that anglophones started misspelling that word in the 17th century, when it was actually spelled "key" from old French "kay".

So it's not so much a word that you go "ah I'm so clever I can pronounce it right", but more one that you go "some people along the way started misspelling this too often and the people who knew better didn't do a good job in correcting them".

And why should I know how to pronounce a word that current anglophones don't know how to spell? It's a tongue-in-cheek question, people.



This is great! I'm just trying to improve my pronunciation for recording a podcast and I'm spending a lot of time on dictionary.com. Hearing many people pronounce a word is great.

@cairo1982 some videos are listed as private, and some start playback at the very end.



Oh, but it's there!

https://youtu.be/fHxO0UdpoxM

It is however touted as 'an easy and fun way to improve your English'..

Village name marketing gimmicks aside, good tool.


You could also grab videos from Pronounciation Book Youtube channel: https://www.youtube.com/channel/UCT9uWfM5hvVlAPzl7KMlSvA


This could also help deaf people wanting to learn how to lip-read certain words -- because you are searching transcripts, you don't have to guess what the person is saying.



http://youpronounce.it/search.jsp?q=gif

1) hard g

2) hard g

3) hard g

4) hard g

5) I closed the tab

The concept and implementation are pretty neat, at least.


Genial! I've used AT&T text synthetizer before to learn to pronounce certain phrases and that works well too.


This is a great idea, with a great implementation and use of accessible resources. congratulations to the creators!


Apparently there were a few minor changes in the search engine and the results given




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: