• It's not very pleasant to create audio right now (recording, then hacking into pieces, exporting to mp3, with some audio editor like Audacity, then uploading the audio at the word cards themselves).

    There will be an updated tool as part of the upcoming wiki, allowing us to simply record audio here on memrise itself. I will wait until that's in place, then see what I can do.

    Posted by ikenaiAndi 4/15/12 (2 years ago)
  • @ikenaiAndi, the audio uploading process isn't perfect, but you actually only have to recod the file and export it as one long mp3 file and then send it to me (via email or dropbox) and I can auto slice it and mass upload it, as long as the words are in the same order as they appear in a wordlist. Which does actually make it pretty easy for you. Let me know if I can help at all with explaining how to do it !

    Best wishes

    Ben

    Posted by BenWhately Staff   4/16/12 (2 years ago)
  • Ahh, I know. I've read the FAQ. Sure, this is a lot better than the process I described above, but it's a band-aid fix and after recording some 100 words, I will end up with a some megabytes large file I have to transfer.

    With a tool to record in Voocaroo-esque manner, I will be a breeze to just add audio when I have the time. I can do 20 or 50 or 100 and it would be very convenient.

    I would love to contribute audio here and will do so, because audio is a HUGE help, especially for a language with complicated to read and pronounce compound words like "Verwaltungsrechtsverfahrensordnung". ;)

    Posted by ikenaiAndi 4/16/12 (2 years ago)
  • (Ha, and imagine me, a native speaker, to fail to say "Verwaltungsrechtsverfahrensordnung" - it happens a lot! __)

    Posted by ikenaiAndi 4/16/12 (2 years ago)
  • (among a list of 200 words I'm recording - what to do, start over?__)

    Posted by ikenaiAndi 4/16/12 (2 years ago)
  • yes, that process is still not ideal, but it is ok for the time being. But we will get a much better in-browser recording option online soon.

    Thanks

    Ben

    Posted by BenWhately Staff   4/16/12 (2 years ago)
  • I'm looking forward to such an easy option. :-)

    Posted by Zirien 4/16/12 (2 years ago)
  • @IkenaiAndi I record sets of about 100 words in one go. If I make a mistake I pause, then say the word again. It's quite easy to cut the wrong pronunciation using an audio processor. I also usually leave ten seconds of silence at the beginning to give Audacity's 'remove noise' filter something to analyse.

    Posted by gloda 4/25/12 (1 year ago)
  • 1st one done, more to come soon, I promise

    Posted by hathacline 4/30/12 (1 year ago)
  • I agree with Ikenai, audio words are a huge help. I'm learning a 6000 vocabulary japanese course, there are just 50 audio files overall. Question: Anyway there are computer voice programs that can read every word correctly? Memrise could install them?

    Posted by Celicity 5/1/12 (1 year ago)
  • Nope. The best ones around can't produce natural speech. Some come pretty close, but not close enough.

    You're learning Japanese, which is an even more complicated thing. Depending on which syllables follow which, sounds shift from one to another.

    As an example: 島 しま island 国 くに country 島国 しまぐに island country

    Here, the KU of Kuni changes to GU. Computer voices do recognize this is some cases, while in others, they treat compound words like their single parts stuck together. しまくに is not しまぐに unfortunately. You wouldn't want to learn wrong pronunciation "because it's better than nothing". I think nothing is better than wrong in this case, as you can produce the sounds for this yourself when you've become a bit more familiar with the phonology of the language.

    Also, Japanese words have a different pitch (up/down of sound) depending on regional dialect. No computer voice for Japanese I've checked so far gets Standard Tokyo Dialect right. I can not speak for any other language, since I don't learn any, but I do know that German, English and Russian have stress (instead of pitch), not indicated by the written form. Other languages, like Mandarin, Cantonese, Vietnamese etc have tones, but for the Chinese languages, the Pinyin are actually helpful - I can imagine, that in such a case, a computerized voice could be of some worth. I know however that computerized German, English and Japanese sounds awful.

    Posted by ikenaiAndi 5/1/12 (1 year ago)
  • Do you know that old internet joke "I have no idea what you're talking about, but here's a bunny with a pancake on its head"? That's basically the same thing as putting up AI voices as a replacement for "the real thing".

    Here it is :) http://www.flickr.com/photos/hilarious/252849242/

    Posted by ikenaiAndi 5/1/12 (1 year ago)
  • oh, okay. I see. Robodl made a specifically list, for German speakers. Memrise could that do: Making lists for the missing audio files. Categorized them for the right native speaker. Maybe things would getting easier to search.

    Posted by Celicity 5/2/12 (1 year ago)
  • I would love to help out with the audio files too - although I do not have a lot of time and thus would somehow prefer to wait till there is this "easy" option everyone is talking about __"

    Posted by neko-chan 5/5/12 (1 year ago)
  • I would also be interested, I also wait for the easy option:)

    Posted by TL-Pandaling 5/6/12 (1 year ago)
  • Guys, please help :) It seems that the easy option for uploading audio is not going to appear any time soon. I think that phrase "easy option" have misled you. Current solution is also quick and easy. It's just not as convienient. Vielen Dank im Voraus!

    Posted by ponchoosh 7/18/12 (1 year ago)
  • I made some German recordings for the courses http://www.memrise.com/set/10020921/montag-morgen-2/ http://www.memrise.com/set/10036490/-549/ in the ドイツ語 topic (German for Japanese people). Finished both.

    I must say, the sound uploading procedure is still a major pain the ass and I'd like to have a proper system in place the day before yesterday.

    Posted by ikenaiAndi 7/20/12 (1 year ago)
  • No way, example sentence testing is way more important. It will solve so many problems. We can already add audio, and it's usually not necessary with many languages after learning pronounciation rules.

    Posted by Robodl95 7/20/12 (1 year ago)
  • Example sentences are a cool thing, but think again. They need to collect hundreds of thousands of those.

    A recording feature is fire and forget.

    Both are way more important than stuff like quotes, score screens and other fluff.

    Posted by ikenaiAndi 7/20/12 (1 year ago)