Proposal talk:Record and upload sound clips in web application

Hi, please feel free to give feedback and to edit proposal as long as it follows the original intent of this project. Grenavitar 07:42, 15 August 2009 (UTC)

Mammoth

Supposedly this is a mammoth config which works on Ubuntu, but I haven't tested it yet:

-- The C compiler identification is GNU
-- The CXX compiler identification is GNU
-- Check for working C compiler: /usr/bin/gcc
-- Check for working C compiler: /usr/bin/gcc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detected version of GNU GCC: 43
-- Boost version: 1.38.0
-- Found the following Boost libraries:
--   date_time
--   filesystem
--   system
--   thread
-- Found OpenSSL: /usr/lib/libssl.so
-- Found TBB: /usr/local/lib/libtbb.so
-- Found FFMPEG:
/usr/local/lib/libavcodec.so;/usr/local/lib/libavformat.so;/usr/local/lib/libavutil.so
/usr/local/include
-- Parsing 'config.hpp.in'
--
-- Resulting configuration for :
-- -----------------------------------------------------------------
--  Will MAMMOTH be built as a shared library?  : Yes
--  Additional checks even in Release        :  No
--
-- -----------------------------------------------------------------
--
-- Configuring done
-- Generating done
-- Build files have been written to: /usr/local/src/mammoth

75.55.199.5 19:03, 20 August 2009 (UTC)

Also, "Make a symlink from vod_service to libmammoth-vod.so and you should be fine." -- http://sourceforge.net/mailarchive/forum.php?thread_name=4B140B39.2050708%40wichers.nu&forum_name=mammoth-users should help too. 99.27.201.47 22:03, 3 December 2009 (UTC)

http://rtmpd.com/ , Kamaitachi, rtmplite, or RTMPy are all better than mammoth. 99.191.74.229 21:31, 4 December 2009 (UTC)

Merged from Proposal talk:Audio offerings

>Pure audio can convey much information in relatively few bytes

Audio need much more bytes than text, even if the text uses more words. --Platonides 14:05, 7 August 2009 (UTC)

I'd be happy to see more audio on WP. It would also be great if we made use of Librivox and included excerpts of their audio books in the articles about those books and their author. Bu I'm not sure this needs the mobilisation of a significant part of the community which this idea of a 'strategy' implies to me. Is there an audio portal? An audio WikiProject? I think those would be a good start. --Bodnotbod 15:45, 13 August 2009 (UTC)

Well, I think, that audio in English and other major Wikipedias would be very useful. I can see a great use in the education of foreign languages, when you can here a native speaker and together learn it in the text. If there are then hundreds hearings from different articles - depends on the interests of learners - it help with strengthening their pronunciation. But! There are two technical problems. One is with article development. Every time it is very difficult to find out a version of article, which is recorded and more over for users who doesn't any wikipedia. The other problem is with quality of the recordings - usually these people doesn't have appropriate hardware for such a work.

The solution can come with MediaWiki development and local Chapters programs supporting by hardware and workshops people interested in recording.--Juan de Vojníkov 07:45, 14 August 2009 (UTC)

Accessibility of audio, listening and recording

I recognize that the choice of .ogg was driven by its being a free format in ways that .mp3 is not. I think the in-line players currently available help it to be accessible, but I also think it could still be easier and better supported.

If contributing to a wiki is intimidating, contributing audio is doubly so. To expand the audio offering, we should certainly do whatever we can to remove obstacles from recording, uploading, and reviewing audio files.

I have contributed a large number of audio pronunciation files to Commons and Wiktionary. (We need much more audio pronunciation. Here is a pronunciation rendered in IPA. Do you know what simple, everyday word this is? /ðiːz/ Wouldn't you find it easier to hear it pronounced by a native speaker?} Even with bot uploads and a program called Shtooka that allows mass recording of word lists, the process is pretty tedious.

I have lately played with the implementation for uploading audio pronunciation here: [1]. Things I like about it:

• The recordings are reviewed by other users.
• It is very easy to record, to change the recording, and the file is automatically uploaded and placed in the correct location.
• The word recordings are automatically connected to the words. Putting audio onto Wiktionary currently requires at least three steps: Audacity, Shtooka, or some other recorder; Commons; and Wiktionary.
• It's easy to advance automatically through the request list, and the software reminds users to check the requests.

Things I don't like:

• The recordings are not posted until they are reviewed.
• The site is specific to German and English, and it does not appear to be freely licensed.
• It is not (currently) clear to me how to change misspelled words in recording queues.
• This implementation probably would not be suited to recordings of longer articles.
• A word that is skipped is not subsequently excluded, meaning that it keeps asking for words that I didn't wish to record the first time.

--Dvortygirl 06:01, 12 August 2009 (UTC)

Yes, it is difficult. I use to record for Czech Wiktionary. But it wasn't easy. Usually I used other programs than Audacity and then converted my recordings from something such as wav to Ogg Vorbis. Than came a very tiering cutting of the recording to smaller recordings by words, uploading to Commons (at that time just Commonplace could be used - with bugs). Now it is much worse for me. As a poor student I have a "poor" computer. My microphone doesn't work. Other computer stopped to communicate with us and the recorders are quite expensive (about 4 000 CZK ~ 222 USD).--Juan de Vojníkov 07:45, 14 August 2009 (UTC)

Could we can do much about what users' computers can and can't do? Perhaps we could; microphones, at least, are not that costly. Perhaps a sound scholarship or audio scholarship would be something to look into, or simply some sort of technology stipend to help provide the equipment needed for worthy projects. We have many "starving students" helping out with projects, and I have heard from various people that they lack the technology (or financial means to obtain it) to contribute works they otherwise might.

I haven't yet explored this site in detail, but it appears that they have also had some success in creating a collection of multilingual audio pronunciations based on user contributions from all over the world. If we were to develop such an application, it would be worth looking at what sites such as these are doing well, and what they are doing poorly. I like that this one records geographical data for each user and links it specifically, since languages do vary greatly by region. This information is linked only loosely, if at all, to files we currently use. I also think there should be some sort of rating or feedback system built in, along with an easy mechanism for correction. I periodically get notes about recordings that have various problems, and it's not much fun to try to track them down and re-record them. —Dvortygirl 21:07, 22 September 2009 (UTC)

Pronunciation recordings in the major languages

As an educator from Wikiversity, I see that major gap is in recordings of word pronunciation for Wiktionary. On en.wikt it is Dvortygirl and others, but what about Spanish, German, French, Russian, Chinese and others - nothing. The recordings can get quality to Wiktionaries, but also help Wikiversity learners of foreign languages. So, this proposal would be very good, but there is missing technical support from WMF - we need software development and than programs animating people to do recordings supported by the help on the level of hardware lending.--Juan de Vojníkov 07:45, 14 August 2009 (UTC)

I think this could be forked into a project asking for a browser based method to record and upload sound files. If this all could be done in the browser (and it should be able to rather easily for wavs under 20 seconds and have them encoded server side) then I am sure many more people would spend some time to help fill out pronunciations. Grenavitar 07:25, 15 August 2009 (UTC)

Proposal:Record and upload sound clips in web application I have just made this proposal here please comment if you see any use to it... or even if you don't :) --Grenavitar 07:43, 15 August 2009 (UTC)

Un-merged

The Proposal:Audio offerings describes a very wide range of services, many if not most of which would be enabled by the ability to upload, but some of which, like text-to-speech (which was also merged with the audio offerings proposal) wouldn't necessarily be. 99.60.0.26 23:09, 24 August 2009 (UTC)

Impact?

Some proposals will have massive impact on end-users, including non-editors. Some will have minimal impact. What will be the impact of this proposal on our end-users? -- Philippe 00:15, 3 September 2009 (UTC)

You can split impacts into direct and indirect. Direct for viewers will be nothing, they will know nothing about this change. Direct for editors will be that they have a new interface to learn to record sounds... but they can also ignore it if sound is not what interests them. Indirect for viewers and editors will be the same... that is with it becoming easier to create audio content there will be more audio content -- any type that requires short spoken word such as pronunciation for Wiktionary / Wikipedia names or language books for Wikibooks will be greatly improved. This will definitely have a huge effect on the usefulness of Wiktionary and Wikibooks languages, and will be helpful for articles with foreign words/names in Wikipedia that don't yet have pronunciation clips. Making it easy for any editor of any language to upload a sound clip will be a big step from how it is now... that the editor must 1) know how to pronounce the words and 2) know how to record, encode, upload, name their content. Grenavitar 14:16, 8 September 2009 (UTC)