Hor roadennoù a zo didan lisañsoù Creative Commons a-bep seurt.Muioc'h a ditouroù
Ma vezit plijet gant an endalc'had-mañ, soñjit e-barzh ur roadenn arc'hant.
Lañser
I am happy to have found this site -- I tried a couple time to make something like this but don't have the technical skill. I have a few comments/questions. I suspect all have come up before, but didn't see answers in FAQ.
* is it possible to search the Wall?
* When looking at a sentence, can one see all the tags that have been applied to it? E.g. #3451959, [have you ever noticed how many near-anagrams 6 first digits in pi come up in daily life or is it just me?] I think I got there by following "Compound sentence" tag. Now, I'd like to know if it has been tagged by a French tag similar to "Compound Sentence", but I don't know/can't guess how a French speaker would tag it -- can I see all the tags applied to that sentence?
* I'm generally interested in longer sentences -- it it seems like an obvious feature to add to the "Advance Search" page (At least/At most X words). Is it not feasible?
* It seems to me like set-up with Shtooka tools is a hindrance, and the demo of how to get crystal-clear recorinds is a bit intimidating. I don't even own a microphone! So, I don't know if your goal is to have all recordings very clear -- I'd see a value in a variety of situations, like talking with other voices in the background, talking with a mouthful of food, talking with uhms, pauses, excitedly. (My interest is as a language learner, not an archivist). Have you hashed over this topic and decided a goal of Tatoeba is to have recording uniformly high quality? And would you be interested in easier to use audio-gathering tool at the expense of lower sound quality? (and probably also at the expense of attributing to a specific tatoeba user without a fair amount of programming.)
* do you solicit/accept "help pay for hosting" money contributions?
> I don't even own a microphone!
I'm using my webcam to record sentences for now.
Donations - https://tatoeba.org/eng/donate
Thank you!
* is it possible to search the Wall?
Not really, However, if you download the weekly export of Wall posts, you can search them offline if you know how to.
* When looking at a sentence, can one see all the tags that have been applied to it?
All tags on a given sentence are shown on the right side of the sentence's page.
* I'm generally interested in longer sentences
You can use the advanced search to find the longest sentences, though it's not possible to easily find sentences with "At least/At most X words."
Here is a pre-filled in advanced search form set up to find the French sentences with the most words that also have English translations. Just go here and input a French word or phrase, or skip inputting a French word and click the "Search" button to find the longest French sentences.
https://tatoeba.org/eng/sentenc...rt_reverse=yes
Thank you * 3
> would you be interested in easier to use audio-gathering tool at the expense of lower sound quality? (and probably also at the expense of attributing to a specific tatoeba user without a fair amount of programming.)
gillux made a proof-of-concept tool for recording audio a while ago when the question of using mobile devices for recording came up: https://tatoeba.org/eng/wall/sh...#message_33176 Was something like that what you had in mind?
What I envisioned was an independent web page, not so much to enable mobile as to get rid of the need for installing shtooka, naming files, emailing, etc. I was thinking of attracting sentences from casual contributors, including cell phone users without a PC.
Using the Web Audio API as gillux mentioned, I'd picture a simple site aping shtooka recorder. User would select the language to read in, and the web site would pick a batch of sentences that don't have audio in that language.
I haven't ever written a usable web site, so I don't really know what's involved technically. And a site that attained popularity might start to incur expenses -- but I guess that would be a good problem to have.
This is a project that does somethings like what you are suggesting.
https://commonvoice.mozilla.org/en
You can listen to some of their sentences to get an idea of what the sound quality of the files are like. You can also try recording a few sentences.
Intriguing. I wasn't aware of that project.
The recording UI is something like what I had in mind for Tatoeba to gather sentences. I like the 5 sentence at a time unit.
For a hypothetical flow -- a similar site could gather 5 sentences, using a stock of tatoeba sentences without recordings, create a zip of between 1 and 5 mp3s (depending how many sentences a user skipped), and generate a URL for the user to give you to incorporate. If users have to log in to tatoeba to give you the URLs, maybe it mitigates the possibility of flooding your in queue with junk.