menu
Tatoeba
language
Register Log in
language English
menu
Tatoeba

chevron_right Register

chevron_right Log in

Browse

chevron_right Show random sentence

chevron_right Browse by language

chevron_right Browse by list

chevron_right Browse by tag

chevron_right Browse audio

Community

chevron_right Wall

chevron_right List of all members

chevron_right Languages of members

chevron_right Native speakers

search
clear
swap_horiz
search
nava nava October 9, 2014, edited October 9, 2014 October 9, 2014 at 8:21:15 AM UTC, edited October 9, 2014 at 8:25:59 AM UTC link Permalink

Question/ discussion re adding Audio. In Hebrew (and perhaps in other languages as well ??) we can often read sentences correctly in more than one way. For example, one way is when we pronounce a word as female gender and another as male gender. Because today it seems that Tatoeba supports only one audio per sentence, when I record a sentence I need "choose" which way to read it, and the community misses out on the other ways to read it as well. They may think that this way way is the only way that is correct and other is incorrect, etc. Or they may never know there is another way of pronunciation for that sentence.

How do you suggest we can solve it?

1. Can we support more than one audio file per one sentence?

2. If not, another option is to write sentences in Hebrew with "nikood" (ניקוד ) to sentences, and then for each sentence add just one audio. This means the number of sentences in Hebrew will increase significantly. Also it would require a lot of work on adding nikood for past sentences. I can do it for my own sentences, but for all the rest it would be each for his/her own sentences. So this means a lot of cooperation and extra work for everyone. Another problem with adding nikood in Hebrew is that there are many words in Hebrew that are spelled differently with nikood vs. without nikood.

Given the issues with option #2, I'd recommend option #1.
What do you think?

{{vm.hiddenReplies[20624] ? 'expand_more' : 'expand_less'}} hide replies show replies
gillux gillux October 9, 2014 October 9, 2014 at 9:35:12 AM UTC link Permalink

Hello nava,

First, thank you for recording sentences in Hebrew. :-)

Allowing multiple audio files per sentence is a frequent request which is unfortunately yet to be implemented. There are only few developpers working on Tatoeba, and a lot of things to do.

From what I understand by reading the Wikipedia article about Niqqud [1], I think you shouldn’t add them since “In modern Israeli orthography niqqud is seldom used, except in specialised texts such as dictionaries, poetry, or texts for children or for new immigrants. ” On Tatoeba we should keep sentences as natural as they can be. We shouldn’t alter the authenticity of languages in any way, whatever the reason.

In the mean time, I think the best we can do is to let you record sentences in every way they can be read. This way, we’ll just add one of them, and add the others once we support multiple audio files per sentence. I don’t know how easy this would be however, maybe we could give you a list with alternative ids like 1234_1 1234_2 1234_3 for sentence 1234. How can we tell the number of alternative versions for each sentence?

[1] https://en.wikipedia.org/wiki/Niqqud

{{vm.hiddenReplies[20633] ? 'expand_more' : 'expand_less'}} hide replies show replies
nava nava October 9, 2014, edited October 9, 2014 October 9, 2014 at 11:16:21 AM UTC, edited October 9, 2014 at 11:17:56 AM UTC link Permalink

Thanks a lot gillux. It's my pleasure. I'm recording the sentences also for a spelling tutorial for the 1,000 most frequently used words in Hebrew that I'm developing for LerGO (nonprofit EdTech venture I'm working on). It's very nice to be able to build upon and leverage the great work that has already been done before by many contributors of Tatoeba.

I agree about your point re Niqqud. I'm now recording the other versions of audio as you suggested. I'm marking them mostly with sentence ID and at the end either F (for female) or M (for male). When Tatoeba supports multiple audio per sentence we can make use of these too.

CK answered that this is an issue that is not just for Hebrew but for other languages as well (e.g. different accents, dialects). So I hope it will be resolved soon :)

{{vm.hiddenReplies[20634] ? 'expand_more' : 'expand_less'}} hide replies show replies
Lenin_1917 Lenin_1917 October 17, 2014 October 17, 2014 at 11:50:50 PM UTC link Permalink

תודה על ההגייה שלך
:)

Selena777 Selena777 October 18, 2014 October 18, 2014 at 10:55:25 AM UTC link Permalink

Having two audios (male and female version) for the same sentence can cause some confusion with translations into the languages, that have difference for male and female speakers. For example, you have a Hebrew sentence with two audios (male and female) and somebody translate it into Russian.
In Russian there are different ways to say the same thing for a male speaker and for a female speaker in some cases. So, there will be a kind of confusion: which Hebrew audio matches with the Russian male sentence, and which one matches with the Russian female sentence. If there are two different Hebrew sentences, it would be much easier.

{{vm.hiddenReplies[20684] ? 'expand_more' : 'expand_less'}} hide replies show replies
gillux gillux October 18, 2014 October 18, 2014 at 12:15:40 PM UTC link Permalink

> In Russian there are different ways to say the same thing for a male speaker and for a female speaker in some cases. So, there will be a kind of confusion: which Hebrew audio matches with the Russian male sentence, and which one matches with the Russian female sentence. If there are two different Hebrew sentences, it would be much easier.

I don’t think it’s a good approach, we rather should make it clear we’re *not* translating spoken but only written sentences. To me, audio is just an extra help for learners, it doesn’t fixes the sense of sentences in any way. The rational behind that is that people should be able to translate without requiring to listen to every sentence in addition of reading it, which is just impossible since only a very few sentences have audio, and I don’t see that changing any soon, because typing will always outspeed recording. The problem you mention for Russian applies to any sort of information that can only be transmitted using voice: intonation suggesting a question or exclamation, little pauses between parts of a sentence…