The content is about 7 Minutes but the transcript end at 3:45 Minutes. The rest is missing.
Could you please tell me where this item exists, in which category. Thanks.
It is under Type “Podcast”, Category “LingQ” and also it is under the newest items.
This is from my blog and that is all the text there is. I continued speaking freely in the podcast.
I imported that item from Steve’s blog. I searched the Library and found no one has shared it. Considering some one might need it and this would save their time, so I shared it, even the transcript is not complete. It happened in some of the Steve’s and others podcast too. To me, it didn’t spoil the pleasure at all.
But this is reminding me to expand it in text next time.
It would be nice, Edward, if you complete the text.
Vera and Edward,
I do not think it is necessary to complete the transcripts for all the texts in our library. There are some excellent content items in Spanish for example which have more audio than text. That is still useful, as Edward himself said. We cannot ask Edward to transcribe this content and we cannot follow up all content that is loaded up to ensure that it is all fully transcribed.
On the other hand Edward, it might be worth mentioning in the footer that this content is not fully transcribed.
By the way Edward, when you do upload other content, from other sources, I hope you always make sure that it is free of copyright.
From my view I’m disappointed when there lot of differences between audio and transcript. I think it should be mentioned in the description of the content.
It’s a point of quality. The great benefit of the library is the combination of audio and text. Text without audio or audio without text is easy to find but LingQ is good for having the combination. And it is good to have a good quality to get satisfaction.
But I’m looking forward to the changes in future when library is extended and content could be evaluated and commented.
I had a look at one Russian item where the interview was 15-16 minutes long (maybe even longer than that), but the actual transcript ended after 8-9 minutes. Somewhat annoying. On the other hand, it would take forever to save every unknown word if everything was transcribed (possibly several A4 pages of dialogue), and the day has only 24 hours…
Anyway, I would have appreciated if it had been mentioned in the description (just as Vera suggests).
I agree to Vera’s comment because the great benefit from LingQ is to have text and audio!
If that is not more (or often) the case I could listen to other material in the Internet.
The written form is for me very important because I am working easier with the eyes than with the ears. What I have seen is better in my brain.
I think that we should try to mention in the description whenever the transcript does not match the audio. However, it is time consuming for us to be constantly watching and editing content.
The content is free. We rely on our members to share content. We do not want to discourage them either. If the occasional item is not complete this is a minor annoyance to a member, a member who very often is not paying, and so I prefer to just leave things as they are for now.
When we have an evaluation system for content things will be better. Meanwhile, as Jeff says, the day has only 24 hours.
I should add that some users may enjoy listening to 15 minutes of something even if only half of it is transcribed. I agree that in an ideal world everything would be perfect. As I have said before, learning to deal with imperfection and incertitude is a very important to language learning.
I don’t suppose there is software that would speed up the transcription process? Voice to text translation, which we could then check and correct?
Helen, here are some youtubes I told about:
Microsoft Vista Speech Recognition Tested - Perl Scripting - YouTube - 10-minute demonstration, how a programmer guy tries to dictate a simple Perl script to the computer (Windows Vista). I was laughing as mad.
Windows Speech Recognition Macros Demo - Dictation - YouTube - this is rather successful demonstration how it’s supposed to work.
Windows Vista speech recognition - the whole story - YouTube - how it performed on the live show.
Steve, I understand that you don’t want to discourage members sharing content. But you should encourage them to do it right. It discourages me to add correct content when you say it doesn’t matter. I’m asking myself why I did great efforts to make correct transcripts! Other member’s only copy and paste text from other websites. They don’t thinking about if it is free of copyright and it is easy to copy and paste. On the other side I transcribed a lot of content by myself, what is a lot of work, and get the same points in award. I think a minimum of correctness in content should be and could be expected when members are earning points for content.
Maybe it is enjoyable to listen 15 minutes of something without transcript or without full transcript. But for such content I didn’t need LingQ. There are thousand or millions of podcasts in English without transcript. Therefore I didn’t need LingQ!
World is not perfect but shouldn’t we try to make a better world? Or should we say, it doesn’t matter?
If I have only a part with transcript there are some possibilities:
I shorten the mp3-File fitting the transcript.
I make the full transcript.
I mention it in the description of the content.
The content is free. But I’m a paying member and I pay for LingQ as a complete system and not only for some functions. The completeness is the great benefit of LingQ in my eyes and you lost this benefit if the content isn’t good. I’m interested in good and correct material and I expect a minimum of quality in content. I can understand when you mentioned that the day has only 24 hours. I don’t expect that you prove content before it is shared. But I expect that you didn’t say it doesn’t matter! And I expect that you react if a member indicates when there are faults in a transcript. For me does it matter and I’m sure for some other paying members, it matters too.
Steve, I really understand that there are faults in the library. You can’t prevent all faults. But I didn’t agree when you say that it doesn’t matter. You should encourage members not only to create content, you should also encourage them to provide good content.
I’m looking forward for the evaluation system for content. I think this is a really good idea.
I want to add that I enjoyed listening to the Russian audio file in its entirety although only half of it was transcribed. But I also want to add that I probably would have uploaded the interview in two parts (each fully transcribed), if I were the provider.
Back to my studies…
At the moment, voice to text doesn’t seem to be quite there. And, in fact, if there are a lot of mistakes in the transcript it can take almost as long to go through and edit the transcript as it does to transcribe it from scratch. However, it would be neat to try, at some point, to have some speech to text system create some transcripts and ask our members to review them and see how it all works.
For the future enhancements list!
You are right that it is much better to have matching text and audio. In an ideal world all content would be like this. At the moment, we simply don’t have the resources or the time to go through and identify or fix every little problem in the content. If people don’t like something, they can delete it.
We have to spend our time improving the site so that we can provide things like a rating system for content and special access for volunteer content editors. These features are coming. Please just be patient.
I have to write that Vera is writing exactly what I think too and what bothered me (…it does not matter).
I create the most of my items myself, think long about what can help the learners, record it, controlle it.
Sometimes are anyway faults in it and I am thankful to receive a tip for control.
But this all is a lot of work - not only copy and past!
Form a place that want to have the title “the best place for learning” I am expecting a good quality.
I think for me the words from Mark are suitable too - I have to be patient.
Corr.: From a place that want…
We prefer content to match. However, there is excellent content in Spanish and Portuguese for example where the audio is longer than the text. Many learners still can enjoy this content.
We do try to correct obvious typos and errors. The issue of content where the audio does not match the transcript is one that we will be deal with when learners can evaluate content and leave comments. The solutions is not, in my view, to remove content that some learners may enjoy listening to.
Ultimately the value of any content item depends on each individual user, and we will be finding ways to have users express themselves on this. Meanwhile, the content is free and users will delete items they do not like, or avoid those collections. On the other hand, the will go back again and again to collections they like.
As I often say with this kind of question. Please be patient. Also, not everything we do at LingQ can please all of the people all of the time. That is why we need to collect the views of a majority of users. We will be better able to do that in a while.
I should add that I consider the quality of the content that Vera and Irene have produced, and also contributed from other sources, for German, to be outstanding.