Automatic splitting of imports, dictation improvements

We pushed some changes and bugfixes early this week. Here are some of the improvements we pushed.

For quite a while we have received requests to make it easier to import long lessons without having to split them up manually. Also long lessons do suffer from poorer performance than shorter lessons just due to the larger volume of words that must be compared all the time. Therefore, we made the decision to automatically cut off all texts at 2000 words and create a new part or parts for the remainder of the text automatically. Give it a try and let us know what you think!

We also improved the Dictation activity by having the audio play automatically when moving to the next card, and on the answer screen making the Enter key send you to the next card. Therefore, you can simply type in your answer and hit Enter to operate the test now. Of course, you can still use Tab to replay the audio and the number keys to set the status before moving on.


Regarding to the performance I think shorter lessons are much more convenient. If I import lessons for myself, I prefer shorter lessons too.

There is a problem, if the original audio is longer and you share a link to the audio only. Have you considered this? Sometimes I get only permission to share a link to the audio only and not to import the audio to LingQ.

What happens to older lessons? Are they effected too?

What happens, if I edit an older lesson of someone else? Is it possible without causing problems?

I’ve checked now some of my German lessons where I’ve spend a lot of time for transcribing that are defenitely longer :frowning:

Ow! That’s what it was. I though my lessons were doubled and removed a part of new imports…

We don’t split the audio automatically, just the text. The full audio gets attached to both parts. You would have to manually split the audio if you wanted to match it up with the text and then re-upload it for each part.

As for existing lessons, they are not affected. Only new imports will be split.

This change (split imports) is great! A real improvement!

1 Like

I experienced this automatic splitting of texts a couple of days ago when I tried to upload a long text/short story with a 30 minute audio. I thought the system was hiccuping or something. You must have been testing this change before finalizing it, huh? In any case, I had to delete the lesson because this splitting of text would not match the audio. I finally managed to upload an entire course, containing fifteen lessons. I’m glad I did so before this final update.

This update will make it difficult to import audiobooks from sites, such as Librivox and Litteratureaudio.

Glad to hear it!

1 Like

Yes, we did have it up for a few days before announcing it just to make sure everything was running smoothly. In terms of audio importing, if you want the audio to match the text, you will have to edit the audio before uploading it or edit it and replace it once the text has been split. In fact, the long chapters from Librivox will have to be split and imported that way. This will make sure the experience for the user is better even if it takes the provider a little longer to import. That way hopefully people will be more likely to continue studying additional parts.

Mark, as I’ve mentioned above I’ve got only permission to share a link to the audio from some providers and not to import the audio to LingQ.

What happens if I have to correct a typo in a long text. Will the system allow me to keep the length of the text or will it try to split it too? That would make corrections much more time consuming. Or will you only check and split if the lesson is new?

Vera, this change only affects new imports. You can make changes to existing lessons.

What is the maximal length of a single non-splited lesson?

2000 words

Thank you. Good to know :slight_smile:

Is it just me or all the paraghraphs are merged into one when using this auto-splitting importing?
I imported a large body of text (~70 000 words) and the result was ~35 lessons where there are no new lines.
I like the fact that the splitting is automatic. I like the fact that the naming is automatic. However, without paragraphs (new lines) reading in a foreign language is challenging…

I’ve just noticed that, paragraphs and line breaks of the original text disappear when using this auto-splitting importing.

That doesn’t seem to be happening to us. Can you direct us to the text that you imported so we can try and reproduce the issue? You can send it to support if you like.

2000 words isn’t enough for me and many other members. I propose you double it. Related thread:

1 Like

I whole-heartedly agree.

We appreciate the suggestion but we are going to stick to our limit of 2000 words. The fact is this is a limit that works best in terms of performance which is why the change was made in the first place.