Importing audio files with subtitles

Support & Feedback Forum
1

I could do that with previous importing UI but, I don’t know how to do it with current importing features.

Is it gone?

There was a all-in-one style importing where you could upload texts and audio together.

Why was it removed? I wish it could be restored back.

2

This is my complete process:

  • Click on “Import Lesson”.
  • Click on “Text Files & Ebooks”.
  • Drag in .srt subtitle file then press “Review & Import”.
  • Enter title, select course & level, then press “Import”.
  • The lesson appears. Select “…”->“Edit Lesson”.
  • Click “Add Audio” to add audio file.
    Click “Add Video” to link to a YouTube video.
    Click the blue circle in the thumbnail to replace the thumbnail image.
    Add “Original URL”, “Description”.
    Click “Save”
  • Then I press “Lesson Text” to edit text and timestamps, “Save & View Lesson” and repeat until it looks good.
  • Then I share the lesson, add a lesson Category, and click “Save”.
3

For SRT files, it seems you still edit the timestamp. lingQ does not preserve the timestamps from SRT files?

4

@Hsingh LingQ does preserve the timestamps from SRT files. Well, it probably truncates timestamp resolution to 0.1s whereas SRT files have millisecond resolution, and this may be why I sometimes feel that I need to adjust some of them by 0.1s in LingQ.

I actually generate those SRT files myself using Subtitle Edit, and test and edit them there to be perfect. But I also know that if I don’t adjust them before importing them into LingQ, then they often start slightly too late. So, I shift them all to start 150ms earlier, then I expand them all by 300ms, or as much as possible without creating overlaps. This makes for a better experience for the learner on LingQ. But even then, sometimes on LingQ during testing I find a subtitle starts or ends slightly off, so I adjust it by 0.1s. Sometimes even 0.2s. How much I do this depends on the material and how much I care about having it perfect.

Adjusting of timestamps is not easy on LingQ, and thus very time consuming. I’ve been begging them for small easy changes to make editing them a lot faster and easier, but nothing’s been done yet. That’s one reason why I searched for an external tool, found Subtitle Edit, customised it and learned how to use it, and when appropriate, do a lot of editing in Subtitle Edit first.

5

Perfect timestamps are very important for sentence mode. The other modes its fine to be off.

Thanks for sharing your process. I tried SRT long time ago and the timestamp didn’t match or lingQ did it own thing. I gave up at some point. I used Aegisub for subtitle edit. I will try again when I get the chance.

There is also Rooster lesson editor for visually adjusting timestamps but it get very slow when loading large files. Now I am using very high bitrate mp3/m4a files only, due to better sync quality (low bitrate files have erratic behavior), so my audio files are huge.

6

Yes, I totally agree that perfect timestamps are very important for sentence mode. They also greatly improve the learning experience in all other modes. The vast majority of users can’t be bothered correcting timestamps because it is just way too slow and cumbersome. We all would do it a lot more if it were made a lot faster and easier!

That’s to be expected when we’re basically stuck with buttons that can only adjust in 1 second increments (which I rarely use; Please give us 0.1s!), and having to constantly manually enter timestamp numbers.

And the need for time wasting manual entry and/or copying and pasting of multiple timestamps when splitting long sentences into two is also dreadful! Instead make it easy and instant, and just use the current playback pause position as the end of the first sentence and the beginning of the second sentence!

LingQ please save us from this!

7

Regarding your sync issues:

I have researched this. In summary, Aegisub generated SRT files may be prone to sync issues.

VBR MP3s can also cause sync issues. I only upload 48kbps M4A files to LingQ, which is the smallest size that YouTube provides. And critically, unlike MP3, there is no re-encoding. They sound fantastic despite the tiny size, and they load instantly when users use my lessons because they are so small. And there are no sync issues!

Paste this into a few AIs (they all respond differently):

I want to edit or create subtitle files and import them to LingQ along with audio files to create lessons and don’t want the text and audio to be badly synced. The source may sometimes be YouTube. Please discuss Aegisub vs Subtitle Edit, and MP3 vs M4A/AAC. Consider the “Subs line up in Aegisub but are off everywhere else?” reddit thread.