FC Pro X - Importing caption files, only part of the captions show, not all

As a test, I generated a (synthetic) caption file with more than 5000 captions and it imported without issue (it took a bit, but came out perfect).

I then modified it and got the error also.

I think I have a general idea of what is causing the problem: FCP does not take captions where the duration is too small. If your duration is, say, 1 millisecond, then it will be less than one frame and that causes an error.

In your case, that happens with caption number 117. It's timed to display for... 1 millisecond.

There may be more of those, so just fixing this one may not be enough, but at least we have a handle on the problem.

What is the caption format?

Try resetting FCP preferences.

Try opening a new library and importing. Any change?

If the problem persists, would it be possible to share the captions file or at least a portion of it that includes the area where it is failing to import?

If the program producing the captions cannot be tuned to avoid these too short ones, we can always adjust the file afterwards. After all, it is only text in it. There are far too many captions to just try to find the problem ones visually.

One could import into FCP, note where the problem appears (it is the first caption after the last that imported correctly), and manually edit it. But this is still a nuisance, if there are many of those. I found three in your file, and then I quit searching.

A little script can help, though. At most, we could concoct a script that takes a file and spits out a new file with some adjustments to timing to avoid the problem.

I propose an intermediate solution, since I do not have the time to program and test a full thing.

We may try to build a little script that identifies the problematic captions, then they can be fixed by hand.

What do you think? I may try to do it if you think it is useful.

Luis will likely provide something easier to use, but I hacked together a python script with the help of ChatGPT.

In Terminal.app you'll first need to install the srt python library:

$ pip install srt

Then create this python script using something like vi in Terminal or a simple text editor. I named it srtchk.py and set it to check for captions that are less than 2 milliseconds. Change the line that sets the name of your SRT file. You can save this script wherever you want. I just put it adjacent to the srt file in a temp directory.

import srt
from datetime import timedelta

def check_caption_duration(file_path):
    with open(file_path, 'r') as f:
        srt_data = f.read()

    subtitles = list(srt.parse(srt_data))

    for subtitle in subtitles:
        caption_duration = subtitle.end - subtitle.start
        if caption_duration.total_seconds() < 0.002:  # 2 milliseconds
            print(f"Caption {subtitle.index} is less than 2 milliseconds.")

# Replace 'subtitle_file.srt' with the path to your SRT file
srt_file_path = 'subtitle_file.srt'

check_caption_duration(srt_file_path)

In Terminal, I navigated to the temp directory that has the python script and your SRT file and ran it with:

$ python srtchk.py

It found these in your SRT:

Caption 117 is less than 2 milliseconds.
Caption 1150 is less than 2 milliseconds.
Caption 1542 is less than 2 milliseconds.
Caption 1914 is less than 2 milliseconds.
Caption 2086 is less than 2 milliseconds.

Thank you for posting the file.

I have been experimenting with it, and I am getting errors as well.

I was suspecting that there might be some invisible character that would be invalid, but I have confirmed that it is not the case. It is a bit weird, because there does not seem to be one specific caption that is invalid, but it always fails around the same one. I am still trying some variations and if I arrive at a useful conclusion I will report back.

Thank you, Terry!

I have learned something new thanks to you...

I was coding this by hand (and it was mostly working, but still had to polish), and of course I should have though that "there is a Python module for that!"

I believe that your solution is just the right thing.

With that output it is easy to open the srt file in a text editor, and make an adjustment to the timings.

You're very welcome Luis. Finally, the student can offer some assistance to the master!

Hopefully @mercury290 can use this. Though I think the best solution would be for Descript to allow a user to define a minimum caption duration. 1 millisecond is pretty useless. Even 10 milliseconds is not great.