New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
SVT site: "spoken subtitles" audio selected as default audio when available (workaround provided) #5164
Comments
Further information... I just discovered that SVT also serves another separate audio variant called "tydligare tal" ("more audible speech" - with a lower background volume and more distinct voices, intended for those hard of hearing). This is even newer and is only available for very little material so far, but they write that more is on its way. This variant seems to (always? not sure) be prioritized lower than the standard audio by yt-dlp default format selection, so it is not currently as big a problem. It is marked similarly to "uppläst text" but with the identifier "sv-x-tydligaretal". Here is a page collecting SVT material with "tydligare tal" available: https://www.svtplay.se/kategori/tydligare-tal I've included an example below (video format information cut for brevity).
Default selection behavior:
|
And here is an example where both "tydligare tal" and "uppläst undertext" are provided for the same item:
(video formats information cut out) |
More on proposed change in behavior... Also a note to add that I don't know if yt-dlp has special cases for channels/fronts closely related to SVT Play such as SVT24, Öppet arkiv, Kunskapskanalen and so on, but it might be that these also suffer from the same issue, or not, I haven't tested them. |
Another note to add that while adding Note the error:
Formats list:
Just to show that it works without
|
It’s probably because the |
DO NOT REMOVE OR SKIP THE ISSUE TEMPLATE
Checklist
Region
Sweden (most items) and global (some items)
Provide a description that is worded well enough to be understood
Background
Since some time SVT has started to provide "uppläst undertext"/"uppläst text" audio (I think the English term could be "spoken captions"?), which means that you can select an audio track where the original audio is overlaid with a computer voice reading the subtitles aloud. The idea is that this will help viewers who have difficulties seeing/reading the subtitles. This is a very niche use and not what the common user would want or expect.
Technical details
Because of how SVT serves their video/audio formats and how yt-dlp's default format sorting works, for those items on SVT that has "uppläst undertext" audio available, yt-dlp always prefers that audio.
It seems that the "uppläst undertext" formats are always marked with language = "sv-x-tal" (compare to standard Swedish language code "sv"), which facilitates the workaround of adding [language!=sv-x-tal] to the format selection string.
However, this workaround is not easy to find for the user and it should not be expected that yt-dlp prefers to select these very niche audio tracks when just as good quality normal audio is available.
Resources for further investigation/testing
This is a page collecting all the material on SVT with "Uppläst undertext": https://www.svtplay.se/kategori/upplast-undertext
Most of the interesting material is only accessible from within Sweden but there are some that are globally accessible and would be suitable for testing outside of Sweden, e.g. https://www.svtplay.se/15-minuter-fran-sapmi .
Note that the "uppläst undertext" is only added for where SVT only has burnt-in subtitles, where there are soft subtitles available you will not hear the "uppläst undertext" (probably since the target audience would have equipment to access the soft subtitle text without it being added to the audio).
Suggestion
I propose that yt-dlp's behavior is modified so that the default handling of audio selection on SVT websites is to select "normal" audio rather than prioritizing "uppläst undertext" audio.
Provide verbose output that clearly demonstrates the problem
yt-dlp -vU <your command line>
)[debug] Command-line config
) and insert it belowComplete Verbose Output
The text was updated successfully, but these errors were encountered: