First, my setup is Windows 7 64-bit, Firefox 100.0 64-bit, licensed VDH 7.6.3a1 beta,
CoApp 1.6.3, VLC 3.0.16 Vetinari, ffmpeg downloaded from
ffmpeg.org on 2022/3/17.
I am going to be throwing around some ideas that are covered in another thread. You can
find it by looking over here:
https://groups.google.com/g/video-downloadhelper-q-and-a/c/BzPLK2YyL-s
You can find the particular topic in question by doing a string search on "cannot
download" in there & clicking the link you'll find.
To start, I went to your rsi page. VDH presented me with what looked very much like
separate audio & video. I downloaded the 2 variants & indeed, I got a 1280x720
video-only file & an audio-only file of 48kHz sampling rate & 192kbps. They did play
fine in VLC using VLC's synchronous playback feature. It seemed like the audio was badly
out of synch near the start of the file but got a bit better within a couple of minutes.
I couldn't quite figure out whether the audio was ahead or behind the video. I'm sure
the fact that I don't speak Italian made things more difficult, although if I had Rachele
Bianchi Porro teaching me, I'm quite sure I'd learn Italian pretty quick. I didn't play
the thing all the way through. There did seem to be a bit of jumpiness in the video
(what you call non-fluid), but it didn't seem to me to be overly annoying.
Then I went looking in the Network Monitor to see what I could find. Not too
surprisingly, I found an HLS manifest, a .m3u8 file with the string "master" in its URL.
It did turn out to be the master manifest for this item. I ran the manifest through
ffprobe & got the results you can see in attached file ffprobe.txt. It turns out there's
3 Programs in this manifest. Each one consists of an audio stream, a video stream, & TWO
timed_ID3 data streams. Not one timed_ID3, like every other instance I've encountered of
this type of data, but two. Two, that's a first. I've posted about this type of data a
few times in here, discussions you can find by searching this forum. This type of data
is ignored by ffmpeg & it was my understanding that VDH ignored it now as well. Possibly
not. In any case, you can see from the ffprobe output that audio stream 0:0 is shared by
all 3 Programs. The video stream in each program has a different video resolution.
Curiously, stream 0:1 of timed_ID3 data is shared by all 3 Programs, while the second
timed_ID3 stream is unique to each program. I downloaded streams 0:2 & 0:0 via ffmpeg
into a single MP4 output file. Doing it this way avoids downloading the timed_ID3
streams. The resulting video looks much smoother & the audio synch seems pretty good.
If you look closely at the ffprobe output, you'll see it complaining about an unsupported
subtitle stream. This is inexplicable. Using the information provided in the ffprobe
output, I reconstructed the URL for the subtitle manifest. I then used ffmpeg to
download the subtitles. I don't know why ffprobe had any problem with this. As part of
my download of the subtitles with ffmpeg, I ran ffprobe on the resulting .vtt file.
There was no problem with that. I've encountered this before with subtitles. I can't
explain it. When I played the video I got with ffmpeg, VLC displayed the subtitles.
They were in Italian so they weren't helping me understand things any better. The
subtitles seemed to be in perfect synch with the audio I was hearing & with the lips of
the people speaking. Although I didn't understand any of it, the words of the subtitles
looked sort of like the words I was hearing. I didn't sit & watch it all the way
through. I just sampled it at intervals & it seemed to be good all the way to the end.
So. It appears that whatever solution Michel put into VDH to ignore timed_ID3 data isn't
complete. I think the fact that there's 2 streams of timed_ID3 data in each Program is
causing the problem. VDH needs to allow for the possibility that there could be multiple
streams of timed_ID3 data. I think my results point in this direction. So if you want
to get this video now, you're going to have to learn how to use ffmpeg. Or just continue
to use the tools you have already told us about. We'll just have to wait for a new VDH
to see whether it handles this any better.
Speaking of a new VDH, over the past several months, Michel has dropped little comments
here & there about small things he has fixed in some upcoming release. One thing that
was solved that I can think of is the glitch that displaying Details on an in-flight
download in the blue dot status display removes the percent completed & time remaining
from the display. But there have been a few other things. I'd like to see a new VDH
beta with these things fixed. Beta. I don't want an official release on the Mozilla
site. Their installer is brain dead. It still insists you uninstall VDH before you
install a new version. Garbage! A new beta is indicated.