I don't work for VDH. I have posted that in this forum on numerous occasions. I DO NOT
WORK FOR VDH. I'm just another user, like almost everybody else who posts here. Only
Michel (who posts here as mig) & Jérôme (who has posted above in this thread) work for
VDH. Michel is the developer of VDH. As far as I can tell, he is the ONLY engineer
working on the product. Jérôme helps out by monitoring this forum so Michel doesn't have
to. So people should NOT E-mail me anything. If you have something to say that
progresses the discussion, you should post it here, as a regular post, so the maximum
number of people can read it & chime in here. But it doesn't matter because I have not
received any E-mail from you, Matt. That's good. Don't retry it. Just post whatever
you have to say here. Use Reply all, NOT Reply to author.
That's out of the way. Now we can deal with the issue at hand. I first reported a
problem about multiple concurrent VDH downloads in this thread:
https://groups.google.com/g/video-downloadhelper-q-and-a/c/lCy0hw8t-uE
I had a hard time finding that one, it's so old. But the topic has come up a number of
times since. I just searched for those discussion & found a few that are on the topic.
Problem is they are also to a large extent off the topic. That other thread I've just
cited is mostly still relevant but I have learned a few things in the 4 and a half years
since. (No kidding? It took me only 4 and a half years. I'm so smart.) The short
story about the multiple concurrent downloads is this. An MP4 includes something called
a moov atom. I am no expert on the thing & what I know is actually subject to challenge
from someone who does know. It is my understanding that the moov atom is metadata about
the MP4. It includes information like resolution, frame rate, duration, & I don't know
what else. It is information required by playback applications (like VLC). Michel has
explained in here somewhere (good luck finding it) that he keeps the moov atom in program
storage while VDH is downloading something. He does not write it periodically to a file,
not the target file of the download, not some temporary side file. This means that it is
subject to management by the operating system's paging mechanism. When the operating
system is Linux in one of its plethora of mutations, I gather the operating system does a
good job of managing its paging algorithm. Michel develops VDH on Linux so he thinks
this architecture is a good idea. But out here in the real world, there's a lot of folks
relying on Microsoft. For example, I run Windows 7. The paging algorithm in Windows is,
shall we say, less robust. As I describe in that thread I've just cited, it is easy for
me to bring my system to its knees due to thrashing on the pagefile. It's so easy that I
have managed to avoid doing it in a very long time. I will either single-thread my
downloads in VDH or I will use ffmpeg instead. Ffmpeg isn't subject to the thrashing
problem because it manages the whole process of downloading in a different way. That
sounds like I know what that different way is. I don't. I just know that I can run
multiple concurrent large downloads with ffmpeg & the thrashing problem doesn't occur.
The only conlusion is that ffmpeg does things completely different from VDH. In that old
thread I mention my slow Internet connection. Since that time, I have subscribed to a
different ISP & my bandwidth is now 100 times what it was. But that does not affect the
thrashing problem in VDH. That is still present.
In any case, if you are trying to push multiple large downloads through VDH at the same
time, you are probably hitting this thrashing problem. When the thrashing occurs, it's
likely that only one of your downloads completes successfully, and that only after an
unreasonably long delay. The rest of your downloads probably time out during that long
delay. When that happens, VDH usually just prematurely terminates those downloads
without writing the moov atom. Usually, not always, but usually. When a file doesn't
have a moov atom, it is not simply impossible to play it. The tools I found that claimed
to be able to repair videos all told me that they couldn't repair such interrupted
downloads because there were no moov atoms in the files. Catch-22. I speculate that
some form of this problem is occurring when you have a truncated audio track..
So don't try to download multiple concurrent videos with VDH. Single-thread them. It
doesn't matter what your bandwidth is. Single-thread your VDH downloads. Go into your
VDH settings & look for these 2:
Max concurrent downloads
Max concurrent stream downloads
Set them both to 1. That's how you single-thread downloads in VDH.
It is my understanding that the MKV format does not rely on a moov atom. I have no idea
about WEBM. Still, to be safe, I recommend single-threading VDH until we hear from
Michel that he has changed the architecture of VDH. I would not hold my breath waiting
for that. I reported that thrashing problem 4 and a half years ago. The VDH
architecture has not changed.
Despite all of that, I am still waiting for somebody here to post some URLs of videos
that you downloaded with truncated audio tracks. Amy McLaughlin has teased us with a
promise to do that. I hope she follows through. By URLs I mean URLs to the web pages
where one of us can download the same videos. I don't need to see one of your videos
with a truncated audio track. That is no help. I need to do the same download you did &
see if I get the same result. Long-winded treatises (I can write those with the best of
you, present post being proof) are not really advancing us towards a solution. Post some
URLs. Let me (and other people, that's why you should post & not send private E-mails)