NO AUDIO HLS VIDEO (m3u8 / .ts stream)

[lol]

No Audio after downloading video. nomatther what settings i triy. anyone can help?

[jc vdh]

hi,

please give a URL example

jerome

[lolme]

its a porn URL thats why he not provide! but i also figured out the same issue with some videos in past! video downloads but NO AUDIO comes with the video! i loaded up the file video in video edit software but no audio displayed!

[meme]

Problem is: there are 2 Links! (1 is video and 1 Audio) would be nice if download helper would auto merge this2 together

[Walerie Schwenzer]

After left-clicking on the VDH icon one can see thatt the download queue alternates between MPT2 snippets of short and bigger size. I suppose that's a sign that video and audio are BOTH detected. At the end of the fast moving queue there are several .mp4 solutions where you can choose the download. But this download does NOT contain any audio.

So I suppose that VDH doesn't unite the two streams of snippets correctly. Any takers ?

[Walerie Schwenzer]

important UPDATE:

I just found out that VDH outputs video AND sound but in DIFFERENT .mp4 files: I clicked on the mp4-file with resolution info and saved it with VDH; the I clicked on the mp4-file in the list WITHOUT the resolution info and saved it under a different name.

The first download contained the video WTHOUT sound and the second download contained ONLY the sound. But so far I don't know how to remix both streams after the download. I am using VLC as video viewer.

[Walerie Schwenzer]

To answer my own question about muxing the audio and video streams afterwords: A solution with ffmpeg can be found here:

https://superuser.com/questions/1116326/replace-audio-sync-save-all-to-a-new-video-file-vlc

Just tried it out: It works !

If only VDH could put it all together under its OWN hood ;-P ...

[meme]

thank you very much! this is a good advice! it would be nice if there is an option to murge /remux them automaticly within the VDH! you know another Softare / Addon wich can do this?

i discovered a software named Jdownloader 2... this software can downlad video and audio separately and we can murge them together with third party apps. but i does not find out if there is an option to download and murge in one step....hm

[jc vdh]

yes, ffmpeg is the good idea.

Could be an interesting feature, thanks for the feedback.

jerome

[mig]

Merging Audio and Video together to make a Audio+Video file is easy and we have all the support in VDH for that. This is what we do for YouTube ADP variants.

The problem is that on many sites that stream audio and video separately, we cannot automatically identify the streams as being the audio and/or the video of a same medium.

If we can identify a pattern to match both stream types than we can do something, but it won't be easy if no one posts sample URLs.

[Wild Willy]

It's not necessary to merge the video & the audio into a single MP4 to permit playback of the audio in sync with the video.  VLC can play video & synchronized audio from 2 separate files & it looks just like it's all one file.  I mentioned this in this other thread: https://groups.google.com/forum/#!searchin/video-downloadhelper-q-and-a/no$20audio%7Csort:date/video-downloadhelper-q-and-a/8V2cRB-bcK4/eNUK07BNAQAJ.  The instructions are on the VLC Wiki site, which you can find by hitting F1 when you are in the VLC window.

But one lesson to learn from that other thread is that you have to do some guessing on your own to figure out whether the site is offering the video & the audio in separate streams.  I think my pal MgFrobozz (who posted about it in that other thread) has figured something out about it.  He might have more to say about it here.  For now, I agree with Michel.  I have looked at the m3u8 files the Metropolitan Opera is feeding us every night & I can't tell by looking there which are the video streams & which are the audio.  As far as I can tell, the audio stream is a chunked mp4, same as the video stream.  The manifest shows video resolutions on the entries for the audio streams.  It's all very confusing to me.  So I just download them & leave them as separate files.  When I want to watch the opera, I just tell VLC where the 2 files are & let it put them together during playback.  Leaving them as separate files allows me to just listen to an opera without watching it should I feel like doing that.  As for captions, that's also hit & miss, as is also discussed in that other thread.

Just to emphasize this all, my download of the video stream of tonight's opera, Don Carlo, just completed while I was composing this post.  The download of the audio stream completed a while ago.  It took only 24 minutes.  The download of the video stream took 3 hours and 17 minutes.  I have a not particularly fast Internet connection.  These were the first downloads I have done using VDH 7.3.10a1 beta, which is just to say thanks for the great work.  It still works flawlessly.  Separately, I downloaded the captions in less than a minute, but that file is only 74K.  It would be nice if there were some way VDH could recognize these but, as I described in the other thread, it's still possible to manually locate the m3u8 manifest for the captions & use ffmpeg directly yourself to download that stream into a single .vtt file.

[mig]

Jerome suggested to add the following feature to VDH: having a tool to merge the audio and the video. It would be the responsibility of the user to download both stream types into 2 separate files and then to specify into VDH what file is the audio and which one is the video to launch the merge. This sounds doable to me and a good compromise between automation and capability.

[meme]

i think that wouldnt be a big problem for you to add this feauture! sites wich offer video streams in 3mu8 (video and audio separate) are spliting both files also in two separate URLS! so if VDH could give the option to INSERT some links to download an merge automaticly somewhere in the addon, this would be an idea!

VLC does so that is right, unfortunately NOT every video with audio / video split will work with vlc!  JDownloader also recognize audio and video URL but unfortunatelly its not been able to merge and download them!

[Gary Vellenzer]

Thanks for the tutorial on joining met video and audio. Any idea on how to get the subtitles also?

[jc vdh]

to me the idea is to provide a very simple interface for a user who managed to get audio and video tracks separately and does not know how to merge them.

jerome

[meme]

VDH does NOT upport audio and video download / merge options, right? ord did i something wrong? will this feauture be added soon?

[jc vdh]

VDH aggregates videos when it knows how to do it.

Sometimes it does not find the aggregation info, and then don't.

jerome

[meme]

VIDEOS: YES, AUDIO: NO! 

NO AUTO MERGE AUDIO N VIDEO AFTER DOWNLOAD

[MgFrobozz]

Sorry, it's been a while since I've chimed in. As you've noticed, brightcove changes their packaging occasionally, and I've found I needed to make changes to the python script to keep up with it. (Things were radically different for the At-Home Gala).

If they've really baffled me, I usually do the following to regain my bearings (this is for google-chrome, but is easily adapted for firefox):

• Click on the chrome 3-dot menu (right-hand side of the google-chrome)
• Click on "More tools"
• Click on "Developer tools"
• In the same tab, browse to metopera.org and click on the "Watch Now" button
• In the toolbar for the developer tools, click on "Network"
• Scroll to the top of the Network display. Each line is a request/response between your (client) machine and brightcove's (server) machine
• Click on the left-hand top item in the Network display. This will bring up a right-hand display with Headers, Preview, Response, etc tabs
• Select the Reponse tab
• Select the top left-hand item (again), then use your down-arrow key to chug through the items until something of interest appears in the response window. Most of what you see will be javascript code and/or html or "No response available"
• Once you see a response starting with a curly bracket "{" and some data (eg "{"description":null,"poster_sources":[{"src":"https://cf-images.us-east ..."), you've found a json (JavaScript Object Notation) file. Each of these contains good information about the clip. Copy the stuff in the response window, pretty-print it (I paste into https://jsoncompare.com/#!/simple, setting the beautifier to 4 spaces, and then hit "Lint" to get that), and you'll be able to start figuring out where everything is.
• Note that using curl or wget on the url for that entry won't work; you need to include the policyKey for the clip in the curl header in order to do that.
• Doing this for tonight's opera (Aida, with Leontyne Price!), the left-hand line which shows "6152432857001" (which is the videoId for this clip) contains that information. It includes urls for the vtt tracks, and also includes the url of master.m3u8, which is the top-level manifest file.

[Geevee]

Frobozz,

For the Met nightly Opera, once you are on the network tab, enter m3u8 into the Filter field. You will get three rendition m3u8s. the first is video, the second audio, the third English subtitles. If you reload the page, you will also get the master m3u8.

You can get the URL of each m3u8 file by right-clicking and selecting "copy URL".

You get this without even starting the playback.

If you want non-English subtitles, start the playback, select the language you want, and you will get another m3u8 for the subtitles in that language.

(As it happens, today's Aida is ancient and has only English subtitles available)

[segmentation fault]

Michel said:

"The problem is that on many sites that stream audio and video separately, we cannot automatically identify the streams as being the audio and/or the video of a same medium."

...and Geevee said:

"For the Met nightly Opera, once you are on the network tab, enter m3u8 into the Filter field. You will get three rendition m3u8s. the first is video, the second audio, the third English subtitles. If you reload the page, you will also get the master m3u8.

You can get the URL of each m3u8 file by right-clicking and selecting "copy URL".

You get this without even starting the playback.

If you want non-English subtitles, start the playback, select the language you want, and you will get another m3u8 for the subtitles in that language."

Now, what Geevee said suggests the following feature request for Michel:

How about this: almost every site that offers videos comes with a player. Tell the user to start the player with the resolution, language and subtitles he needs. At the same time, monitor network traffic internally, just as the "network tab" of the browser's Developer Tools (F12) does - and filter the .m3u8 filenames that get requested from the player. These 3 files will then be the right (because hand-chosen by the user!) audio, video and (possibly) subtitle files. You can identify which is which by tentatively downloading the first .ts file of each and doing some file inspection (something like the "file" command in Unix, which checks the "magic" byte at the start of the file in order to infer its type). Once you know which .m3u8 is for video and which for audio, pass them to ffmpeg, doing something like

ffmpeg -i audio.m3u8 -i video.m3u8 -async 1 -c copy complete-playlist.m3u8

This does not produce a complete .m3u8 file, as I thought, BUT: it creates new .ts files that do have video+audio: hundreds of complete-playlistNNN.ts files! All you have it to concatenate them into one .mp4 file, e.g.

cat complete-playlist[0-9].ts > 0-9.mp4

will produce a .mp4 video from the first 10 "combined" .ts files thus created - including audio! 'cat'ing them all in the right order will produce the whole video, including audio (and, if given, possibly also subtitles).

[Wild Willy]

This is an old enough thread. Some things have changed. For one, the Metropolitan Opera
is no longer offering their free nightly streams, more's the pity, because they are
supposed to be getting back to live performances in their theater next month, which is
good news. For another, VDH now has a feature for merging 2 separately downloaded files,
one being video without audio & the other being audio without video. You still have to
guess which file will be which on the VDH menu, but once you've guessed right & done the
2 downloads, you can use VDH to merge them. Personally, I'm happy to leave them separate
& play them using the VLC synchronous playback feature, as I describe here:

https://groups.google.com/g/video-downloadhelper-q-and-a/c/YwvzcYm-fP0

Sadly, there is still no support in VDH for subtitles (captions).

Many of the manifests I've looked at show multiple audio-only streams with associated
video-only streams. These relationships are explicitly identified in the manifests. VDH
should, but still does not, structure its menu in such a way that you can easily
associate the multiple -- I emphasize -- multiple audio streams with certain video
streams. Some given audio streams are associated with multiple video streams of varying
resolutions & bitrates. VDH really could very easily reflect these relationships in its
menu but it does not. We are left to either guess, or try them all to find out, or
interpret the manifest ourselves. But some sites don't use manifests so you can't always
get what you want. (Who knew Mick Jagger uses VDH?) As for captions, when they exist
they also might appear in the manifest. VDH does not pay any attention to them at all.
When captions are not in the manifest, I'm not sure how VDH would detect them. Players
that support captions that can be turned on & off do something to support that so I
should think there is something VDH could do about captions even in these cases.

I have suggested that VDH use ffprobe to determine properties of streams for display on
the VDH menu. In general, this would allow VDH to properly identify streams that have
only video, only audio, or both. ffprobe also identifies stream duration, an attribute
that sometimes is & sometimes is not displayed in the VDH menu. ffprobe also identifies
video resolution, an attribute that sometimes is & sometimes is not displayed in the VDH
menu. ffprobe also identifies associations between audio-only & video-only streams,
something VDH today gives us no help whatsoever in identifying.

I have seen sites on which VDH displays the size of the variant before you download it.
I'm not sure where this information comes from. It doesn't seem to be consistent but it
would be nice if we could get a file size for every variant all the time. I suspect this
is one of those things that VDH must rely on the web site to get & since there's no
international standard for these things, maybe VDH can't always get the information. But
I am just speculating.

Things are possible. We can only hope they are in the works.