E-AC3 7.1 stream detected and treated as E-AC3 5.1

[YS5tsXOJJAhjJDPS]

Dear all,

I have found a problem (or at least something I don't understand) in VLC 3.0.12 on Windows 10 x64 Enterprise (PC) with a 7.1 surround speaker configuration:

I have a M2TS file with a h264 video track and several sound tracks in various formats, one of them being an E-AC3 7.1 track. VLC recognizes that track as E-AC3 5.1, and consequently sends only 5.1 PCM channels via HDMI to my surround receiver. To be precise, it actually sends 7.1 channels, but both surround rear channels only contain silence.

If I pass through that track in original format to my surround receiver via HDMI, the 7.1 sound works as intended, because my receiver can decode that format. However, I would prefer VLC recognizing the format correctly and converting it correctly to 7.1 PCM streams, and sending these PCM streams via HDMI.

Of course, I am willing to cut 50 MB or so from that file and upload it somewhere to assist in analyzing the problem, but I first would like to know whether VLC at least in theory should be able to decode that format at all. Cutting, uploading and analyzing does not make any sense if VLC is not (yet) able to decode it

Thank you very much in advance!

[Lotesdelere]

It should theorically work fine.

Please open Tools -> Messages (set Verbosity to 2) before you start the playback, and then paste the full resulting log here or on Pastebin.com if it's too long.

Also upload a short sample of your original problematic file to Zippyshare.com or to WeTransfer.com and then post the link to the file here.
https://www.zippyshare.com
https://wetransfer.com

If needed cut it with DGsplit and read here about how to do it:
https://forum.videolan.org/viewtopic.ph ... 35#p193335
100 MB max should be enough.

[YS5tsXOJJAhjJDPS]

Thank you very much for your help and for confirming that VLC should be able to decode that format.

The log file is a little lengthy, so I have posted it here ((file) names obfuscated):

https://pastebin.com/pw2e7PWr

Please note that I have set it to expire after one week.

Regarding the upload: That seems to be difficult at the moment. I have an appropriate file, but when I direct my browser to https://www.zippyshare.com, I just get a "403 Forbidden"; they seem to be down right now. As for wetransfer.com, it seems that we get storage only after payment, which I would like to avoid at the moment; there is no storage in the free account (even not when registering).

May I ask you to have a look into the log file first? Maybe you can see what is going wrong from that; if not, I'll try zippyshare again (hoping that the failure is temporary).

Thank you very much!

[YS5tsXOJJAhjJDPS]

In the meantime, I have found out that zippyshare.com is blocking users from certain countries, my country being among them. Perhaps I can upload the file via TOR or via VPN.

Before doing that, I'd like to know whether the cause of the problem can be seen from the log file.

Thank you very much again!

[Lotesdelere]

You don't need to register on Wetransfer.com, just click on the "+" button for adding files, then on "..." button for getting a link to the file.

[YS5tsXOJJAhjJDPS]

Thank you very much again!

This is the link to the example file:
https://wetransfer.com/downloads/82c6df ... 819/61100f

I am curious about whether VLC behaves as I described, or whether it is a mistake on my side

[Lotesdelere]

Yes, I can reproduce the issue:

Code: Select all

Codec : A/52 B Audio (aka E-AC3) (eac3) Channels : 3F2M/LFE

You should create a new ticket on the VLC Trac with your above report, the link to your sample file and the link to this thread:
https://trac.videolan.org/vlc/

[YS5tsXOJJAhjJDPS]

Thank you very much again - I'll do that (current situation with your link: "maintenance ongoing, please check back later")

[YS5tsXOJJAhjJDPS]

Well, it seems the videolan people try hard to keep users from filing bug reports.

The bug tracker has been moved to code.videolan.org, which obviously was the reason for the downtime. Fair enough, and probably a good decision. I registered there and expected that finally I'd be able to file the bug report, but imagine my surprise when I learned that they insist on two factor authentication.

This is the point where I have lost patience. I definitely will not install any app on my mobile phone just for being able to file a bug report, nor will I register any device (which probably would make tracking me around the net even easier) or buy any piece of hardware for that.

So I politely ask you if somebody else would be willing to file that bug report or somehow bring that problem to the developers' attention. If not, I can live with the situation because my surround hardware can handle that format, and passing it through works. Plus, there are other media players, too. Mileage may vary for other users, though. Some people probably will spend hours searching for the cause why they only get 5.1 on their neat 7.1 equipment, and then learn that they can't work around the problem because their surround receiver doesn't support E-AC3 7.1.

In the meantime, the pastebin with the log file has expired. I have created a new one, this time with two week expiration. It is here:

https://pastebin.com/BU2bh3Da

This will be my last message and last action regarding that subject, with the following exception: If somebody who is willing to resolve the bug asks me to upload the sample file to wetransfer (or elsewhere) again, I'll do that and reply to that request.

Regards, and thank you very much!

[Lotesdelere]

The bug tracker has been moved to code.videolan.org, which obviously was the reason for the downtime.
I registered there and expected that finally I'd be able to file the bug report, but imagine my surprise when I learned that they insist on two factor authentication.

I definitely will not install any app on my mobile phone just for being able to file a bug report, nor will I register any device (which probably would make tracking me around the net even easier) or buy any piece of hardware for that.

Yes, I agree with you.
I see no reason why I should use some foreign private company services and/or give my phone number or any else personal for helping a non-profit organization on my free time.
Non sense.

[mederi]

I think the situation is not that bad. I installed Google Authenticator (you know, Google anyway) and it works offline. It is possible to disable the Two-factor authentication in Account settings.

[YS5tsXOJJAhjJDPS]

For me, it is impossible to disable 2FA. Whatever action I want to perform after having logged in, it is always the same page which reappears and tries to force me to register a device. That is, I can't go into account settings because it shows only that page whatever I do. But I won't set up 2FA. Period.

By the way, that page also has the following hint in a red popup: "The global settings require you to enable Two-Factor Authentication for your account." While I don't know what "global settings" are, I am quite sure that I can't change them. So, obviously, normal users can't disable 2FA. Probably only the administrator of that bug tracker can change it.

[Lotesdelere]

I installed Google Authenticator (you know, Google anyway)

I do not want to use Google or any other private company for my accounts.
Especially when it's about to help free and open source software.

It is possible to disable the Two-factor authentication in Account settings.

But for that you need to log in first, hence to use 2FA

[Stef_An]

I do not want to use Google or any other private company for my accounts.
Especially when it's about to help free and open source software.

I know, that this is a workaround, but it works: I successfully used oathtool in Linux to create a 6 digit PIN from the long code on the 2FA setup page. The whole article about it is here:

https://www.cyberciti.biz/faq/use-oatht ... ation-2fa/

At the end, all you need is a command like oathtool -b --totp 'N3V3R G0nn4 G1v3 Y0u Up' and it generates your PIN. The only problem - it works just for the current session. So after logoff, on the next logon, one have to repeat the PIN generation. But on the bright side: one don't have to install some extra software on the phone or somewhere else.

On Windows, WinAuth should do the trick.

[Stef_An]

I know, that this is a workaround, but it works:

Afterwards, I found this solution on the wiki.

So now one has 3 possibilities:

1. As in the wiki, just save the long code from the right of the QR code and use it with oathtool to manually generate the PIN every time one needs it.
2. Use some full-GUI solution like OTPClient on Linux or WinAuth on Windows or Linux (here with wine). One can create a local saved database and let the PIN be generated every time one needs it. There is no need to install some extra software on the PC or smartphone and sign-up for another account on another service. And the data is saved locally, not in some cloud. One has to, on the other side, setup such database on every device, where one likes to log in into the GitLab of the VLC.
3. As a deviation from the both other scenarios: one just deactivates 2FA every time before one logs off on the GitLab. This way, next time after the log-in, one would be asked to set up the 2FA again and would be once again shown the long code to create a PIN. There is no need to save any information locally. One just uses some program, GUI or not, to generate the PIN as in the set-up step. But: don't forget to deactivate the 2FA, if you don't save the code(s)!

In any scenario: save the recovery codes, for the case you lose the code or need them otherwise. And: either save the code right to the QR code in some file/program database or on the paper, or don't forget to deactivate the 2FA before logout.