metadata remove breaking extra audio tracks

[jaykae89]

hi im not sure if it is a bug or suppose to do that, but the remove the meta data feature is breaking the extra audio tracks added to a mkv?

[boranblok]

First of all, the remove metadata function is really deprecated and not recommended to be used. It breaks a lot of usability even when it would work 100% correctly. People cannot use the files to reseed torrents, some mkv metadata is misinterpreted by plex, etc.

So if this is not essential to your use case I would strongly recommend not using this functionality.

That said, the following is in essence the logic performed:

Is it an MKV? If so execute mkvpropedit --set title="" "filename.mkv" then execute ffmpeg -i "input.ext" -map_metadata -1 -vcodec copy -acodec copy "output.ext"

I could make this a parameter instead of these fixed commands, but for a feature I'd rather not see be used anymore I dont want to spend too much energy.

I would suggest if this is really essential that you move to a preprocessing loop where you move files to a separate folder first, have your own batch processing script pick them up from there, strip metadata as you see fit, and then have it dropped into the watchfolders of the uploader. This gives you much finer control over the proces.

[jaykae89]

hi boranblok thanks for your reply, the main reason for removing the meta is so other indexers cannot put the parts together and form the releases

im not sure if the meta being there puts the real release name in the meta or the guid ect, either way nzedb indexers can track and make a release from it

so it would be good to have both, or maybe you can advice what line for me to edit that will allow me to still use remove meta but not break the audio tracks

or if there is a naother way to stop the other indexers merging the posts i could not strip the meta data

process is

mkv goes into watch,your app packs and uploads it over multi newsgroups, using =UseObfuscation=yes

[boranblok]

hmmh, you might be going about this the wrong way.

Most indexer scraping nowadays goes via an nzb grab and store. No obfuscation method short of encryption is going to solve that.

The uploader already distributes all files out over multiple usenet groups and the name obfuscation makes the post have a guid for name. Reconstructing a valid nzb out of those separate pieces is a rather daunting task so most indexers don't bother.

In any case, obfuscating the meta data does not really help, if they can assemble all pieces of a release then the file is readable and contents can be seen.

[jaykae89]

sure i understand that but from the tests i ran yesterday the same indexers could only grab the releases where meta data was not stripped, any with it on they did not index, so that rules out there grabbing the .nzb?

[boranblok]

hmmh, weird situation, it might be that stripping the metadata renders the file not being processable by the indexer hence getting dropped.

In any case, you can try the posted ffmpeg command and see if the output is what you see. If this is the case this command should probably be tweaked.