Open nebukazar911 opened 6 years ago
Hello! Thanks for trying out ffproc, and the super well-detailed bug report.
I think I know the issue: the (fre)
tag & stream ID on your audio track. ffproc has a line:
if ( track["language"] == "eng" or track["language"] == "und" or track["language"] == None):
which excludes that track from processing.
That dates back from my original use of ripping DVDs, where I did want to bring in only the english track. Maybe my choice was a little short-sighted :smiley:
My proposed fix would be to add a config option somewhere to disable this check, and also add a configurable list of languages that you do want to extract. Does that work for you?
Indeed, that would definitely work! Most of our shows aired in Quebec are in french, but I would also like to rip other stuff with foreign languages; if we could get that option, that would definitely be awesome. Thanks for the quick turnaround, much appreciated 👍
Alright, I've pushed a commit which should hopefully fix this for you. The "default" profile shows an example of the new feature being configured. You'd just have to set "ignore" to true, or add "fre" as an allowed language.
What I intend to implement a little later on is extracting/transcoding audio tracks per-language, so if you had both English and French, your output would have both as well. I'm tracking that in #13. In Ontario at least, I haven't ever seen a TV broadcaster do that, but I have some DVDs that have this configuration.
Let me know if this works/doesn't work for you.
Also, I'm not sure how much ffmpeg
you know, but the output will contain the lines:
Stream mapping:
Stream #0:0 -> #0:0 (h264 (native) -> h264 (libx264))
Stream mapping:
Stream #0:0 -> #0:0 (mpeg2video (native) -> h264 (libx264))
Stream #0:3 -> #0:1 (ac3 (native) -> aac (libfdk_aac))
Stream #0:1 -> #0:2 (copy)
or similar, which lets you verify it's pulling the right data out of your source, and your output has the right tracks, without letting it transcode all the way through (because 2.7fps movie-length transcodes suck just to find out ffproc
did it wrong :disappointed: Sorry about that)
Awesome! I’ll try this out later this week and let you know how it goes.
Cheers!
On Mar 3, 2018, at 12:00 AM, Alex Roth notifications@github.com wrote:
Alright, I've pushed a commit which should hopefully fix this for you. The "default" profile shows an example of the new feature being configured. You'd just have to set "ignore" to true, or add "fre" as an allowed language.
What I intend to implement a little later on is extracting/transcoding audio tracks per-language, so if you had both English and French, your output would have both as well. I'm tracking that in #13. In Ontario at least, I haven't ever seen a TV broadcaster do that, but I have some DVDs that have this configuration.
Let me know if this works/doesn't work for you.
Also, I'm not sure how much ffmpeg you know, but the output will contain the lines:
Stream mapping: Stream #0:0 -> #0:0 (h264 (native) -> h264 (libx264)) Stream mapping: Stream #0:0 -> #0:0 (mpeg2video (native) -> h264 (libx264)) Stream #0:3 -> #0:1 (ac3 (native) -> aac (libfdk_aac)) Stream #0:1 -> #0:2 (copy) or similar, which lets you verify it's pulling the right data out of your source, and your output has the right tracks, without letting it transcode all the way through (because 2.7fps movie-length transcodes suck just to find out ffproc did it wrong 😞 Sorry about that)
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.
Looks like this fellow Quebecois forgot this existed. I'll try to run it on a couple french canadian movies I have laying around, see if I can reproduce the issue :) I also have a metric ton of japanese anime.
Hi, This might not be bug but maybe just a config issue...
I'm currently running Plex and have the DVR function enabled (hooked up on a HDHomeRun Extend). I'm recording shows and movies but since most of our devices are iOS based (eg: atv, iphones, ipad, ...), I have to optimize these records to get directplay.
I've been trying to use ffproc to do mass conversion. So far, I'm able to do conversions, but for some odd reason, the audio track seems to be stripped.
The current ffprobe's output on the TS file is:
And the ffproc conversion's output:
And finally the ffprobe's output on the newly created file:
The current profiles.json looks like:
Are we missing something ? Is there any particular reason why the audio track(s) would get stripped ?
Thanks!