-
Notifications
You must be signed in to change notification settings - Fork 335
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Audiograms include 81m11s long "Apple Text" component #89
Comments
Hmm, I'm not sure where to even begin debugging this. It certainly seems like the Hindenburg export settings would be the culprit. Does this ever occur with an MP3 exported from a different source? In terms of it being "intermittent," are you sure the conditions are identical? Can you take the same file and run it twice and get different results? Or is two different files exported from Hindenburg? My best guess would just be that whether your export includes that metadata depends on exactly how you export it or what time range you have selected. Can you attach one of the mp3 files that produces the bad result? This may be related: |
I'll test a different source (and may simply re-export the same MP3 from Audacity). The conditions were as identical as I could make it - those two MediaInfo outputs were from the same audio file and used the same pic. Here is the MP3 file I used for them: |
I tried with a different MP3 file (again, from Hindenburg, but taken from an edit which didn't ISTR use chapter headings), and that gave me a correct 28s length Audiogram. Also, putting the offending audio files through Audacity (which strips the extraneous information whilst also decoding & recoding it) seems to fix it too. And I can predictably reproduce the bug with the offending MP3 now. So I have a workaround, at least. |
After using Audiogram successfully for months to produce half-minute videos, it's increasingly often made them 81m11s long.
Interrogating the MP4 files, I see that the 81m files have these two components, and the files which are the correct length don't. You'll notice that some chapter marks that only make sense for my full podcast have been exported from my audio editor (Hindenburg) into the clip MP3 file, and then imported again by Audiogram; but the strange thing is that this is intermittent. At the end of this, I'll paste the full Mediainfo of two Audiograms created from the same MP3 file, one that works and one that doesn't.
Curiously, the "long" Audiograms are accepted by Facebook, but not by Instagram.
Text
Version:
Audiogram 0.9.5
Uname -a:
Linux orwell 3.2.0-23-generic #36-Ubuntu SMP Tue Apr 10 20:41:14 UTC 2012 i686 i686 i386 GNU/Linux
Full Mediainfo Output for two Audiograms created from same clip & bg image.
Expected:
Buggy:
The text was updated successfully, but these errors were encountered: