Using Audacity 2.0.6 with the ffmpeg library I notice that when mpg files from a DVD or wtv files recorded OTA are loaded, there is a section at the start of the stream that is blank.
It increases the length of the stream anywhere from seconds to minutes depending on the running time of the source.
If the same audio is extracted using ffmpeg cli and loaded, the audio shows the correct length.
The mpg files are generated by VOB2MPG, the wtv files are from Windows Media Centre.
Anyone have any comment as to the reason.
+ Reply to Thread
Results 1 to 9 of 9
There is a time code, but I think its refers to the time of the frame in the original DVD that this was re-authored from.
Time code of first frame : 01:24:15;09
Time code of first frame : 10:22:43:31
Do you have a file that DOESN'T have a "time code of first frame", does it do the same thing?
No, all the mpg and wtv files have a "time code of first frame".
Other formats don't have a problem (but they don't a "time code...").
If I take the mpg which came from VOB2MPG and let ffmpeg re-mux without re-encoding, it loads into Audacity without issue (it still has the original "time code.." thingy).
I ran into a similar problem with Audacity. I was extracting extras from a Blu-ray video using the .mpls files and tsMuxeR and extracting as a .ts file.
I dropped the .ts file into Audacity to convert the LPCM audio to AC-3, then I added that and the VC-1 video to MKVMerge and converted to a MKV container.
The audio was way off. I noticed when checking the MKVmerge'd file with MediaInfo that the audio length was one minute longer than the video.
Re-doing the process with Audacity, I saw that blank part at the beginning that was also exactly one minute long.
I extracted the extra again with tsMuxeR as as a demuxed version with separate video and audio and dropped the just the audio file into Audacity and it
showed the correct length with no blank area. I converted to AC-3 with no problems.
Then I put the new AC-3 file and the VC-1 video into MKVMerge and re-muxed it to a MKV. Everything in sync and no problems.
ffmpeg was also used to generate the AC-3 file, but the blank area was there before this, so I don't know that ffmpeg was the culprit. But the extracted .ts file did play back with no blank section. I wasn't aware that ffmpeg was reading the LPCM file, but something in Audacity seems to have added the blank area. I also didn't think that Audacity had a problem with muxed files.
Yes, if I extract the audio before loading into Audacity there's no problem.
In my particular situation, I want to open the recorded .wtv file to get the audio and remove all but the talky part in the centre channel( "Grimm" causes problems for a family member inasmuch as the mood music overpowers the actors speaking) then add it back to the video.