Please report bugs in the Avidemux forums to get some feedback and determine the nature of the problem.
If it involves non systematic crashes, please also do the following:
gdb avidemux2 r
Do as you usual, when it crashes, go back to gdb and take the output of
That plus the last lines printed by avidemux2 will help a lot.
You can attach your log files and screenshots to your posts via the “Add file” button. One exception is the quick reply form, which doesn't offer such funcionnality (yet). If that doesn't work, you can use pastebin or pastebay to post large text content. Posting in the thread itself is tolerated but makes reading the thread painful, and could turn some helpers away.
If for some reason you don't want to (or can't) upload a huge sample on the web, you'll want to cut a piece of the file (ie. grab the first xx MBytes of the file).
On windows, you can use DGSplit for this purpose. Check the “Stop after N chunks” box, for 1 chunk, to speed up the process. The chunks are numbered starting from 0. Your cut sample should be named
For .TS and for .M2TS files you can use Cut TS sample
Alternatively, if DGSplit doesn't work for you or if it doesn't work with your OS via an emulator, the split command does this perfectly (where XX is the number of MBytes you want to cut) :
split -bXXM <input file> <output prefix> –verbose
This will actually split the entire input file into chunks of XX MBytes. You only want the first part, which should be named
You can freely Ctrl+C the process after the command has told you the nnnnnnnn-ab sample is being created.
The split command is available natively under *nix. For Windows, you can grab the GnuWin32 CoreUtils. It contains the split command mentionned above, along with many other nifty tools from the *nix world. Once the “Complete package, except sources” is installed, don't forget to update your PATH environment variable so it points to “C:\Program Files\GnuWin32\bin;” (or whatever path you defined during installation).
Using tools like XtremSplit or other fancy file splitters is a no-go in this precise scenario since they usually require all the chunks to be present for the data to be reconstructed, which won't be the case.