Strange... I don't know how the audio data would get changed. It's not easy to alter PCM audio data "accidently" without
really fouling-up the audio... Do these files sound similar?
If the files sound the same, I wouldn't worry about it. If they sound different, choose the best-sounding one.
I
think foobar2000 can compare the audio data (or get a checksum of the audio) while ignoring the file header. (I don't use foobar, so someone else can help with that... assuming it's actually possible.)
I have a m2ts file with video and LCPM audio.
Is there possibly more than one audio stream? (Like a file from a DVD with both LPCM and Dolby AC3 streams?) Audacity (and I think VLC) should be able to open various formats and save them as WAV. I'm not sure if TsMuxer can convert, but a simple de-muxer should simply be making an m2a file... If it's making a WAV file, it's doing more than de-muxing... but it should be just writing a header followed by the original PCM data.
For a start, file size are slight different.
That could simply be a difference in the header, or perhaps some silence-padding on the end, depending on how the MPEG frames are decoded.
When i open the wave file in Audacity, the waveform are slightly different as well.
How are they different? If you can see a difference, you should be able to hear a difference....
If i perform a binary compare, the whole files are totally different.
(if its a offset issue, i normally only see differences at the start or end of the file)
If there's an offset,
all of the bytes will be offset and a binary compare will show at all bytes are different. If I compare ABCD to xxABCD, the 3rd element in the 1st series is 'C' and the 3rd element in the 2nd series is 'A'.
None of the data matches. (All of the data
after whatever is causing the offset will be mis-matched.)
Use http://hank315.nl/ (http://hank315.nl/) DGMPGDecode package. DGDecode should demux to a .pcm file for you. You can then convert that using SoX, NicLPCMSource() + SoundOut() in Avisynth, etc.