Now this is odd. I'm using DTS Pro Surround Encoder software (which includes Packer, Encoder, and Encrypter) to convert multichannel PCM to DTS.wav. It can generate 5.1 DTS files at 44, 48, or 96kHz. The first step is packing the 6 mono wav files of a track into an '.agm' file (a format DTS uses as input for its Encoder). To do this you select the appropriate mono files one by one and assign them to L,R,C,Ls,Rs,LFE, then 'pack' them with the Packer module.
I ripped some 5.1 96/24 PCM tracks from a DVDA I own, intending to convert to 96/24 DTS. But the Packer input file opener didn't recognize the mono .wav files...the directory they are in showed 'empty' and when I typed the name of a .wav file I *knew* the be in there, it generated a 'file format not supported' error. After fussing around a bit, I fortuitously found that if I run the files through Audiomuxer and convert them to 'standard wav' with 'resampling' at 96kHz (i.e., the sample rate they were already at) , the filenames were now visible in the selection window. After thus resampling all the files, I was able to proceed with the conversion to 96/24 DTS with no flaws. (btw they're also visible if I resample them in the more common meaning, i e downsampling to 48kHz)
Does this imply something like a big endian/little endian .wav issue, or is it something else?
Different types of WAV header, maybe?
You can open WAV files in MediaInfo to determine what's different.
(Or you can load those files in audio editor or player (e.g. foobar2000) and re-save as WAV)
I've encountered something similar with "WAVs" written by eac3to: even though I specify .wav as the output extension, it doesn't appear to write a header, i.e. it seems to just write raw PCM. I say "seems" because CoolEdit 2000 pops up the "Import Raw PCM" dialog, but fb2k correctly recognizes them as 24/48 WAVs (downmixes from DD5.1/DTS5.1 tracks), and if I "re-encode" to WAV with fb2k, CE2K can read that correctly. Interestingly enough, if I have eac3to write to individual mono WAVs, CE2K *will* recognize those.
My guess would be that your DVD-A ripping program is doing something similar when it writes its output files, and also that Audiomuxer, like fb2k, can "interpret" whatever is/isn't being written as a header to the ripped WAVs.
Do you have fb2k kicking around to see if it can recognize those ripped WAVs?
Edit: There seems to be an echo in here.