Well, I figured it out.
Seems ET is pretty particular about the format of .wav files.
I discovered that if I put the railgun news in for mine,
it worked fine. So, it was obviously a problem with
my particular .wav.
I did a quick Google for wave file formats and found this:
Then I did some hex dumps on the .wavs that worked
and didn’t work.
It seems ET has problems if the format chunk isn’t
exactly 16 bytes long (excluding the tag and size).
e.g.
The railgun header started like this [format chunk size in red]:
0000000: 5249 4646 b6de 1600 5741 5645 666d 7420 RIFF…WAVEfmt
0000010: 1000 0000 0100 0100 2256 0000 44ac 0000 …"V…D…
0000020: 0200 1000 6461 7461 92de 1600 0000 0000 …data…
whereas mine started like this:
0000000: 5249 4646 d669 1400 5741 5645 666d 7420 RIFF.i…WAVEfmt
0000010: 1200 0000 0100 0100 2256 0000 44ac 0000 …"V…D…
0000020: 0200 1000 0000 6461 7461 b069 1400 6308 …data.i…c.
Those extra two bytes in the format chunk seem to be the problem… i.e. if they are there, it can’t find the “data chunk”
part of the .wav and thus I get the “No data chunk”
error above.
So… before writing a program to change the format chunksize to
16 and toss the two unused bytes (0000), I thought I’d
see what “sox” did. I did:
sox infile.wav outfile.wav
and ran xxd again and, fortunately, the format chunk
header is the same size as the original’s (16).
0000000: 5249 4646 d469 1400 5741 5645 666d 7420 RIFF.i…WAVEfmt
0000010: 1000 0000 0100 0100 2256 0000 44ac 0000 …"V…D…
0000020: 0200 1000 6461 7461 b069 1400 6308 fd08 …data.i…c…
zipped up my new file (with its script) and Viola! It works!
I guess depending upon the sound programs you use,
you may or may not have this problem… I was using
Creative Labs Wave Studio (and loading it up in
Windows Sound Recorder and then saving again
didn’t help).
Anyway… thought I’d share that in case someone
runs into similar problems.
SiliconSlick