Finally got my Taranis and got around to playing with this. I noticed this difference when inspecting the files (V911.wav is generated on OS X 10.7.1, the other one is from the Taranis). The original files shows 16 bits per sample which is not shown on the generated files. I tried looking at the man pages for say and experimenting, but I can't seem to be able to force this setting (it seems to only work for pcm). UPDATE:xtrmtrk wrote:The Mac command tools include a very nifty "say" program. I've tried using it like this:
say -o ~/loiter.wav --data-format=ulaw@32000 --channels=1 -v Victoria "Changed to Loiter Mode."
It's great because with a command line you can create many wav files very quickly using any voice on your Mac saying anything you want. The problem is it doesn't work. I've compared the files created here which don't work to those from Audacity which do work and I can't tell the difference. I've examined them with several different tools and all of the header and codec info seems the same - WAV, MU-LAW, mono, 32000hz and 8 bit.
Attached is a sample file I created. Can anyone tell me why it won't play? I'd be grateful.
Was able to learn a bit more and get what seems like a correct file, but can't test till I get back from work. The (builtin) afinfo util gives detailed information on audio files, this is for a default Taranis sound:
Code: Select all
$ afinfo flapsup.wav
File: flapsup.wav
File type ID: WAVE
Num Tracks: 1
----
Data format: 1 ch, 32000 Hz, 'lpcm' (0x0000000C) 16-bit little-endian signed integer
no channel layout.
estimated duration: 0.717594 sec
audio bytes: 45926
audio packets: 22963
bit rate: 512000 bits per second
packet size upper bound: 2
maximum packet size: 2
audio data file offset: 56
optimized
source bit depth: I16
----
say -o test.wav --data-format=LEI16@32000 --channels=1 -v Samantha "Testing audio"
Looks good so far, will report back once I test it.
UPDATE 2: It works perfectly.