I’ve converted multiple files at once, but I haven’t tried thousands of files and I haven’t tried that exact conversion. And, I’m not sure if the 32-bit WAV format is integer or floating-point (and as far as I can tell it doesn’t give you an option).
32 bit WAV can be either integer or floating point.
The conversion to either 32 bit float or 32 bit floating point WAV can be done with SoX (http://sox.sourceforge.net/), but that is a command line program so it would require a little bit of scripting to handle batch conversion.
Any particular reason why “32 bit” PCM .WAV and not 16 bit?
I am actually doing an environmental acoustics project in which I have to compare certain .ogg files. However, the own software of the company I am working in doesn’t accept neither .ogg files nor floating point .wav as input. For that reason I need to convert those .ogg files into .wav PCM and I need them to be exactly the same as the original .ogg files. As the dynamic range of the files is very high not even with 24 bits PCM I can get the exact similitude.
I have tried with a few files and with 32 bit PCM I get the desired files. The problem is that I need to batch processed a huge amount of data. I think I would try with SoX.
Ogg format cannot be converted directly to 32 bit float format. SoX (or any other transcoder) can only decode to 8 or 16 bit, then convert that to 32 bit float.
In the case of SoX this can be achieved with one command:
Converting from Ogg to 16 bit WAV (assuming that it is done correctly) will produce identical audio data to converting from Ogg to 32 bit float PCM (I’ve just tested that to confirm, and the result was a difference of 1/infinity = 0.00000000000000…)