I am also using Chip Audette's F32 library. There are differences in the timing hardware for 4.x that are not supported, and are not the same as for T3.6, say. So, I suggest you use the PJRC teensy 16-bit input/output and then use Chip's AudioConvert_I16toF32 and AudioConvert_F32toI16, see
https://github.com/chipaudette/OpenAudio_ArduinoLibrary/blob/master/AudioConvert_F32.h This also allows you to use other CODEC options not included in Chip's F32 library.
Still, the compiler won't be happy until you
* Remove input_i2s_f32.h and .cpp
* Remove output_i2s_f32.h and .cpp
and because of need for simple fixes
* Remove synth_pinknoise_f32.h and .cpp
* Remove synth_whitenoise_f32.h and .cpp
I think the rest will compile.
The conversion blocks are fast and make intermixing of the float and integer stuff. I just posted an example with similar routines built in so that they would just look like 16-bit objects:
https://forum.pjrc.com/threads/61261-Generalized-FIR-Filter-Block
If you are using an object from the F32 library, I would look it over and compare it with the similar file in the Tympan library, which may be more up-to-date. See
https://github.com/Tympan/Tympan_Library
I'm also curious as to your radio projects, as I have been fixing up some radio specific blocks for the F32 library with T3.6 and T4.x. These will be posted soon, I hope!
Bob W7PUA