First of all: The EBU Technical Recommendation R98 is dated 1999 and I
don't see any updates available. In 1999, I was digitizing at 16bit,
44.1 or 48K because that was available at the time (OK, 24bit was
available for a major investment in hardware). Nowadays, I record
everything at 32bit float, even the lower sample rates. AFIK, BWF will
document whatever rate & depth you record at. I use iZotope RX for
injest and it writes the metadata to BWF. At WB we used both ProTools
and Pyramix and both did the same although everything was 24bit/48K.
Perhaps I'm not understanding the question.
Corey Bailey Audio Engineering
On 8/6/2018 7:37 PM, Richard L. Hess wrote:
> For the BEXT chunk, the highest bit depth permitted appears to be 24
> bits. How does one code 32 bit fixed and 32 bit floating point? Are 32
> and 32FP acceptable? Or 32Float?
> Also, the standards for sampling frequency stop at 48000, but I assume
> 96000, 192000, and 384000 are acceptable.
> see: https://tech.ebu.ch/docs/r/r098.pdf