On 20.11.2014 18:34, Jurgen Kramer wrote:
DSD is a bit odd wrt LE/BE. Both formats are allowed and used. The device (XMOS based/Marantz+Denon) want the following order of samples: L1 L2 L3 L4 R1 R2 R3 R4. Where Lx/Rx are 8-bit DSD samples, L1-L4 are packed to 1 32-bit sample.
What you described is precisely interleaved big endian U32 sample format. L1 is the MSB byte and L4 is the LSB byte of big-endian U32.
So "L1 L2 L3 L4" is equal to DSD_U32_BE. The currently advertised DSD_U32_LE would be "L4 L3 L2 L1".
I do think real BE will not work.
Why? Your description precisely matches U32_BE. And seems to work correctly for me...
Constructing the U32 by having oldest bit in MSB and newest bit in LSB and sending it out as little-endian results in incorrect output. Adding byteswap to big-endian format results in correct output (either using __builtin_byteswap32() or htonl()).
- Jussi