Question about little-endian audio streams

[amtm]

So as I was going through cleaning up the source code of delaycut I noticed that whenever it was doing byte comparisons of fields, it was always assuming big-endian order for (e-)ac3/dts streams. Now, this is obviously a reasonable assumption, as bot the AC3 and DTS ETSI specs use big-endian ordering, as I would assume that people would be reporting issue if it was failing on little-endian streams they had. I can see that the ac3 spec says that little-endian streams are compliant and according to this (http://wiki.multimedia.cx/index.php?title=DTS#Frame_format) there are little-endian formats for DTS. So my question is, do these little-endian streams actually exist with enough frequency that it would be worth any time coding in support for them?

[tebasuna51]

I only see 14-bit little-endian in dtswav (wav output from Surcode/Suite).
There are dts with a fake wav header to be burned in CD's and play surround with appropriate CD players.
There are tools to convert them in standard DTS 16 bits BE (BeSplit, Ac3Filter tools)