DSD: DoP versus native, and what your streamer supports

Serento plays DSD up to DSD512. What happens to it depends on where it plays.

On a streamer (a renderer), Serento hands over the DSD file exactly as it is on your server. Whether the streamer's DAC then receives it natively or wrapped in PCM frames (DoP) is the streamer's own decision; Serento does not wrap or convert anything. If the streamer takes no DSD, or refuses that DSD rate, and your server offers a PCM version of the same file, Serento hands over that version instead.

On the device itself (this iPhone, iPad or Mac as the output), Serento is the player. With a USB DAC that accepts it, it sends DSD as DoP v1.1: the bitstream unchanged, carried in PCM frames. On iPhone and iPad, a DAC that cannot take the full rate gets the DSD lowered to a rate it accepts, still as DoP; on a Mac it gets PCM instead. PCM also goes to a DAC that accepts no DoP rate at all, to every USB DAC on an iPhone or iPad with 4 GB of memory or less (DoP needs more working memory than those devices can spare), and to any other output: the built-in speaker, headphones, Bluetooth or AirPlay.

The format badge tells you which of these you are getting: tap it (on a Mac, hover over it) to read BIT-PERFECT, DSD OVER PCM, the DSD rate the DAC receives when it was lowered, or DSD TO PCM.

Seeking within a DSD track is not universally supported. Some renderers, Rockna's among them, accept a DSD stream happily but reject or ignore a seek request inside it. That is a limitation of the device's firmware, not something Serento can work around, because there is no way to seek a DSD stream except by asking the renderer to.

If a DSD album shows its format badge without a rate, the server did not report a trustworthy sample rate for it. Serento probes the file header to recover the real rate rather than displaying one it does not trust.