Downmix multichannel DSD

I have a number of files which show up and do not play . They are playable on Roon.

Can you tell me which music server do you use? MinimServer, AssetUPnP, Plex, other?

Serento does not intervene in the playback. It only tells the streamer what files to play from the server. In this relationship, the server and the streamer must work together to play the file correctly.

If the server cannot read multi channel DSD then it cannot send the file to the streamer.

Roon acts like a man in the middle and does something to your files. Roon controls what the server does and what the streamer does because they wrote the software for both. That means that if they want to downmix DSD files before reaching your streamer they can.

I only wrote software for the controller app. In your context, with Serento, the streamer has its own software and the server its own. I cannot control any of them, I can only make sure Serento plays well with both.

If you are using MinimServer or AssetUPnP there are ways to convert the files to the desired sample rate or channels using their provided solutions (MinimStream for MinimServer and the built in converter in AssetUPnP).

Lorin.

@JKND68 Are you looking to downmix MultiChannel DSD to 2 Channels or to MultiChannel PCM / WAV / Flac?

TLDR: Roon will probably have been transcoding the music on the fly to lower quality (Nerd explanations follow…)

In either case, I’m not currently aware of any way to downmix DSD which doesn’t involve turning it into PCM for either - The software / tooling just isn’t there.

I spent a while looking at this recently due to having some DSD128 and DSD256 files, but I have 2 streamers - 1 plays only PCM and the other also does DSD64.

In the end I was able to get the files to play on the older streamer by forcing a conversion to PCM for all DSD for that streamer but, ironically, not on the DSD Native streamer as I couldn’t configure MinimStreamer to both allow DSD64 AND not higher. To get around it I looked at converting the files down locally to DSD64 and then realised you can’t actually do that at all, so I converted them to PCM 24/88 instead and gave up - I had far more DSD64 content that I didn’t want degraded.

DSD is a bit of a minefield. If Roon was handling this I strongly suspect it’s either using DOPWAV (DSD over PCM) or down-converting to 24/96 or 16/44.1 PCM. I think for DOPWAV it would need support on the server and streamer side, so I would expect Serento to “just work” if that was the case.

I’m using Plex, my streamer is Eversolo T8.

Yes I’m looking to downconvert multichannel DSD to two channels DSD or even any pcm variant.

Since I’m running on Mac, as you know it runs DOP , DSD over pcm.

Maybe I can try AssetUPnP?

If the server software runs to DOP you should be good with stereo as long as the Eversolo supports what it throws at it. Multichannel is, as far as I know, wholly different - you can’t DoP MultiChannel to 2 Channel - that needs hefty encoding and channel merging with verspecific tools down to a 2ch source.

I had real problems with DOP as the bitrate worked out as something like 32/384 and my streamers just couldn’t handle WAVs at that rate - my more modern one would handle FLACs at that rate but WAV OR RAW PCM of that size would just be too much bandwidth.

It looks like the T8 is a transport that tops out at 24/192 so the DoPWAV rate may be too high for you. I think @lorin.bute uses Asset so he may be able to advise on that front.

I’ve done some reading up on this to try and understand what’s going on under the hood with the Roon comparison you mentioned @JKND68 - this post in the Roon community is very enlightening.

Ignoring the does it / does it not work back and forth, the images posted show the steps Roon goes through - It grabs the source, converts DSD to super high rate PCM, converts it again, remixes it, converts it again and then sends it to your device.

From an audiophile perspective, none of this is good stuff - 5.1 etc to 2.0 is never lossless, mixing and volume levelling is really important and inconsistent between different source types, which can lead to clipping and other nasty artifacts. A 5.1 downmix will never be right for a stereo soundfield because the mix wasn’t engineered for it. I have a vinyl in a David Bowie set with some tracks from the SACD Mix of Heathen and it’s just flat as hell in Stereo.

Serento’s very “hands off” explicitly to keep optimal quality between the source and your ears.

I recommend keeping dedicated 5.1 and 2.0 mixes - I manage this nicely with Serento now by representing all my Multichannel content like an SACD - I have a disc 1 with a Disc Subtitle of either PCM Stereo or DSD Stereo - depends what I’ve got - and a second disc in the album with a Disc Subtitle set as 5.1 Surround, then I choose which mix to play based on the device.
I don’t have any devices that play multichannel DSD, so all my Multichannel discs are FLAC files - I have written a load of scripts for converting extracted DSD, DTS-HD and Atmos MKVs.