also try minimum phase
I’m just starting to play around with DSP settings and my first finding is linear introduces some “fuzzyness”
1 Like
FWIW, I like minimum phase for small ensembles (better transients) and linear phase for large groups (better sound stage). I saw Jussi recommend this a while back, and it works for me.
That computes. I tend to listen to small ensembles most of the time by far.
right: same here
Another vote for Minimum Phase here…more pleasing with more types of music for me
1 Like
Played around with it a bit last night. I’m using Synology DS713+ for core, SOtM sMS-200 endpoint to Chord 2Qute. DSD ruled out for now due to use of Synology DS713+. By the way, I’m getting amazing results over a WLAN (Netgear Orbi). I have DS713+ connected upstairs to Orbi router via ethernet and the sMS-200 connected downstairs to the Orbi satellite via ethernet. I’ve never has a drop in connection, except when testing DSD.
Tried max PCM (384 on the 2Qute) with variety of 16/44.1, 24/44, 24/48, 24/88.2, 24/96, 24/176.4 and 24/192. Initially, I liked what I heard. Then I turned DSP off and on again. After awhile, I wasn’t so sure upsampling was actually an improvement. Different, but not necessarily an improvement, except for 16/44.1 files (CD rips and Tidal) where I actually preferred the upsample. Next tried max (power of 2). Basically, the same. Files were upsampling to 352.8 or 384 this time. I also wasn’t loving the sound, again except for when playing 16/44.1 files. Need to test a bit more, but perhaps the 2Qute doesn’t really benefit from upsampling??
This morning I manually configured to ONLY upsample 44.1 material to 176.4, minimum phase filter (which I also preferred). Also, enabled volume leveling…which I love, love, love!! This sounded really good. Again, need to test a bit more, but my preliminary tests reveal only 16/44.1 benefit from upsampling…at least in my system.
played a little today with mine and found what was making some difference is linear or minimum phase but upsampling itself did not
as far as i remember… the 2Qute upsamples everything something like x1000 times so… you are not bypassing its internal filters as it happens with DACs that use off the shelf chips 
1 Like
I’ve closed this topic while we moderate some posts.
Hopefully we can reopen it and get back on topic.
8 Likes
I’ve deleted a number of posts that were off topic or attacks on other users. I did a blanket delete.
I apologize if one of your posts was deleted and you don’t feel it should have been, but I’m not going to spend a lot of time on this.
I’m going to reopen the thread. Please keep it on topic.
The topic is DSP Up-Sampling features in ROON 1.3
Thanks
Greg
3 Likes
Is there a benefit to Roon’s DSP up-sampling with a PCM DAC?
Yes definitely, I’m using a Holo Spring DAC and switch to NOS (No Over Sampling, this by bypass hardware the 8x over sampling) mode and use Roon up sampling from 44.1k to 352.8k. Between the two, I prefer Roon PCM up sampling and sound a lot better.
It is amazing Sonica can do DSD256 in DoP, this requires PCM transport stream of 705.6/768kHz which many of the DACs are out of reach. Does doing a DoP at such high sample rate take up a lot of resources when done in the software level, i e, encapsulation?
What CPU are you using to upsample to DSD256?
Here you go…
I also run RoonCore on my ReadyNAS_516, but it taps out at DSD128. I have ordered an upgrade (CPU and RAM) and hopefully will allow the same up-sample/convert my workstation does at DSD256. I prefer to run all servers on my NAS.
I am really enjoying the OPPO, I have a Bryston BDA-3 and I think the OPPO will take it’s place in my main listening room.
Wow, it needs to run almost to max 4GHz+ at 19% CPU resources! This is really power hungry when convert to PCM to DSD256. I saw low processing index 1.2, anything lower than that, it is going shutter the playback. I can’t imagine if you need to convert to DSD512.
Thanks, it is good to know that an i7-3770K will work, though evidently just barely.
Yeah, it seems like it would require Roon supporting the use of multiple cores on a single stream to make DSD512 feasible. Otherwise it looks like I’d have to stick with HQPlayer.
My i5-6500 can do the same at around 1.1x to Oppo Sonica @ 256. As 512 is not able to be tested yet I can’t say it will work but based on what I have heard then unlikely unless more cores are activated. A 6700K might cope with current limitations on cores