Streaming clarification

[markush]

I have been using an EtherRegen with Farad Super3 power supply (top configuration) as well as galvanic isolation via a Sonore OpticalModule Deluxe V2 (also with a Super3) and Finisar SFPs as well as Audioquest Diamond Ethernet cabling all the way.
But what I really would recommend (there is a 30 day free trial period anyway) is the Network Acoustics Muon streaming system which acts like a noise filter.

Regarding upsampling:
Has anyone directly compared DSD64 vs PCM 192/24?

[daniel.avasilichioaei]

...I thought that everything sounded cleaner but "too clean" in a sense. One of the things I like about my system is how natural voices sound, especially the sense it delivers of voices coming from breath. Tolstoy's voice lacked that sense of breath and I thought the sense of chest tone had diminished. I liked the cleanness of the sound, the voice sounded a little sweeter as a result, but that added sweetness came at the cost of losing my sense of her as a breathing person. As a result I lost some of the sense of "naturalness" that I like and that's not a price I like paying, even if things sound a little sweeter and cleaner in some ways.

This is exactly my impression, not only for PCM to DSD conversion, but for any DSP I tried in Roon. Something of the naturalness and purity of the sound is lost...

I have the same impression when I activate the Devialet's DSP (for example, SAM).

[David A]

This is exactly my impression, not only for PCM to DSD conversion, but for any DSP I tried in Roon. Something of the naturalness and purity of the sound is lost...

I have the same impression when I activate the Devialet's DSP (for example, SAM).

Interesting. I have a bit of a love/hate relationship with SAM. When I got my 130 Pro nearly 6 years ago, before the CI board and the FW update that turned it into a 140 Pro, I was using a pair of Dynaudio Contour 1.3 SEs and I quite liked SAM. When I upgraded my speakers to Focal Sopra  2s I had difficulties with SAM. I've got an awkward room, it's L-shaped with 2 large open archways into it and no place where I can place the speakers in a symmetrical relationship to side walls because of the shape and openings. As I played with speaker placement for the Sopras and also with the placement of my physical acoustic treatments which are free standing I discovered that the effect of SAM was strongly dependent on what the room was doing to the sound in the bass region. Some speaker/acoustic treatment configurations made the use of SAM both highly noticeable and highly objectionable to me while SAM had a positive result with no drawbacks I could notice in other configurations provided I kept the SAM level very low. It took a long time but I eventually found a setup where the speaker and acoustic treatment placement worked really well without SAM and using SAM at 0%, just for the phase correction benefit, gave a subtle positive benefit without any drawbacks I could hear. That's the way things are running now.

I do think SAM is a useful feature with positive benefits but that its benefits are highly sensitive to room and speaker setup. Devialet's limited info on SAM suggests that you can just dial in the level setting for best results but I found that simply is not the case. SAM is said to work below 150 Hz and room effects dominate what we hear in those regions, exaggerating some frequencies and reducing other frequencies. SAM applies phase correction and the speaker's phase response varies immensely within the SAM range. Put the 2 together and the results are unpredictable. As I experimented over time I found setups I liked where SAM had a positive effect and other setups I liked where it had a negative effect. In setups which I didn't like, SAM didn't "fix" the things I didn't like.

SAM is definitely not a tool that always delivers good bass and it doesn't always make good bass better.  If you've tuned your room setup to get a result you really like then SAM may be just the right finishing touch or it may be the wrong finishing touch but if your room setup isn't giving you the result you want then SAM is not going to fix that. In my view SAM is worth trying if you like the result you're getting without it but you may find it doesn't work for you, at least with your setup at that time. If you're not getting a result you like without SAM then you definitely need to start playing with your setup in order to get a result you like because SAM on it's own won't fix your problems.

[audio_engr]

Arup,

Many thanks for the reply. You convinced me to try it but I've gone back to streaming PCM as PCM at its original resolution.

The first thing I noticed when I tried the DSD conversion was the drop in level. I'd been expecting the 6 dB drop associated with the DSD  to PCM conversion in the Devialet, I deal with that when I stream the DSD64 files I have on my Nucleus+ to my 140 Pro, but I'd forgotten about the 3dB loss in the Headroom settings which I also copied from that post on the Roon forum.

Once I had the level back up to my usual listening level I had mixed feelings. I don't play classical very often and rarely orchestral music. Mostly I play small group jazz. I put on an album by Viktoria Tolstoy, a 96/24 high res download and tried that. I thought that everything sounded cleaner but "too clean" in a sense. One of the things I like about my system is how natural voices sound, especially the sense it delivers of voices coming from breath. Tolstoy's voice lacked that sense of breath and I thought the sense of chest tone had diminished. I liked the cleanness of the sound, the voice sounded a little sweeter as a result, but that added sweetness came at the cost of losing my sense of her as a breathing person. As a result I lost some of the sense of "naturalness" that I like and that's not a price I like paying, even if things sound a little sweeter and cleaner in some ways.

I'll add a caveat to that comment. There's a part of me which thinks that what I should do is to live with the change for a week in order to get used to it because when I do that I stop hearing the differences and I start hearing the sound that I'm getting. I need time to become accustomed to a change in sound so that I can appreciate the change for what it offers rather than simply noticing what it changes. When I make changes I usually go that way but with some changes, and this was one of those occasions, I seem to have an immediate reaction which causes me to put things back the way they were. I've learnt over time that I can usually trust those reactions. I may try it again in a day or two and try to stick with it for some days but at the moment I don't want to do that. I'll see if I still feel the same way in a couple of days after letting what I heard work on my subconscious in the background.

At present I can understand why someone might like and prefer converting PCM to DSD before streaming to the Devialet, it does have some attractions for me, but those attractions seem to come at the cost of the loss of something else that I really like.

RE the EtherREGEN 2:  on of the changes is that the single ethernet port on the B side is going to be a gigabit port rather than a 100T port. That concerns me. I'm still not convinced that Devialet have completely solved the RAAT/ethernet problem and Devialet's last recommendation on that issue was to use a 100T connection to the Devialet. That 100T port on the current model is one of its big attractions for me. Most of the time I'm running RAAT over ethernet without problems but for some reason every now and then I seem to have a week or so when the RAAT connection keeps dropping in and out and my network connection between  the Nucleus and the Devialet becomes unstable. I turn things off and back on and eventually things go back to being stable again for months. It's frustrating because I can go into Roon's Audio settings menu and I can see the Devialet available for Airplay and also for connection using Devialet's own AIR protocol but the Roon Ready connection just doesn't show. Even worse, I can sometimes see it come back into existence in Roon's Audio settings menu and then just disappear again a couple of seconds later on occasion. I'm reluctant to go back to using a gigabit connection to the Devialet while this sort of problem still exists, even if it is sporadic and only occurs at intervals of several months.

David

David, 

So good to hear from you. 

For my DSD64 setting in Roon, I use Smooth, minimum phase setting. And I do not use Headroom management as I've never had a case of overload. 

We humans whenever we sit down for an audition session, each time our frame of mind is different thus we interpret the sonic picture painted differently each session. A week's time is a great-great idea to allow for consistency and also to figure out if it works for you. Mind you, its different for each one of us; what may float my boat may not for you and vice versa. 

I understand your reluctancy for etherREGEN Gen-2 and I would similarly lean in that same direction as yours. It took ages to fix the 'dropping' issue and now that its done, none of us want it back. 

As for DSP functions whether in Roon or SAM in Devialet, it just didn't work for me at all. SAM with my speakers and in my setup was a strictly no-no. I only use "Sample Rate Conversion" as a DSP function in Roon and even any of the other DSP functions in Roon just didn't work for me as it severely robbed the naturalness of the sonic reproduction. 

- Arup

[audio_engr]

I have been using an EtherRegen with Farad Super3 power supply (top configuration) as well as galvanic isolation via a Sonore OpticalModule Deluxe V2 (also with a Super3) and Finisar SFPs as well as Audioquest Diamond Ethernet cabling all the way.
But what I really would recommend (there is a 30 day free trial period anyway) is the Network Acoustics Muon streaming system which acts like a noise filter.

Regarding upsampling:
Has anyone directly compared DSD64 vs PCM 192/24?

I got the Ferrum Audio Hypsos (2 units) very recently - one for my Roon Nucleus and the other for my etherREGEN. I also use Finisar SFP and AQ Diamond from Roon Nucleus to etherREGEN and AQ Diamond from etherREGEN to Dev 440 Pro. 

I have to say that once the Hypsos settled in (I also use a Shakti Stone AIR over each Hypsos and IsoAcoustics footers too), it was just absolutely wonderful esp. when you can slightly alter the DC voltage that Hypsos allows and fine tune the sound to exactly zero-in to what you like. 

Directly comparing DSD64 vs. PCM 24/192 ? 
Yes - I have and results vary depending on which Mastering Engineer did the job. For Classical music (that has become my primary music of late), most of the time DSD64 wins.

[zdenes]

@audio_engr the answer to your original question is, your Dev is your streamer, it has a streaming board inside, called Core Infinity by Dev. You Nucleus is your server.

I have quite an elaborate streaming front end outside the Dev using the AES/EBU input (see my signature). I find this much better than the CI (yes, mine doesn't have it but I borrowed a Pro unit to try and decided against the upgrade).

On the topic of upscaling, I tried all possible combinations, such as upscaling everything to DSD64, upscaling everything to 192kHz, or leaving at native DSD or native PCM. I found after a lot of listening "tests" that I prefer the last one and in fact, I convert DSD to PCM in the server. My server runs HQ Player under Roon.

Devialet immediately converts every incoming signal to 40bit/384kHz for internal processing and DSP. I think that happens even if you don't use any of the DSP functions. However, since the two Devialet units are connected through spdif or hybrid AES/EBU that can only handle up to 192kHz and no DSD (as output), I guess that first it upscales/downscales to 192kHz or leaves PCM as is before it reaches the companion unit. Anyhow, all these conversions involve a lot of maths with filter functions, etc and I found that using them in series is compromising the sound quality. At least that is my preference.

[David A]

I would argue that @audio_engr's Nucleus is the streamer for grammatical reasons.

The suffix "er" after a noun indicates something that pereforms an action related to the noun it follows. Killers kill and streamers stream. A streamer allows you to initiate connections to a source of files, choose the files you wish to play, and direct those files to the device which is going to process the digital stream. The "streaming" board in the Dev receives streams but it doesn't control them or direct them elsewhere. It performs none of the control functions associated with a streamer.

When @audio_engr streams music from his own music library he uses Roon control software to connect with the Roon core software running on his Nucleus and that core will connect the drive on which his music is stored and stream the chosen music to his Dev. If he uses Tidal or Qobuz then he will also be using the same software and his Roon Core on the Nucleus to connect to the Tidal/Qobuz server and direct the stream to his Dev. Along the way Rool will also use theNucleus' processor hardware to do things like converting PCM to DSD. The Devialet is simply the end point receiving the stream. That doesn't make it a streamer, it's not a doer but only a receiver.

If the Dev was a streamer it would provide a menu system of its own which allowed you to select and control the stream. It has no such menu system.

[maxijazz]

Yep. CI card is endpoint/renderer.

[BoyScout]

Sorry,
But for those that only use roon as a source (my case) what´s the best set up?
I don´t listen from any other format. Simple system, Tidal via roon to Dev straight to my ears, but i´d like to optimize the way i´m listenning too.
David, i saw the setup you mentioned from roon forum but, unless you tell me the opposite, looks unnecessary in my case.
Thanks.

[listener]

IMHO it is some kind of taste already within this price range, but the extras I tried, first best results / improvements:

- use Roon DSP function for a room correction (less than 500,-)
(it needs a laptop, microfon and "Audiovero Acourate" software - then it corrects time and frequency range. Software creates direct filter for import to roon - it needs some time to learn the program or a consultant that does the job for you)
- amount of improvement vs cost
- amount of improvement vs cost
- use active (never passive) current regulator
(this is more costly, in my case it delivers a better soundstage and benefits that become clear when you move it away again)
- for the same money, you would get a used second 220, I never tried that
- improvement of network router like EtherRegen or other
(have done these improvements but it depends a lot from your streaming device and the result differs on each setup)
- as a last comment - if you think the streamer in your setup is in its class adequate already best for that system?