I primarily use the Innuos streamer as the source for the Rossini Dac over USB. Without getting into an extended discussion regarding preferred connections for now, saying only that I have a basic understanding that dCS wants to break the connection to isolate for sound quality, but more to the point to ask about what is it exactly that triggers the USB input to reset itself.
As this issue is getting my attention lately, I’ve been playing a variety of sources from the Innuos to see if the file location has an effect on the switching action of the USB connection. And it does seem to.
I can hear the dCS unit when it does switch between tracks so this is something I’ve been listening for preceding a drop out delay at the start of the next track. At which time but not always, up to about a second or two of music is eliminated/ skipped and the track may begin during music playback accompanied by brief distortion. There does seem to be some correlation. I need to make sure it’s only happening proceeding a switch or resetting of that input. But so far it doesn’t seem to switch when I’m listening to albums or from random tracks within an album from Qobuz which is my only external streaming source with the Innuos.
For now I’d venture to say that truncation only happens when playing consecutively tracks from different albums within an Innuos playlist, when it becomes a notable distraction or occurs at all. I’ve been listening to and creating playlists of late so this may the reason it’s getting my attention.
The reason as ask for some technical confirmation of this is so that I can pass this information along to Innuous and that dCS or Innuos might consider this issue as somthing that might be resolvable.
I know that dCS considers it dacs to be best used with its own streamers and this issue points to that advise. I can only say that I have preferred USB for listening reasons alone, dependent on cable, to the direct network connection. Since both these two brands are now respected and often see together at shows, myself and perhaps others would be interested in a fix for this issue.
Thanks,
Rossini ( or other dCS units) is only audible when a new track format is encountered and the unit has to reset itself to change the processing to fit. You can usually hear a relay switch.
This is unlikely to occur playing a commercial album as all of the tracks have the same format. However things are not the same with a playlist where every track could be different moving from , say, 16/44.1 to DSD to 24/96 etc.
I am guessing that you are having this issue because when such tracks are played consecutively Rossini has not time to reset before play starts. This is suggests that you have Buffer disabled. Please look at your Rossini settings ( settings/Device) and make sure that it is engaged ( see page 39 of the user manual).
if you prefer USB to network connection subjectively that is fine if unusual but I wonder if there is actually a further issue to be dealt with, however that lies beyond this topic.
Pete, unless I’m mistaken, the Buffer setting only applies to the synchronous interfaces, not the USB streaming interfaces.
From John’s description of the problem, seems like some compatibility issue with the Innuous USB out (I don’t recall any similar issue with Melco/USB on my dCS systems).
I’m pretty sure there are other Innuos users on dcs.community, but rarely using USB, but maybe one of them can try and see if there’s any issue.
Yes, perhaps other Innuos users can help. I can recall some USB issues that I had many years ago when I ran a PC based computer audio system. However this turned out to be an early version of Qobuz and the issues disappeared with a subsequent software revision. However these were not similar in nature to john’s. Nevertheless that USB system also used JRiver MC and that certainly had a buffer option and it was needed.
Check your DAC is operating in USB Audio Class 2 mode.
If you have a ZENmini Mk2, ZEN Mk2 or ZENith Mk2, do note there are incompatibilities with some USB controllers on DACs, particularly with some versions of XMOS.
These 2 settings from Innuos Statement’s menu might also be involved in your issue:
Thanks everyone for chiming in so quickly. The combined input seems to have addressed my issue so far.
Thanks Pete, for the explanation from the dcs side of things. That proved to be the crux of the issue. The delay with buffer settings on in Mosaic is variable according to the intruction manual “ 0.72 with 44.1kS/s and 0.16 with 192kS/s data. “. So a change from a low to high frequency format might not always give the Dac time to switch before music begins. That seemed to bare out to not always being enough delay after I changed my setting to buffer ON. But my testing was brief so I can’t say whether Anups comment about not applying to an asynchrones conventions bore out.
Erno, My Rossini was set to USB Class 2 but the buffer was OFF. With the buffer ON things seemed to improved but maybe not completely. I got the first note but didn’t always get the initial attack. The controls in the settings of my Innuos Statement streamer however offer custom Sampling Rate Change Delay settings of 100, 250, 500 or 1000 so I choose 1000ms or 1 second which might be a little long combined with the dCS delay but with that delay in my brief testing I could always be sure I’m getting the full attack and anyway I find I enjoy a little more space to transition between different music / musicians in my playlists. I also switched the USB latency to NORMAL from LOW. Don’t know how that’s different in effect from Sampling Rate Change Delay.
Parr - “ if you prefer USB to network connection subjectively that is fine if unusual but I wonder if there is actually a further issue to be dealt with, however that lies beyond this topic. “
This is an interesting question that I’ll just touch upon. I found the connection to the Dac to be very noticeable as in a definite improvement in micro details that help define the space or “ air “ in my soundstage and a loss of harshness when switching between inputs, from network to USB through the Innuos vs network straight into Rossini, but was dependent on using the Shunyata Sigma USB cable. But that was in my previous space so it would be perhaps too difficult to try and test this again. My new space is much smaller and was a difficult adjustment for me and my system requiring a change in speakers to reduce glare / brightness and try to regain some that micro detail in the soundstage, not altogether successfully. I’m not sure I can now tell the difference between the two inputs now, so I’ll either have to try an address this problem with treatments and or bide my time until a larger space is available for me.
Anup “ I’m pretty sure there are other Innuos users on dcs.community, but rarely using USB, but maybe one of them can try and see if there’s any issue.”
Thanks for reminding me of possible options. When I first got my Innuos Statement I thought I would be able to use the network connection from it to the Rossini but I wasn’t able to make it work. It’s been several years but I think Nuno was helping me and we couldn’t seem to make it happen so he suggested sticking with USB. Even with my current reduced sound quality I would be interest to understand how that’s possible, especially since you say that USB is not the common connection with owners of both products. I’m surprised to learn that !
I wonder if I should open another question in that regard.
Wait - Not So Fast
As I’ve been listening further for the last several days I have noticed that the settings that had seemed to give me good results, ie enough delay on the Innuos audio setting of the maximum of 1 second ( SAMPLING RATE CHANGE DELAY ) and the buffer on in the mosaic Dac settings is proving to be insufficient. For example in a playlist I have a track at 16/44 followed by another at 24/192 which is noticeably truncated. Changing either setting to buffer off or lesss innuos delay only makes matters worse. So it appears that that one or the combination of both delays remains inadequate. I find it curious that the dCS Rossini instruction state that the delay for higher rate data 24/192 gets significantly less delay than 16/44, 0.72 vs 0.16. I suppose that means that switching from lower to higher rates happens more quickly ie less delay than the reverse, switching from higher to lower. I bracketed the 24/192 track with the 16/44. The truncation is still slightly present even on the switch to a lower rate, though it is of course a different track. This problem also reveals the lack of much of a silent space within the track itself of course this is just one example. It’s odd because previously, last week I seemed to be getting clean results.
This is really a technical question whose answer depends on what circumstances the buffer was designed for. I suspect that the answer may lie with dCS techs so I will tag @Phil .
Can you please drop an email to [email protected] … I’m not sure why I didn’t see this thread until @PAR flagged it.
We’re not doing anything odd with the USB input (we’re not disconnecting it on the fly) and we’ve not had any similar issues reported previously with Innuos (and other streamers / players) via USB so I think that the issue you are having is a localised one that is specific to your system.
Ok, is there anything off the top of your head you can think of that I might check on the dCS or the Innuos side, before I start an inquire over at their site ?
Our UK distributor has been using Innuos servers with dCS kit for a number of years and so they should definitely work and work correctly.
Can I ask what you are using as a USB cable between the two devices? If it isn’t the cable that was supplied with your Rossini can you dig that out and confirm whether you get the same issue with that as you do with your current cable?
I’ve been using the Shunyata Research Sigma USB since 2018 or so,
but have only recently been creating extended playlists. I’ll have to did out the Rossini box to see if it’s still in there. Will check though I doubt that Shunyata is doing anything to go out of USB’s basic specs.
Have listening with the dCS cable now. Because it is track dependent, I assume because the presence of silence proceeding the music or the track fades in I’m looking for definite beginnings, like a first note or length of first note. I am detecting a missing note. I should now change inputs on the Rossini to network and try the same track that way.
Just repeating the same track on the network input, finding that it seems to include the missing note. Playing music in sequence is a bit different however with the Mosaic app not possible with playlists ? ,because the track is in the middle of the playlist and unlike in the Innuos player I can start a playlist anywhere - the next track is played rather than just a single track when in Mosaic. In which case I have to play a low res track followed by then selecting a high res track. It’s too time intensive to wait until the player plays the right combination of tracks, unless I’m not understanding how the player plays playlists. In either case though or just as with the Innuos app., I can hear the Dac switching something between track frequency changes.
I don’t seem to be missing the note over the Mosaic network player.
I had thought that switching noise was a part of the USB mechanism magnetically reestablishing galvanic isolation but it’s the same sound over network.
.
