? ? 
I assume @Katzky means that they hope the Vivaldi replacement (whenever it comes) doesn’t need a separate box for the streamer. I’m with them on that, unless it’s a massive sound quality upgrade to keep it separate.
1 Like
Thanks Ben, that seems a reasonable interpretation. Of course, from my point of view, if the successor is in the line of Bartok, Rossini and Vivaldi One in that upsampling cannot be disabled I will not be a customer.
Good point — had forgotten about that, Pete. I’d love an update (hopefully it’s just firmware) that allows the whole range to disable upsampling.
I’ve not yet had the pleasure of that unique-to-Vivaldi comparison!
1 Like
Hello Pete @PAR, I’m happy to report that Vivaldi Upsampler, Clock and DAC are now playing under my roof as well. And they look superb on the titanium rack. Complementary industrial design.
I might have read it from you elsewhere - what was your reservation about Upsampling? And maybe you have comments on the following experience relating to the upsampler? I have read the long and thorough exposition by a fellow forum member in another thread about his learnings in setting upsamplers. That was illuminating but question marks remain for the actual practical application. (And there is one sentence in the paragraph that relates to the future of the Upsampler drawing from Bartók.)
I am having a very steep learning curve when it comes to setting the Upsampler together with the DAC in order to achieve truly great results. Only few rather intuitive patterns yet. Some recordings I enjoy most without upsampling some become awesome when going from 44.1 KHz to 48 KHz or maybe 88.2 or 96 KHz. Some when going from 48 KHz to 192 KHz and various other empirical combinations. And there might be elements of recording quality or music complexity in determining the best setting regardless of bit depth or input rate. It’s different from the Bartók, as there I could only choose DXD or DSD. On the Bartók the DSD setting was - at least for me - a valid catch all. It might be, that the algorithms in the Bartók are different from the ones in the Vivaldi Upsampler - maybe more forgiving? Just a wild speculation based on comparative listening experience upgrading from my already highly pleasing prior setup with Bartók and Vivaldi clock. DSD is in my ears not a satisfactory catch all setting on the Vivaldis when aiming for optimum results. The tactic so far has been to set the dCS recommended filters for Upsampler and DAC per input and output rate of the Upsampler. They are then remembered by both devices. And then experiment just with Upsampler output rate and DAC mapper 1 or 3 for recordings I want to enjoy in full and in depth. Two minutes of work for an album long reward 
The equivalent of putting vinyl through the record cleaner and lowering the needle onto the first groove? Or maybe just 
… For some recordings the upsampled output sounds sublime as opposed to a really good result sans upsampling. And it’s often not the very high output rates or DSD even. Somewhat of a riddle for me still.
Also, do we know whether or how the upsampling algorithms change for different input and output rate combinations? And finally, you switch Upsampling fully off via the Upsampler Clone Mode, no? Otherwise it would upsample from 16 to 24 bit depth?
That’s a more accurate way to describe my motivation. Doubt I’m in the target audience for the new Stravinsky but it would be weird for a dac released in 2023 to not include an Ethernet input. Almost as weird as a company still using Din connections in the year 2004.
Combining too many functions into a single box would take me off the target market; I value the Vivaldi stack precisely because of its “disaggregated” architecture. 
That said, I can understand the need for a better inter-chassis communication interface/protocol than using AES as a backchannel. dCS has their work cut out for them, but if anyone can, it’s dCS; they have more experience with digital interfaces than pretty much any other high-end [audio system] vendor out there.
1 Like
I am not necessarily against upsampling per se. It is, after all, necessary for some modern DAC processes e.g. in Ring DAC. I am specifically talking about the layer of processing that occurs when the optional ( on Vivaldi) " Upsample" function is selected. This also relates only to dCS products as I have been using them exclusively for over 22 years and have therefore had no opportunity to listen to any competitor’s implementation of upsampling.
As I have had the same reservation about such upsampling for 20 of those years it cannot be the result of a different algorithm implemented specifically in Vivaldi.
My problem is with the spatial inconsistencies that are introduced immediately that the upsample option is engaged. I see this as a factor only of relevance to those of us who are classical music listeners as other genres of music do not have bona fide soundstages in a similar manner to classical i.e. are recorded in such a way as to resemble what would be heard at a live event . Incidentally I seem to be unusual in having this sensitivity and therefore often have doubts about it. So I do switch on upsampling now and again to see if my opinion has altered or if any other equipment change has affected it. Coincidentally I carried out this process only two days ago , once again confirming my feelings as the vocal soloist was given incoherent placement ( was she centre stage or slightly to the right or somehow both at the same time?) , the orchestra no longer had depth or layering front to back but was now rendered in a single plane behind the vocalist. I stuck it out for at least 45 seconds 
.
BTW I have had this opinion through three generations of dCS equipment, four different pre-amps, three complete changes of cable loom and three iterations of active speaker. So I think I can dispense with the idea that it is a problem caused elsewhere by some odd combination of gear. It is also the same whatever source I use; silver disc, network replay from Qobuz, UPnP or internet radio.
I am jealous of the majority who can enjoy whatever benefits they may find in upsampling without finding a downside.
I would advise that you experiment with different settings of filter, mapping etc, until you find combinations that suit you for most music then leave them alone. The objective is to listen to music and not to press buttons. I will add one thing in this regard though. As I mentioned my problem with upsampling dates well back and one thing that emphasised my preference for not using it was the introduction during an upgrade to the second dCS generation of an asymmetric filter for 44.1Ks/s. This is still filter 5 for that resolution. I was recommended to try it many years ago by a senior person at dCS and I pass on that recommendation.
Yes, clone mode switches off upsampling completely including the padding that creates a 24 bit word from a 16 bit one. That, however, reminds me that though I did not change ( upsample) sample rate I still changed 16 to 24 bit with Paganini. It sounded a little better (for no obvious reason) but I have found it unnecessary with Vivaldi. There is a clue about my issue in that if I do not engage clone mode but choose the sample rate to match the incoming data e.g. I end up with 24/44.1 for a 16/44.1 input I do not hear the spatial inconsistencies.
2 Likes
Bit confused as to what the upsampler does. If you turn off upsampling then what is the point of using it?
Hello @Katzky Vivaldi DAC has no streaming component inside. Unlike Bartók. Vivaldi DAC can not take a Qobuz account login and play music from Qobuz. Neither can the Vivaldi DAC address an UPnP server or Roon core and play music from it. In the dCS product family you could use the discontinued network bridge to enable streaming. To stay in that terminology - the upsampler becomes a pure network bridge (streaming player) when upsampling is off. I have no first hand comparison between upsampler and network bridge. Others here reported that the upsampler sounds markedly better than the discontinued network bridge.
As Marco explains the Upsampler is not just an upsampler. It is also the network player part of the Vivaldi range.
But you need the optional “volume control” box.
The volume control box is either the Vivaldi DAC or a pre amplifier.
Thank you Pete @PAR, that’s what I will aim for. I’m developing a better intuitive feel for what the different settings do to different music. The ‘pressing buttons’ phase is down to half a minute now. 
I listen to headphones only, so the depth of the soundstage is hard to perceive for me. I did encounter recordings though, where upsampling made placement of vocals fuzzy or even slightly wandering left and right. Moreover, when some records were upsampled the music presentation became too orderly or more boring.
Does padding mean the eight extra bits are filled with zeroes? My understanding was you can represent any given waveform that much better with more available values - 24 bit instead of 16 bit (subdivision of the Y-axis of the waveform whereas the sampling rate determines how often you slice the waveform - subdivision of the X-Axis). I only looked at Andrew’s article on the Ring DAC but not further. Would an upsampler in some form interpolate between the values a 16 bit word delivers to create the 24-bit word?
Attention to detail!Yes it does but whether or not there is interpolation in this specific case needs technical advice from dCS as James Cook’s series of articles has not really touched on this subject ( so far) and Andrew’s was really to explain that DXD is a resolution in its own right and not a method of upsampling. I would think however that using interpolation would offer only a small gain, if any, as a 24 bit word offers a 144dB dynamic range which for music is an over specification and effectively means that part of this capacity can contain no musically relevant data. Paul Miller ( Group Editor AV Tech Media, President EISA etc.) has speculated that simply using 24 bit words with 24 bit processors provides a benefit in processing “ease” and it is that which may result in a sonic benefit.
I dug a bit deeper and pulled the following from here
(…)
That doesn’t give us the answer to the question whether Vivaldi Upsampler does any processing beyond adding zeroes in order to get from 16 to 24 bit depth. It tells us however there is value in upsampling/processing the bit depth beyond adding zeroes.
1 Like
Thanks for searching that out. I am happy that I am still learning things in my seventies.
3 Likes
The Rossini manual indicates that upsampling is optional.