Ahh, I get what you mean; You regularly upsample everything to DSD/DSD2, but would like DXD to be bypassed.
It’s not really similar to the “DSD Pass-through” feature though - that’s bypassing a whole format, whereas you’re requesting the bypass of just one particular PCM rate.
If I understood it right, if this feature were there you did not have to switch to DXD upsampling (from ur regular DSD/DSD2) in case the input is DXD. Is that right ?
I obviously did not expect that you to express any kind of commitment, but some feedback on whether or not it would be difficult to do, pointless because of something I missed etc…
For example, with such high resolution streams, what is the value added of the upsampler ?
In my case I find very little interesting music released in higher resolutions. So upsampling to DXD or DSDX2 provides me with (IMHO) better SQ for CD quality releases.
If I understood correctly the explanation from @Phil, the Clone mode cannot be used for this requirement you have. The Clone mode can be used just to switch the upsampling to happen either in the Dac (when On) or in Upsampler (when Off).
Given that understanding, to me it is a valid feature request. Why either of the Dac and Upampler should do any kind of Upsampling if the input itself is DXD ?
I shall also extend this requirement for Rossini. Doing unnecessary upsampling does not make sense to me if the input is already in the highest form.
If you look at the information regarding how the Ring DAC works then you’ll see that no matter what the input format (whether DSD, DSDx2 or DXD) there is always a need for the input audio data to be upsampled for the Ring DAC.
But, if I understand it correctly, Ring Dac is handling DSD inputs without upsampling.
The above understanding is based on this part in manual in the Upsampling section - ‘This feature does not apply to DSD data, which is processed differently.’. But surely ‘differently’ does not necessarily mean that it is not doing some other kind of upsampling/some other processing’.
It’s not really something for us to discuss or make judgements on though - we can (and do) feed suggestions into the R&D guys and we are here to help and answer questions wherever possible but there are things that we really do need to keep out of.
If you are feeding the DAC straight DXD or double speed DSD and nothing else then an Upsampler may well not be of relevance to you…
Dac itself (inside the Vivaldi DAC) is working as 5bit/2.8-6.1MHz according to the papers. That does not match any standard format and I guess it is still supposed to apply some digital filtering (a Vivaldi DAC is not expected to be paired only with the upsampler). So there is always some kind of filtering/oversampling at DAC level.
However, that does not mean it does not benefit from something else “preparing the stream” so any of its “mandatory duties” can be done in the best conditions (specific input format for example). From personal experience, DSP is always a matter of compromises. I would be surprised that it does not apply also to dCS devices…
Sorry, I think that possibly my replies to @Meadow are perhaps getting a bit mixed with your question here - that query was regarding using the upsampling that is already in a Vivaldi Transport and an Upsampler and DAC.
I’d hoped that the link I gave you to the details on how the Ring DAC works would have given you all the information as to what happens with the audio data going into the DAC but yes, even if you send DXD at 24bit/384kHz then it HAS to be further upsampled within the Vivaldi DAC.
If I can quote directly from James’ material…
“PCM data which arrives at the Ring DAC is first oversampled to 706.8kHz or 768kHz. This is then modulated to 5-bits at a rate between 2.822MHz and 6.144MHz (depending on the unit, settings and content sample rate) and fed into the Mapper which distributes this signal to the current sources in the DAC.”
I think it is probably myself who may have misunderstood/misinterpreted ur response to my qs on Clone Mode on/off.
My revised understanding now is -
Scenario 1 - Any PCM data that is below 24bit/384kHz wud be upsampled first to DXD in Upsampler with Clone mode Off. That upsampled data wud be further oversampled to 706.8kHz or 768kHz in the Dac.
Scenario 2 - Any PCM data that is below 24bit/384kHz wud NOT be upsampled in Upsampler with Clone mode On. But it wud be upsampled to required rate and then further oversampled to 706.8kHz or 768kHz - both in the Dac (or will it be directly oversampled to .
Scenario 3 - Any PCM data that is 24bit/384kHz wud NOT be upsampled in Upsampler irrespective of whether Clone mode On or Off. But it wud be further oversampled to 706.8kHz or 768kHz in the Dac.
This has been a very interesting thread to read, because it’s very similar to something I am trying to do:
What I would like to do is be able to set all PCM to automatically upsample to the correct DXD rate (easy), and all DSD to upsample or pass straight through to DSD 128 (easy), and to automatically switch back and forth depending on the source material (not so easy). Perhaps I am missing something obvious, but it seems impossible to have these settings and have the Upsampler remember them:
If PCM, upsample to the correct DXD multiple;
If DSD, pass through the 128 and upsample the 64 to 128; and
If the source material changes, automatically switch to the correct setting.
In fairness, it seems the Upsampler will do everything I want except the automatic 2xDSD upsampling/switching. If I am content to have DSD pass through at either 64 or 128, whichever the source is, then the Upsampler will correctly switch back and forth between Auto DXD (Lock) and DSD without skipping a beat. It if I want the output of my DSD to be be 2xDSD, then everything including PCM becomes 2xDSD, unless I manually switch it back. This is similar to @Meadow ’s request to have DXD “untouched,” and so I included this request in this thread.
SIDENOTE: I could have sworn that, in a previous version of Mosaic, when I set Output to DXD (on the Upsampler front panel) and enabled DXD Lock, in the Mosaic/Audio/Output screen, the sampling rates no longer included 352800 and 384000, but were replaced by a single box containing DXD. Whether that was the case before, it’s definitely not the case now as this screen shot shows.
To me, it would make so much more sense to have the numerical boxes disappear when DXD Lock is activated and replaced with a box that says, well, DXD Lock or even DXD Auto.
Well at least you can come close to what you want. With DSD pass through and output set to DXD, as far as I know you will get:
PCM, upsample to the correct DXD multiple;
If DSD, pass through (64 and 128) - this is the bit you will miss: I don’t think you can get the upsampling for DSD64 to 128 with just the “pass-through” switch
If the source material changes, automatically switch to the correct setting.
As in my case, I would like just DXD to be untouched, from “full upsampling to DSD128” setup, it is not the same story
Actually, Meadow, as I consider it, I think we are talking about the same thing. DXD is just max PCM. If I have a source that is DXD, the Upsampler passes it through without change.
If I have DSD set to pass through, DXD will still be passed through unchanged.
With Output set to DXD and DXD set to DXD Lock, all PCM will be converted to the correct max PCM.
However, if I set Output to 2xDSD, then all sources PCM and DSD—get Upsampled to DSD128. That’s not what I want either. Like you, I just want the Upsampler to leave DXD alone. It will do that as long as we don’t upsample 2xDSD. Am I missing something?
For now, I will leave with DXD Lock and DSD pass through. It doesn’t get me everything I want, but it gets most of it.
[Note that in the Ring DAC, yes, everything gets converted. But that’s the way it works. That’s not part of this discussion).
For sure things are close.
But eventually the outcome will be the same in your set-up. But in mine no as I don’t mind converting 24/192 to DSD128. It is really for DXD that I really don’t get the point (as far as know, most of DSD music available has been at some point converted to DXD for mixing/mastering).
Really the best would be that the upsampler remembers the output format per input format, exactly as it remembers the filter. But I guess it is even more daydreaming