Mar 1, 2017 at 5:55 PM Post #7,787 of 27,067
And I finally discovered from where the hot ordor came...Not the dave, it was an insect on a bulb... lol.
 
Anyway, I suppose this is something to do with the usb connection...It seems to be really sensitive, i just noticed i plugged into a different computer usb out... Freaking strange.
 
Anyway, you are right for Pet Sounds. I saw it played in live by the Beach Boys , and I have to say Dave does an amazing job. Even I prefer the stereo mix, it is even better.
 
Mar 1, 2017 at 6:02 PM Post #7,788 of 27,067
 
What is the truth then? It's been said before that much of the DAVE is now dormant with the Blu2 in the chain. I, for one, don't want to believe this, but what does Chord say?

(Note my wishful thinking.)

 
I have tried to find the post explaining it. It is 1 million taps in total, but I am sure that this is split between the Blu2 and DAVE. Blu2 does substitute for the initial upscaling stages in DAVE, leaving these redundant, but the final upscaling is performed by the DAVE. However this seems to contradict what @highfell heard
 
Here is that post which I kept visualising. http://www.head-fi.org/t/766517/chord-electronics-dave/7455#post_13234615
So it seems like the Blu2 does 1M taps of upscaling, then feeds the signal into the DAVE after the initial 16 ufs stage of upscaling, and the rest of the DAVE upscaling is still used. So a portion of DAVE will be dormant, but I think it relates to only a small portion of DAVE, rather than the majority.
 
That also helps explain why the Blu2 + hugo 2 will achieve less overall taps than the Blu2 + DAVE.

 
I think that's based on a misunderstanding. The WTA filtering is fully done on the M-Scaler, so the DAVE's WTA filter is inactive – thus has the same million taps as with the Hugo². However, from what I get and recall, there's some further upsampling done in the DAVE (and in the Hugo[2] as well) for noise-shaping and whatever purposes (remember the DAVE's superb modulation-noise suppression!).
 
Mar 1, 2017 at 6:06 PM Post #7,789 of 27,067
  Maybe there's no DAC in the world that can separate the instruments from the muddy mono mix of Pet Sounds? (God bless that beautiful album.)

 
  Anyway, you are right for Pet Sounds. I saw it played in live by the Beach Boys , and I have to say Dave does an amazing job. Even I prefer the stereo mix, it is even better.

 
Interested in a stereo* version of the remaining mono tracks? PM me!
smile.gif

 
* read: stereophonized
 
Mar 1, 2017 at 6:06 PM Post #7,790 of 27,067
   
I think that's based on a misunderstanding. The WTA filtering is fully done on the M-Scaler, so the DAVE's WTA filter is inactive – thus has the same million taps as with the Hugo². However, from what I get and recall, there's some further upsampling done in the DAVE (and in the Hugo[2] as well) for noise-shaping and whatever purposes (remember the DAVE's superb modulation-noise suppression!).

 
Damn! Just when i thought I was starting to understand things, you have bowled a googly.
 
Mar 1, 2017 at 6:16 PM Post #7,791 of 27,067
 
  I think that's based on a misunderstanding. The WTA filtering is fully done on the M-Scaler, so the DAVE's WTA filter is inactive – thus has the same million taps as with the Hugo². However, from what I get and recall, there's some further upsampling done in the DAVE (and in the Hugo[2] as well) for noise-shaping and whatever purposes (remember the DAVE's superb modulation-noise suppression!).

 
Damn! Just when i thought I was starting to understand things, you have bowled a googly.

 
Yeah, sorry, Mike! But better late than never. I've had a hard time understanding the stuff as well, it's very complex. Just not as complex as I intermediately thought; my initial idea was much closer: The number of «taps» indicates the sharpness of the low-pass filter – called WTA filter in Rob's sophisticated implementation –, performed by the upsampling algorithm. Other than commonly interpreted the main function of upsampling is always low-pass filtering.
 
Mar 1, 2017 at 6:17 PM Post #7,792 of 27,067
Apparently I have to link this block diagram on every page of this thread:



The 16FS primary WTA stage uses most of the FPGA (well, most of the DSP cores, there's programmable cells that are separate). The second WTA stage, to 256FS, is tiny by comparison. The final linear interpolation stage to 2048FS can't be large (not big enough to be given a size on the diagram). But the Noise Shaper that comes afterwards is so big that it couldn't fit, all on its own, into Hugo's FPGA.

The 16FS primary WTA stage is completely redundant when the M-Scaler feeds 16FS music into DAVE.

Now playing: Bettie Serveert - Balentine
 
Mar 1, 2017 at 6:25 PM Post #7,793 of 27,067
   
I think that's based on a misunderstanding. The WTA filtering is fully done on the M-Scaler, so the DAVE's WTA filter is inactive – thus has the same million taps as with the Hugo². However, from what I get and recall, there's some further upsampling done in the DAVE (and in the Hugo[2] as well) for noise-shaping and whatever purposes (remember the DAVE's superb modulation-noise suppression!).

 
This is why I'm hopeful that Hugo2 + Mscaler will be roughly equivalent to Dave + Mscaler.  My pro forma "What Would Watts Do" acquisition roadmap is
 
Mojo for mobile/office use (done)
Hugo 2 for my 2 channel system (preordered)
Add an Mscaler for my 2 channel system (take my money, please!)
Add digital amps to mscaler in my 2 channel system
"promote" my Hugo 2 to office use, with Mojo for mobile use.
 
Mar 1, 2017 at 6:26 PM Post #7,794 of 27,067
I have tried to find the post explaining it. It is 1 million taps in total, but I am sure that this is split between the Blu2 and DAVE. Blu2 does substitute for the initial upscaling stages in DAVE, leaving these redundant, but the final upscaling is performed by the DAVE. However this seems to contradict what @highfell
 heard


Here is that post which I kept visualising. http://www.head-fi.org/t/766517/chord-electronics-dave/7455#post_13234615
So it seems like the Blu2 does 1M taps of upscaling, then feeds the signal into the DAVE after the initial 16 ufs stage of upscaling, and the rest of the DAVE upscaling is still used. So a portion of DAVE will be dormant, but I think it relates to only a small portion of DAVE, rather than the majority.

That also helps explain why the Blu2 + hugo 2 will achieve less overall taps than the Blu2 + DAVE.


That wasn't quite the post I was thinking of, but almost. At any rate, hmmmm... But dual data mode with the Hugo2 DOES give you 1m taps. But you're not getting 2048x upsampling with the Hugo2. DAVE is also a 20 element pulse array design compared to the Hugo2 at 10e. From my understanding, this makes up a large part of DAVE's small signal integrity and lack of distortion.

All of this (and more, I'm sure) makes the DAVE far from obsolete. Indeed, the DAVE is needed to take full advantage of Blu2's 1mil taps.

So, really, if we think there is a possibility to go well below DAVE's 0.000015% THD, and zero noise floor modulation, and to be able to actually HEAR a difference, I would think that the DAVE will remain the go-to DAC, where all one could hope to do is increase the taps to 250 million to fully render 24-bit audio.

I'm thinking that's not the whole story, though, since it's logical that there would be noise problems inherent in reaching such a lofty peak.
 
Mar 1, 2017 at 6:28 PM Post #7,795 of 27,067
Two things that are immediately noticeable about escalated upsampling even from my Red Ref III is 1) Tighter bass and 2) Greater accuracy of instrument timbre, though I can't say I understand why.

I am still absolutely smitten by the Dave Dac after 15 months. It's quite hard to imagine the level of performance Blu2 M Scaler can brink to the party but I am looking forward to hearing it.
 
Mar 1, 2017 at 6:41 PM Post #7,796 of 27,067
 
   
I think that's based on a misunderstanding. The WTA filtering is fully done on the M-Scaler, so the DAVE's WTA filter is inactive – thus has the same million taps as with the Hugo². However, from what I get and recall, there's some further upsampling done in the DAVE (and in the Hugo[2] as well) for noise-shaping and whatever purposes (remember the DAVE's superb modulation-noise suppression!).

 
This is why I'm hopeful that Hugo2 + Mscaler will be roughly equivalent to Dave + Mscaler.  My pro forma "What Would Watts Do" acquisition roadmap is
 
Mojo for mobile/office use (done)
Hugo 2 for my 2 channel system (preordered)
Add an Mscaler for my 2 channel system (take my money, please!)
Add digital amps to mscaler in my 2 channel system
"promote" my Hugo 2 to office use, with Mojo for mobile use.

 
Are you sure the M-Scaler can drive the digital amp? I don't recall such a statement from Rob (don't exclude it, though).
 
Mar 1, 2017 at 6:47 PM Post #7,797 of 27,067
[...]But dual data mode with the Hugo2 DOES give you 1m taps. But you're not getting 2048x upsampling with the Hugo2.

All of Chord's current DACs upsample to 2048FS. They all have a noise shaper and pulse array running at the same speed.

A key difference between them is the initial WTA filter stage (which is the most important, it determines the overall performance level). That's why it's apparently so obvious when M Scaler is connected to DAVE.

DAVE is also a 20 element pulse array design compared to the Hugo2 at 10e. From my understanding, this makes up a large part of DAVE's small signal integrity and lack of distortion.


So both of these stages in DAVE (noise shaper and pulse array) are far ahead of any other DAC. But Hugo 2, with 10 elements is a big improvement over the 4 elements seen in Hugo (TT and 2Qute) and Mojo.

All of this (and more, I'm sure) makes the DAVE far from obsolete. Indeed, the DAVE is needed to take full advantage of Blu2's 1mil taps.

So, really, if we think there is a possibility to go well below DAVE's 0.000015% THD, and zero noise floor modulation, and to be able to actually HEAR a difference, I would think that the DAVE will remain the go-to DAC, where all one could hope to do is increase the taps to 250 million to fully render 24-bit audio.

Arguably there should be a single stage WTA filter that goes directly to 2048FS.

I'm not convinced that noise floor modulation in DAVE is zero. It's very clear in TT once you compare it with DAVE and I think I hear a residual of the same with DAVE. But the only proof is when you hear a better DAC...

Now playing: This Mortal Coil - Dreams Made Flesh
 
Mar 1, 2017 at 6:54 PM Post #7,798 of 27,067
Are you sure the M-Scaler can drive the digital amp? I don't recall such a statement from Rob (don't exclude it, though).

The Power DAC (I think that's a better name) has a volume control. So when driven by Blu 2 via the DX connection you will have no need for DAVE. All the functionality in DAVE after the first stage 16FS WTA filter is functionality that the Power DAC will need to implement itself. See the diagram I posted earlier. Notice the block called DX Digital Output.

There is a complication with DSD though. I wonder if Rob is going to ignore it completely in Blu 2.

Now playing: Pink Floyd - Vera
 
Mar 1, 2017 at 7:23 PM Post #7,800 of 27,067
The Power DAC (I think that's a better name) has a volume control. So when driven by Blu 2 via the DX connection you will have no need for DAVE. All the functionality in DAVE after the first stage 16FS WTA filter is functionality that the Power DAC will need to implement itself. See the diagram I posted earlier. Notice the block called DX Digital Output.

There is a complication with DSD though. I wonder if Rob is going to ignore it completely in Blu 2.

Now playing: Pink Floyd - Vera


What are you talking about?

I have no idea what he's talking about.

Did you just invent something for Chord?
 

Users who are viewing this thread

Back
Top