FiiO X7 | DXD | DSD | 384K/64B | ESS9018+ Android | WiFi | Bluetooth | 4 AMP modules | Balanced Out |
Aug 28, 2014 at 12:01 PM Post #722 of 18,020
or maybe someone can let us know which DAP supports native DSD?
It seems QLS QA360 does. It has analog volume control (TI PGA2311) implemented using digitally controlled resistor network. IMHO this is the way to go.
 
Aug 28, 2014 at 12:09 PM Post #723 of 18,020
  or maybe someone can let us know which DAP supports native DSD?

 
None, in the strictest definition.
 
But the F880/ZX1 comes closest, because it takes in DSD and does PWM with it (basically the same thing) into an analog low-pass, so very little conversion is necessary and is as close to a true 1-bit DAC as it can get without gutting a 1990s era Sony SACD player. This is basically the same principle as the PS Audio DirectStream DAC. Unfortunately, doing this this way suffers from poor SNR (at most about 16-bit ENOB with today's MCUs and switching capacitors) and low-level resolution for PCM conversion (subjective). The thing about the F880/ZX1 is that I think its MCU can't do PWM at fast enough rates and retain voltage precision, so the DSD data is almost certainly also decimated to a lower data rate.
 
Anything that uses multi-bit delta-sigma DACs with "native DSD support" (CS4398, AD1955, ES9018S/K2M, WM841) will perform some kind of internal decimation of the DSD sample rate to whatever the chip itself works at (e.g. >1.5 MHz for ESS, which is partly why the ES9018 likes >80 MHz MCLK, as the DPLL settings are tight). The WM8741 has a "dedicated DSD mode", which requires an analog low pass and analog volume control, but it is locked into that mode, so very few DACs actually use it, and use WM8741's DoP mode, which decimates DSD into the 24/352.8 form of PCM.
 
The "least native" is anything that uses WM8740, because it requires the SoC to accept the DoP headers and decimate the DSD data into something the WM8740 can read, which doesn't support DSD. This hand-off introduces extra jitter at least by a little bit.
 
Aug 28, 2014 at 12:15 PM Post #724 of 18,020
By the way... Choosing between DX90 and X5, not a least for me was support for USB HDD via OTG. Lack of such support in X5 was no go for me. DX90 works quite well in such a setup, with external power bank or wall adapter. I took my whole library on 2TB WD My Passport Ultra to vacations trip, without having to make hard decision or waste money on several flash cards, and I was happy with it!

If X7 would exceed it by providing enough current via USB OTG (keeping external power supply as an option), for 3-4 hours, that would be great!
 
Aug 28, 2014 at 3:42 PM Post #725 of 18,020
  It seems QLS QA360 does. It has analog volume control (TI PGA2311) implemented using digitally controlled resistor network. IMHO this is the way to go.


For the QA360 to do "direct DSD", it'd have to be run in the mode where the DSD data is passed directly to its switched capacitor filter array, and the subsequent output would have to be analog low-pass filtered for ultrasonic junk. Even though it's possible, doing so would be pretty expensive and complicated from both a software design and circuit design standpoint because the SoC would have to be coded to switch between two different modes of PCM (running FIR interpolation) and "Direct DSD", which completely disables all of that functionality and runs through completely different analog output, forgoing I/V conversion and instead going through an analog low pass. This kind of design requires coordination from firmware writing and hardware design that is not exactly trivial.
 
If it's not done that way, then the QA360 implements DSD just like Astell&Kern with a digital filter and available volume control with only one single analog stage. And QLS just throws in the PGA2311 for fun because some people think analog volume control is better (which is true in certain cases, but not for most).
 
Aug 28, 2014 at 11:00 PM Post #726 of 18,020
Given that the Oppo HA-1 does DSD natively (supposedly), and uses the same chip as planned for the X7, and James has/had an affiliation with Oppo in the past, could you not look in their neck of the woods to see what was done? I'm not sure if the implementation would work given that FiiO's design is meant to be considerably smaller (DAP vs full size DAC/amp - Obviously), but it might be worth looking in to.

Although I'm no electrical engineer I find it interesting to read about the possible challenges, like the ones brought up in the above post.
 
Aug 28, 2014 at 11:42 PM Post #727 of 18,020
Given that the Oppo HA-1 does DSD natively (supposedly), and uses the same chip as planned for the X7, and James has/had an affiliation with Oppo in the past, could you not look in their neck of the woods to see what was done? I'm not sure if the implementation would work given that FiiO's design is meant to be considerably smaller (DAP vs full size DAC/amp - Obviously), but it might be worth looking in to.

Although I'm no electrical engineer I find it interesting to read about the possible challenges, like the ones brought up in the above post.

 
Desktop gear are different story as they often don't use digital volume control because they have plenty of space to implement a good analog volume control which avoids all the issue with digital volume control.
 
Aug 29, 2014 at 1:25 AM Post #728 of 18,020
Desktop gear are different story as they often don't use digital volume control because they have plenty of space to implement a good analog volume control which avoids all the issue with digital volume control.


Yeah, the amp in the HA-1 is full analogue class-A so I wouldn't expect it to be the same. However, isn't there a smaller volume attenuator, or something, to control the output (line level?) from the DAC in a DAP? Again, just spitballing.
 
Aug 29, 2014 at 2:46 AM Post #729 of 18,020
I think it is okay X7 just do whatever AK240 is doing with DSD. After all, this will only cost 700usd not 2500usd. thinking that AK240 is  really overpriced by the day...
 
Aug 29, 2014 at 2:57 AM Post #730 of 18,020
Yeah, the amp in the HA-1 is full analogue class-A so I wouldn't expect it to be the same. However, isn't there a smaller volume attenuator, or something, to control the output (line level?) from the DAC in a DAP? Again, just spitballing.

 
There are a few way to implement volume control, the most common way is to put an attenuator / pot between the DAC's line-out and the amp stage. But smaller pot tends to have channel imbalance issue on low volume, and those good pot tend to be really expensive or just big in size (such as the one in HM901). Therefore they are not the ideal solution on DAP, especially if you want to keep the size reasonably small. Therefore the compromised way is to use the digital volume control built-in most DAC chip. There are a few ways how these kind of control is built onto different DAC as well - sometime it is fully digital, sometime it can be an analog resistor array right at the DAC's output, which is more closely resemble to an pot. So you really have to look at each DAC to tell what kind of volume control it is. There is no one universal solution for all of them.
 
Aug 29, 2014 at 3:15 AM Post #731 of 18,020
there comes with a volume control after ES9018 in X7. but is it means that X7 is 100% native supports DSD, who know ? the question is , who have the right decided what is native DSD ? depended on my knowledge I never find there are any answer about this. 
 
FiiO Stay updated on FiiO at their sponsor profile on Head-Fi.
 
https://www.facebook.com/FiiOAUDIO https://twitter.com/FiiO_official https://www.instagram.com/fiioofficial/ https://www.fiio.com support@fiio.com
Aug 29, 2014 at 3:22 AM Post #732 of 18,020
  there comes with a volume control after ES9018 in X7. but is it means that X7 is 100% native supports DSD, who know ? the question is , who have the right decided what is native DSD ? depended on my knowledge I never find there are any answer about this. 


 james , can we have a better battery life , like 20hrs may be ? 10 is not long enough .
 

Users who are viewing this thread

Back
Top