Audio-gd NFB-10

Aug 17, 2011 at 5:26 PM Post #1,037 of 2,860
I understood where your price came from but you also added cost to show the difference (impacts) to you of going balanced.   I've never heard a k701 balanced but I've read that it responds well to being balanced.  I do not recall if they respond to being recabled.  BUT, the real point was that they are unusually easy to balance since they already have a 4 conductor cable.  The 702 which is easier to switch cables has a 3-conductor cable so you can't just re-terminate those.  Personally, I'd wait on the fancy re-cable. 
 
Also, and I realize you probably don't care, but any re-cable I do from here on out will probably all terminate to a 4-pin xlr and then I'll use an adapter for whatever connection I need.  That's how I have my q-cable setup on my LCD-2, even though I rarely listen to it balanced, and I really like it.
 
Aug 17, 2011 at 5:58 PM Post #1,039 of 2,860


Quote:
Originally Posted by tme110 /img/forum/go_quote.gif
 
Do you really care about digital filters?   I haven't used them but I kind of just took them as toys added to the cheaper models. I could easy be wrong though.
 


That's what I thought at first too. But I think the digital filters offered in the NFB-2/3 (and maybe NFB-5) are different from modifying the output stage circuitry, such as adding OPAmps or Kingwa's "Diamond" output stage.
 
It looks like the digital filter options in Audio-gd's WM8741 DACs are actually done on the DAC chip itself (think this was also mentioned in the NFB-2/3 thread). See this link where it says:
 
"The WM8741 includes...  ...a range of advanced digital filter responses...  ...The digital filters include several selectable roll-off and performance characteristics. The user can select between standard sharp or slow roll-off responses. In addition, the WM8741 includes a selection of advanced digital filter characteristics including non-half band filters and minimum phase filters. This flexibility provides a range of benefits, such as significantly reduced pre-ringing and minimal group delay. The internal digital filters can also be by-passed and the WM8741 used with an external digital filter..."
 
So, in the NFB-1/10, etc, either Kingwa is selecting one of these filter settings as default or he is using an external digital filter. But I don't see any mention of an external digital filter on the product pages... So I think he might be simply selecting which ever one he thinks sounds best as a default (not a bad thing as he apparently makes good sounding DACs
tongue.gif
). If so, it would be interesting to know which one.
 
Even if most are gimmicky, it would be nice to be able to get a fuller functionality out of the DAC chips. And I think some have reported positively about the "[size=x-small]8X oversampling , Minimum phase apodising filter. [/size] [size=x-small] Up to 48KHz[/size][size=x-small] support[/size][size=x-small]" filter [/size]with the NFB-2/3
 
 
 
 
Aug 17, 2011 at 9:20 PM Post #1,041 of 2,860


Quote:
 
It looks like the digital filter options in Audio-gd's WM8741 DACs are actually done on the DAC chip itself (think this was also mentioned in the NFB-2/3 thread). See this link where it says:
 
"The WM8741 includes...  ...a range of advanced digital filter responses...  ...The digital filters include several selectable roll-off and performance characteristics. The user can select between standard sharp or slow roll-off responses. In addition, the WM8741 includes a selection of advanced digital filter characteristics including non-half band filters and minimum phase filters. This flexibility provides a range of benefits, such as significantly reduced pre-ringing and minimal group delay. The internal digital filters can also be by-passed and the WM8741 used with an external digital filter..."
 
So, in the NFB-1/10, etc, either Kingwa is selecting one of these filter settings as default or he is using an external digital filter. But I don't see any mention of an external digital filter on the product pages... So I think he might be simply selecting which ever one he thinks sounds best as a default (not a bad thing as he apparently makes good sounding DACs
tongue.gif
). If so, it would be interesting to know which one.


Uh, yes he does mention what digital filter is used... which is the WM8805 if you want 24/192, or the DIR9001 if you want 24/96 (and he says you get to pick the filter you want). The internal digital filter is only used on the ES version because ES9018 has a 50ps or better filter already.
[size=x-small]"Note: [/size][size=x-small]The S/PDIF interface default applied WM8805 (192KHz support), if you want replace the  DIR9001 (96KHz support) by free ,tell us  while place the order please."[/size]
 
Quote:
Would that "external digital filter" be the Digital Interface perhaps?


DI is not a digital filter. Chips like WM8805, DIR9001, CS8416 are.
 
Quote:
Anyone who thinks the -10 isn't a good SE amp hasn't listened to it.  In fact, and even though I probably have the newest -10 anywhere - I think I'm going to sell mine because I want a ROC instead (becuase I love the amp).  But I'm only interested in this because I already have a NFB-1 and I bought the -10 to use with the -1 but can't.


Agreed. The amp and dac sections are both fantastic. I can even drive my HE-6 out of the single-ended and it still sounds great. Though you might want to sell NFB-1 and stick with NFB-10... not sure that the ROC is that much better to make the price worth it.
 
Aug 17, 2011 at 10:08 PM Post #1,042 of 2,860
Uh, yes he does mention what digital filter is used... which is the WM8805 if you want 24/192, or the DIR9001 if you want 24/96 (and he says you get to pick the filter you want). The internal digital filter is only used on the ES version because ES9018 has a 50ps or better filter already.
[size=x-small]"Note: [/size][size=x-small]The S/PDIF interface default applied WM8805 (192KHz support), if you want replace the  DIR9001 (96KHz support) by free ,tell us  while place the order please."[/size]


Yes. The WM8805 and DIR9001 are digital input filters for the spdif implementation. And perhaps that is what is being referred to by "external digital filter". They are completely separate from the internal DAC digital filters, which is what Audio-gd is referring to for the NFB-2/3/5 when they talk about "[size=x-small]The NFB-5/2/3 have 9 types  digital filter characteristic for users find the best sound of personal sense"[/size]
 
The frequency response graphs illustrating these on the NFB-2/3/5 product pages are copied and pasted from the WM8741 documentation (which you can access by clicking on the PDF file under "DATASHEETS" on the page I linked to in my previous post).
 
My interest is that if Kingwa decided to implement these digital filters in the NFB-2 (their higher end single-ended DAC), then why not for the NFB-10WM? Seems like a reasonable query to me?
 
 
 
Aug 17, 2011 at 10:16 PM Post #1,043 of 2,860
Okay I see what you're talking about now, I think calling it an oversampling filter is much more clear since digital filter could mean a lot of things. I don't believe you can change the oversampling in the NFB-10 since I don't see any jumpers for it like they show for the NFB-2. You could email Kingwa and ask him which setting he chose to use.
 
Aug 17, 2011 at 11:27 PM Post #1,044 of 2,860


Quote:
DI is not a digital filter. Chips like WM8805, DIR9001, CS8416 are.
 


Doesn't it essentially do the same thing? Please enlighten us. 
 
edit: ahh..nevermind. We're talking about the whole upsampling vs. oversampling deal. That kinda makes me wonder, how it would sound, if you upsampled then oversampled again..
 
Aug 18, 2011 at 3:31 PM Post #1,045 of 2,860
A little off topic, but
 
I sent Audio-gd an email with a couple of questions regarding the PSU for the DAC in the NFB-10 and his thoughts on the possibility of implementing the internal "filters" on WM8741 for the NFB-10 as he's done for the NFB-2/3. It's been 3 days and I've not heard from him (though I got the automated message indicating that they'd received my email. However, I sent him an email last week and I got a response within 12 hours.
 
Does Audio-gd normally respond within a day or is it variable?
 
Aug 18, 2011 at 4:27 PM Post #1,047 of 2,860
I'm just starting to read this thread, and right now I'm on page 1
smily_headphones1.gif
. Nevertheless, I would like to ask one question now. Right now I have a nbf-3 connected to a speaker amp through rca. The speaker amp is an integrated one, that can be set to power amp bypassing the volume control, this way the quality improves. So I'm thinking of getting a preamp too. I also have headphones, the audeze lcd-2, connected to a leeman cube amp.
Do you guys think I should seel the BCL headphone amp, the nfb-3, and buy the nfb-10? If so, do you honestly think a nfb-10 version with 3 transformers (in a case like the phoenix) wich COULD be released in the future would be audibly better? What about the ESS dac? Is it much colder than the wolfson? I find the nfb-3  too warm for my taste, but I'm afraid the ess would sound too harsh on my speakers with true ribbon tweeters.
 
Aug 18, 2011 at 4:48 PM Post #1,048 of 2,860


Quote:
Doesn't it essentially do the same thing? Please enlighten us. 
 
edit: ahh..nevermind. We're talking about the whole upsampling vs. oversampling deal. That kinda makes me wonder, how it would sound, if you upsampled then oversampled again..


Essentially, I think you are correct. Oversampling DACs (such as the WM7841) often (always?) apply a filter to try and minimize phasing or other artifacts that crop up during the digital-to-audio-conversion. I don't know much about this sort of thing, but apparently the WM7841 has a bunch of selectable  (3 in hardware mode - which I think is what Kingwa has implemented in the NFB-2/3/5). Kingwa describes them as "brick-wall", "soft-knee" or "apodising" filters, which are implemented at 2x, 4x or 8x sampling rates, giving the 9 total internal digital filter options. (Not sure if these are actual names of these filters, or just names that Kingwa uses to describe them more generically for layman comprehension, as these names do not appear in the WM7841 documentation).
 
Someone can correct me if I'm wrong, but I think what it essentially boils down to is altering how linear the frequency response is using various algorithms to mitigate artifacts caused by digital input (spdif) implementation, jitter or poorly-regulated DAC PSU (the latter shouldn't be an issue in Audio-gd DACs). I think that most of these sonic/ultrasonic anomalies impact the treble, or are at least most noticeable in the treble range and result in things like DAC ringing, etc. Therefore, the largest impact of these filters seems to be how linear or "rolled-off" the higher frequencies are.
 
"Brick-wall" would seem to imply a non or very slightly rolled off or altered linear response.  "Soft-knee" would be a more pronounced, yet more gradual alteration of the treble range - resulting in less "harsh" sounding artifacts should they crop up at the cost of altering the true signal somewhat, and again, I think "apodizing" is an attempt at a happy medium.
 
But again, I don't really know much about this stuff. I would think that oversampling DACs and their post-processing filters would be discussed in more detail somewhere else in these forums. 
 
The thing that seems more unique about the WM7841 is that it lets you choose between a variety of these filters. It would seem to me that in a more carefully implemented DAC, with low jitter input, etc, these filters would be less important. However, they might be more useful when using lower quality sources or transports?
 
The fact remains that Kingwa decided to implement selection of these filters in the NFB-2, and I'm curious as to whether they would also be beneficial in the NFB-10. However, I can see why he wouldn't bother in highest-end DACs as he would be expecting optimal digital "input" conditions, both due to the redundancy of his PSU designs and the expectation that people would be using quality sources.
 
Aug 18, 2011 at 4:59 PM Post #1,049 of 2,860


Quote:
variable, also what did u ask him about the psu, im interested.


OK, thanks. I'll give it a couple more days.
 
I just asked him if he could describe the reasons for implementing linear vs class A PSU for the DAC in the NFB-10, and if, in his opinion, a class A would really make all that much difference. But I think that has mostly been answered here. However, he has already told me that he feels the single-ended out of the NFB-10 vs NFB-2 would be "about the same", even though he has previously given the impression (to others) that the NFB-2 would, theoretically, be the better performing single-ended DAC... So I don't know if I'll get much more useful info from him in that regard. 
tongue.gif

 

Users who are viewing this thread

Back
Top