Are Audio-Gd DAC's all that? - Page 20

Quote:

Originally Posted by Currawong 

That's not far off the price of the PCM1704UK.  It's only a 20-bit DAC though. I'm not sure people would buy DACs with a lower spec (even if, realistically, it doesn't matter for various technical reasons) as people usually want new gear with better specs, not vintage gear, even if it (supposedly) sounds better.

You describe exactly how folks in the 70's went for solid state amps and threw their tube amps in the garbage.  Give it a couple decades,  people usually figure things out. 

Quote:

Originally Posted by Currawong 

That's not far off the price of the PCM1704UK.  It's only a 20-bit DAC though. I'm not sure people would buy DACs with a lower spec (even if, realistically, it doesn't matter for various technical reasons) as people usually want new gear with better specs, not vintage gear, even if it (supposedly) sounds better.

I agree, because there's quite a bit of 24/96 music out now.  Why limit yourself?

Quote:

Originally Posted by IPodPJ 

I agree, because there's quite a bit of 24/96 music out now.  Why limit yourself?

Because you aren't,  its just marketing numbers.  No different than buying an "energy bar" that has the same ingredients as a candy bar :)  

But back to the AudioGD's.   the Ref 8 really has my attention.   Anyone own one?   I'd have been first in line if AudioGD bought the HDCD code for it as he wanted but stupid M$ wanted a crimpling fee for the small business entrpenuer.  Anyway I was thinking of using HDCD.exe with foobar to the DSP1 Ref8,  that would do all of the HDCD decoding except on the fly filtering change.  Thinking with the modern filtering algorithms in the DSP-1 I may not miss that feature.
 

Quote:

Originally Posted by regal 

Because you aren't,  its just marketing numbers.  No different than buying an "energy bar" that has the same ingredients as a candy bar :)  

 

Let's assume that there is no difference in the music at all (even though I don't believe that to be the case, and Dan Lavry has even cited that 24/96 has benefits but beyond that is unnecessary).  There is a difference in the filtering that takes place in the DAC.  And to my ears I was able to prove this.

I have never found any on-the-fly software upsampling plug-ins to make a positive difference in my system, but when I actually upsample the file I do.  There is one program I purchased called Audio Transcoder.  I have taken some of my best quality albums and converted them to 24/96.  There is certainly more spatial depth and image localization when I do this.  I'm not sure whether this is due to my Transporter having more accurate clocking at 24/96, or if it's in the Ref7, or a combination of both but the difference is apparent.  It does seem like there is slightly less bass impact with the 24/96 versions though, but just barely.

It seems as if the Audio Transcoder software is actually filling in the missing pieces with the upsampling, but I'm not sure if that's really what's going on or if it is just how Audacity is rendering the waveform.  Here is a screen cap I took after bringing the files into Audacity.  Either way, the sound coming out of the DAC is different.

IpodJ,  my guess is you like the upsampling simply because the Transporter doesn't have a 44.1 mutliple clock.  Upsampling to 96k puts it at a frequency where a PLL derived timing isn't required.

But we need to get back on topic,  sorry to have derailed things a bit.

Regal: You may be right though. A mate of mine in CA who has the Transporter reckons the 24/96 versions of all his music all sound better in much the same way. 

Back on topic though, since it was more or less prompted by Yikes starting this thread, the Ref 7 loaner will be heading out on Monday, so we should start getting impressions by the end of the month.

Anyway I could get on the loaner list?  Would love to compare to my old school PMD100 3SE

Quote:

Originally Posted by regal 

Anyway I could get on the loaner list?  Would love to compare to my old school PMD100 3SE

I don't see why not...PM Curra and see what can be arranged (I'm assuming Curra is the one to contact).

Peete.

Quote:

Originally Posted by dmac03 

I think that the new NFB1 may be the new best value dac on the lot. Of course we'll have to see how the reviews stack up with other units. However, I base my presumption largely on the fact that the entirety of the circuitry is still utilized during SE output.

So the ref 5 does not? Would this make a real difference?

Quote:

Originally Posted by siess 

So the ref 5 does not? Would this make a real difference?

Sigma delta DAC chips are differential,  so a single chip has a balanced output.  Many DAC makers simply tap the + output for single ended operation,  this is bad practice the S/N ratio and bit depth of the chip cannot be actualized by doing this.  AudioGD correctly does a balanced to unbalanced transform with the S-D chips in order to get the CMRR and complete bit depth of the chip.

However R2R DAC chips as used on the Ref5 aren't differential,  they require 2 chips per channel for balanced and one chip per channel for unbalanced.  It is not bad practice in this case to use only half the DAC chips in unbalanced output because it is how the chips are designed.  It does appear to me (speculation) that on the Ref 8 he is doing a balanced to unbalanced transform with the R2R chips in order to get what would be unheard of CMRR for an R2R DAC,  this would be "going the extra mile"  something I haven't seen anyone do in my personal observation of DAC evolution over the past decade.

Simple summary of a balanced DAC outputing unbalanced:

1. S-D use one chip per channel with both + and - outputs combined and transformed to unbalanced is expected,  taping a single + is bad pactice

2. R2R using half the DAC's in unbalanced is completely acceptable

3. R2R using all the chips in unbalanced by doing a balanced to unbalance transform - extremely cutting edge (charting new territory)

4. R2R using multiple chips in single ended (all the same phase) has been done several times and gives better accuracy (slightly better THD but debatable)

Thanks for that answer regal

Quote:

Originally Posted by regal 

Sigma delta DAC chips are differential,  so a single chip has a balanced output.  Many DAC makers simply tap the + output for single ended operation,  this is bad practice the S/N ratio and bit depth of the chip cannot be actualized by doing this.  AudioGD correctly does a balanced to unbalanced transform with the S-D chips in order to get the CMRR and complete bit depth of the chip.

 

However R2R DAC chips as used on the Ref5 aren't differential,  they require 2 chips per channel for balanced and one chip per channel for unbalanced.  It is not bad practice in this case to use only half the DAC chips in unbalanced output because it is how the chips are designed.  It does appear to me (speculation) that on the Ref 8 he is doing a balanced to unbalanced transform with the R2R chips in order to get what would be unheard of CMRR for an R2R DAC,  this would be "going the extra mile"  something I haven't seen anyone do in my personal observation of DAC evolution over the past decade.

 

Simple summary of a balanced DAC outputing unbalanced:

1. S-D use one chip per channel with both + and - outputs combined and transformed to unbalanced is expected,  taping a single + is bad pactice

2. R2R using half the DAC's in unbalanced is completely acceptable

3. R2R using all the chips in unbalanced by doing a balanced to unbalance transform - extremely cutting edge (charting new territory)

4. R2R using multiple chips in single ended (all the same phase) has been done several times and gives better accuracy (slightly better THD but debatable)

 

 

Sorry for my lacking understanding of English but you said: 

Sigma delta DAC chips are differential,  so a single chip has a balanced output.

and

However R2R DAC chips as used on the Ref5 aren't differential,  they require 2 chips per channel for balanced and one chip per channel for unbalanced.

.......................

I don't quite understand the word "channel". I thought a channel is + & - so 2 channels per output and if this is true then 2x chips per channel is actually 4 chips per output.

Just need a clarification, thanks.

kindest regards, Peter

Quote:

Originally Posted by .Sup 

Sorry for my lacking understanding of English but you said: 

 

Sigma delta DAC chips are differential,  so a single chip has a balanced output.

 

and

 

However R2R DAC chips as used on the Ref5 aren't differential,  they require 2 chips per channel for balanced and one chip per channel for unbalanced.

.......................

I don't quite understand the word "channel". I thought a channel is + & - so 2 channels per output and if this is true then 2x chips per channel is actually 4 chips per output.

Just need a clarification, thanks.

 

kindest regards, Peter

No, channel is Left and Right in stereo.  The + and - are the phases of a single channel.  With unbalanced output only one R2R chip is needed as negative is just ground.
 

ok thanks for clarifying for me regal 

Quote:

Originally Posted by .Sup 

Sorry for my lacking understanding of English but you said: 

 

Sigma delta DAC chips are differential,  so a single chip has a balanced output.

 

and

 

However R2R DAC chips as used on the Ref5 aren't differential,  they require 2 chips per channel for balanced and one chip per channel for unbalanced.

.......................

I don't quite understand the word "channel". I thought a channel is + & - so 2 channels per output and if this is true then 2x chips per channel is actually 4 chips per output.

Just need a clarification, thanks.

 

kindest regards, Peter

Just for the record, 2 chips per channel would not be required for balanced analog output. From what I have read, you gain about 3dB of SNR, provided the chips are matched.