CONFUSED: USB -- a "solved" problem, or back to SPDIF?

[wbh]

I know that USB was -- say, prior to 2007 -- a compromise compared to SPDIF.

Enter Gordon Rankin and his "Streamlength"  technology, adopted by Wavelength Audio and later (licensed by??) Ayre, AudioStream and others. I think BenchMark and Bel Canto and others had their own "in-house" USB solutions (asynch??), but I may be wrong.

I suspected  SPDIF because of objections from way back (late 80s) -- well before USB -- when I2S was being pushed by manufs. like Perpetual Technologies,  Audio Alchemy, et  .al. 

SPDIF bad

I2S good

And that has been the consensus, on forums like diyaudio and diyhifi, for over 15 years.

 

Okay ...settled ....but maybe not! ....becasue a few years ago, Schiit's Mike Moffat was openly stating that SPDIF was the better way to connect DACs because USB is inferior. Not sure whether he was referring to Schiit products, lower-end Schiit products, or all DACs (in general).

Indeed, on certain forums (notably SBAF), members were CLAIMING the same thing .. spdif sounds better.

And, a few years ago .... poof ... we have the solution...

Schiit's Unison USB.

And there are others, of course. Allo's usbridge, etc.

 

So what's the story ...  many of us don't have hours and hours to kill on A/B listening tests. And, of course, there are many variables ... so let's make the issue testable and clear: say you own a Schiit Modi Multi (which does not incorporate Unison)  and connect it via USB or SPDIF, which will sound better -- all else held equal. 

 

 

[JoeWhip]

If you have Unison, that is the way to go. Sounds better than spdif, at least in my set up. If not, then spdif makes sense, although with my Yggy, the gen 5 usb sounded very close to spdif, but not as good as the AES implementation. Unison even bests that now.

[misterspense]

The hard thing in comparing USB and SPDIF is not on the dac side, but the source, there are not many sources that have both connection types.

So stating that Spdif on your dac sounds better (or worse) than Usb, does not tell us that much, because you can then also say, from the same listening comparison, that your cd player sounds better (or worse) than your "fill in your computer or pi, with or without usb decrapifiers".

[JoeWhip]

My comparisons were usb from a laptop vs. spdif from a transport. With Unison, the laptop is the real winner. No contest really. Hopefully, I will get my hands on the Schiit transport by year’s end Which will have spdif and usb outputs.

[wbh]

5 hours ago, misterspense said:

The hard thing in comparing USB and SPDIF is not on the dac side, but the source, there are not many sources that have both connection types.

So stating that Spdif on your dac sounds better (or worse) than Usb, does not tell us that much, because you can then also say, from the same listening comparison, that your cd player sounds better (or worse) than your "fill in your computer or pi, with or without usb decrapifiers".

Yeah .. the folks claiming SPDIF superiority are not clear on transport sources (optical disc player, computer, etc). USB to SPDIF converters can confuse the issue in that (good ones) buffer and re-clock and "clean up" the digital signal just before it enters the dac.

But, frankly, I'm never going back to optical discs (players, xports) as a playback medium. 

But I don't steam either. Everything is played in either a dedicated DAP (for portable; I don't play off phone!) or a second ssd  (native installed, just for music files) on a PC/laptop. 

So many variables!

My Xonar sound card -- on main PC -- does spit out SPDIF, so I'll have to experiment with a Windows or Linux PC as a source.

 

From disc-spinning days, I do recall SPDIF was very good ... for "good" xports and dacs (Say a mid-level Sony CDP, spdif rca coax out,  with Musical Fidelity DAC). But never as good as all-in-one highenders like Naim or Krell or Linn. 

[matthias]

7 hours ago, misterspense said:

The hard thing in comparing USB and SPDIF is not on the dac side, but the source, there are not many sources that have both connection types.

So stating that Spdif on your dac sounds better (or worse) than Usb, does not tell us that much, because you can then also say, from the same listening comparison, that your cd player sounds better (or worse) than your "fill in your computer or pi, with or without usb decrapifiers".

 

Agree,

but as @JoeWhipmentioned the upcoming Schiit CD transport has both connection types so comparisons will be easy.

 

Matt

[R1200CL]

(I only read first post). 

[davide256]

 

On 10/7/2020 at 2:07 PM, wbh said:

I know that USB was -- say, prior to 2007 -- a compromise compared to SPDIF.

Enter Gordon Rankin and his "Streamlength"  technology, adopted by Wavelength Audio and later (licensed by??) Ayre, AudioStream and others. I think BenchMark and Bel Canto and others had their own "in-house" USB solutions (asynch??), but I may be wrong.

I suspected  SPDIF because of objections from way back (late 80s) -- well before USB -- when I2S was being pushed by manufs. like Perpetual Technologies,  Audio Alchemy, et  .al. 

SPDIF bad

I2S good

And that has been the consensus, on forums like diyaudio and diyhifi, for over 15 years.

 

Okay ...settled ....but maybe not! ....becasue a few years ago, Schiit's Mike Moffat was openly stating that SPDIF was the better way to connect DACs because USB is inferior. Not sure whether he was referring to Schiit products, lower-end Schiit products, or all DACs (in general).

Indeed, on certain forums (notably SBAF), members were CLAIMING the same thing .. spdif sounds better.

And, a few years ago .... poof ... we have the solution...

Schiit's Unison USB.

And there are others, of course. Allo's usbridge, etc.

 

So what's the story ...  many of us don't have hours and hours to kill on A/B listening tests. And, of course, there are many variables ... so let's make the issue testable and clear: say you own a Schiit Modi Multi (which does not incorporate Unison)  and connect it via USB or SPDIF, which will sound better -- all else held equal. 

 

 

 

I did own a Gungnir with Unison, USB was always more resolving than SPDIF. I've also owned an Allo USBridge, electrically a  dirty device for its USB out.

 

The 192khz limit of Schiit's DDC's is a handicap... up sampling helps with USB audio resolution.

Most USB issues IME stem from the source device HW and software...electrically "clean" source hardware and good integration are key

to minimizing  detail haze and digital nasties. Microrendu did this right.

[Savolax]

For what it's worth, I own a Singxer SU-1 DDC which is modified to use SOtM sCLK-EX board and I had a chance to listen Schiit Bifrost 2 at home for a while.

I understand these are different products for different purpose and it might not make sense making arguments either way but I found signal qualities to be better with SU-1 sCLK-EX than using just USB input Bifrost 2 with XLR or RCA out to my Audio-GD NOS-11, in terms of cleaner sound (digital hash or whatever you would feel more appropriate).

 

By the time I made the comparison I had JCAT USB Femto as USB output. Powered with UpTone LPS-1.2.

SU-1 sCLK-EX was powered with Farad3 LPS. 

 

I guess my point here is that I did not experience any great revelation regarding Unison USB on Bifrost 2. Actually I felt Bifrost 2 performed better when input with AES from SU-1 sCLK-EX.

Based on this experience I would say Unison USB is far from perfected. I have no other experience with Schiit gear so I have no reference there how it has improved on USB.

 

 

 

[JoeWhip]

2 hours ago, davide256 said:

 

 

I did own a Gungnir with Unison, USB was always more resolving than SPDIF. I've also owned an Allo USBridge, electrically a  dirty device for its USB out.

 

The 192khz limit of Schiit's DDC's is a handicap... up sampling helps with USB audio resolution.

Most USB issues IME stem from the source device HW and software...electrically "clean" source hardware and good integration are key

to minimizing  detail haze and digital nasties. Microrendu did this right.

The Yggy and Gumby actually upsample everything to DXD resolution internally but do not accept native DXD input. So, if you feed the DAC 16/44.1, the megaburrito filter does it’s processing at 352.8. Sending it multiples of 24/48 results in processing at 384. The Yggy front has lights for 1x, 2x, 4x and 8x with the 8x being for DXD input. Maybe one day it will accept native DXD. That would also require an update to Unison which is Limited to 24/192.

[sandyk]

On 10/9/2020 at 3:39 AM, wbh said:

My Xonar sound card -- on main PC -- does spit out SPDIF, so I'll have to experiment with a Windows or Linux PC as a source.

 It will still only be as good as the receiver in the DAC permits. In my case, the Asus Xonar D2x  in the PC is connected to a highly modified Musical Fidelity X-DAC V3 which has good quality internal Isolation transformers.

[davide256]

5 hours ago, JoeWhip said:

The Yggy and Gumby actually upsample everything to DXD resolution internally but do not accept native DXD input. So, if you feed the DAC 16/44.1, the megaburrito filter does it’s processing at 352.8. Sending it multiples of 24/48 results in processing at 384. The Yggy front has lights for 1x, 2x, 4x and 8x with the 8x being for DXD input. Maybe one day it will accept native DXD. That would also require an update to Unison which is Limited to 24/192.

Which doesn’t help you with USB transmission issues. Upsampling to 706/768 at USB source IME corrects vagueness/blurring of musical detail, much happier with a Chord Mojo than I was with a Gumby

[wbh]

6 hours ago, sandyk said:

 It will still only be as good as the receiver in the DAC permits. In my case, the Asus Xonar D2x  in the PC is connected to a highly modified Musical Fidelity X-DAC V3 which has good quality internal Isolation transformers.

I had a MF A324, which I also modded* quite heavily.  At that time (2006-2009??), I experimented quite a bit with pulse transformers and decoupling those SPIDIF 74-logic buffers. Always heard improvements .... but I'm sold on USB from ~2010 henceforth. Not sure why the recent "re-interest"  ... is Schiit, with Unison, trying to keep its customer base? That company, like Tesla, has its share of fanboys (practically all of SBAF). 

That said: I'm sure Unison is an improvement. 

 

* About modding the MF ... aftermarket clock (feed it battery power) and replace all those cheap ceramic caps with Wima polys. My MF was huge (kinda like Denafrips) so much room to add stuff.

[sandyk]

1 hour ago, wbh said:

* About modding the MF ... aftermarket clock (feed it battery power) and replace all those cheap ceramic caps with Wima polys. My MF was huge (kinda like Denafrips) so much room to add stuff.

Will the attached do you ?"😉

It is also powered via an external Linear + and -15.5V supply using John Linsley Hood designed PSU "Add-ons" which have <4uV noise and a simulated capacitance of around 2 FARAD. It uses a .1PPM 24.576MHZ TXCO and added voltage regulation for the analogue ICs . The TCXO is also powered via a low noise PSU PCB fairly similar to a PFM  "Flea" You will also see better components in the DIR9001 time constant area .

[wbh]

10 hours ago, sandyk said:

Will the attached do you ?"😉

It is also powered via an external Linear + and -15.5V supply using John Linsley Hood designed PSU "Add-ons" which have <4uV noise and a simulated capacitance of around 2 FARAD. It uses a .1PPM 24.576MHZ TXCO and added voltage regulation for the analogue ICs . The TCXO is also powered via a low noise PSU PCB fairly similar to a PFM  "Flea" You will also see better components in the DIR9001 time constant area .

I know I also built a Flea clock at the time. Probably installed it in the MF. Has a separate toroid and PSU for that 18v requirement. Over on diyaudio, I think, there was a mod to get rid of the output coupling caps .... that required quite a bit of hacking, but worth it. The MF A3-24 was Stereophile Class-A rated (2002). Dunno what that means. All I know is after mods, it was really singing.

And then I eBay'd it :(

Maybe I should've kept it .... but a Magnavox-mod CD-player project sounded sooooo much better. 

[sandyk]

8 hours ago, wbh said:

I know I also built a Flea clock at the time. Probably installed it in the MF. Has a separate toroid and PSU for that 18v requirement. Over on diyaudio, I think, there was a mod to get rid of the output coupling caps .... that required quite a bit of hacking, but worth it. The MF A3-24 was Stereophile Class-A rated (2002). Dunno what that means. All I know is after mods, it was really singing.

And then I eBay'd it :(

Maybe I should've kept it .... but a Magnavox-mod CD-player project sounded sooooo much better. 

I am also a DIY Audio member. I didn't previously mention that with  a small amount of tweaking I was also able to remove the output coupling capacitors. The NE5532 were replaced with OPA2134, and the better sounding metal can LME49710HA for the outputs. I don't need additional PSUs as the JLH PSU add-on has <4uV noise, and even the small PCB for the TCXO already takes it's input from  that highly stable and low noise source .

 Incidentally, for those that may not know of the MF X-DAC V3 , it was originally supplied with an A.C. wallwart.

 The X-DAC V3 was modified to accept dual polarity DC from the external Linear PSU.

 

[wbh]

I found the pix I used to Audiogon the A3-24 (or was it eBay ... me fergets!) .... anyway, this one has labels. As you can see, the default PSU is quite beefy ... I didn't touch it ... but the mods are labeled for the customer ...

 

I replaced the mostly useless spdif coax out, with a much more useful BNC spdif input (the preferred way to spdif ... yes, superior to RCA)...

Beauty shot of the I/V and output stage. Note the grounded Cu foil over the DAC IC .

 

 

Mostly the digital section. I had removed most of the cluttery mods (Flea clock, etc) in prep for the sale. Note the blue Bourns trimmers ... those were used for the direct-coupled "servo" output ...

[RikkiPoo]

Quote

Why do you include S/PDIF and AES3 outputs when these transport methods cannot handle high bit-rate files?keyboard_arrow_down

It is true that S/PDIF and AES3 outputs have limited bandwidth compared to the full-bandwidth available with USB, Ethernet and I2S. However, there are downsides to having Ethernet and/or USB receivers in the DAC. Both introduce noise into the DAC. USB has the effect of adding a trace of ‘sizzle’ at the edge all sounds, creating some listener fatigue. Ethernet has the effect of rounding the edge of all sounds, diminishing the tension and urgency of real music. The extent of these problems depends on how well these inputs are implemented in the DAC. If you compare these outputs to our S/PDIF and AES3 outputs you will hear the benefits in naturalness and musicality for yourself.

From the Antipodes FAQ page.  My Terminator DAC's USB is turned off unless triggered on by a USB signal.

[sandyk]

On 10/16/2020 at 11:45 AM, wbh said:

I found the pix I used to Audiogon the A3-24 (or was it eBay ... me fergets!) .... anyway, this one has labels. As you can see, the default PSU is quite beefy ... I didn't touch it ..

Interesting photos. I doubt though that the internal PSU would come close in performance to that  of  the external Linear PSU that I use with my X-DAC V3 with it's <4uV noise and a simulate capacitance of around 2FARAD on each supply rail.

 I also bypassed a couple of the onboard rectifier diodes as they were no longer needed.

 The interesting thing though is that originally John Atkinson gave it a very good review, including measurements, but the 3 of us tech types who purchased them as a result, found them quite mediocre. Hence the modifications that were also performed by several other members of a U.K. based forum with similar results as mine. I had to talk a couple of members through repairs after they made a few errors in the analogue PSU area .

 I also had a  direct coupled output, as I also do with my fully D.C. coupled  DIY Class A HA and  Class A Preamp, as well as my 15W /Ch Class A Power amplifier. However I didn't need a servo as DC out was fairly low, and only needed a very minor change of resistor values to get close to 0mV . 

 I do however use a very low noise servo in the Class A HA and Preamp .