USB Nervosa Thread Decrapifiers, pro interfaces, and bears oh my

A typical measurement suite takes me between 6 to 8 hours, to assure quality measurements. Finding discretionary time is always an issue for a corporate techno-slave. When I do find some time I will probably only compare the inferred jitter tests for a given DAC with and without various DDC, reclocker, decrapifier etc in the path. But probably not happening in the very near future. There are other tasks awaiting in the queue.

 

I think we agree. My point is to see USB and alternative computer digital interconnects for what they are and have realistic expectation of what they will get out of trying to optimise USB or investing in USB alternatives. I think both of us speak from experience of wasting money on external USB converters a number of years ago.

Personally I don't approve overinvesting in USB defuckifiers because the system if already limited on both ends. Actually I would say that if you need defuckifiers that is directly indicative that the source and DAC (or at least DAC USB input) are already limiting. I also would warn against jumping into a different solution that is also limited. I tend to challenge the label defuckifiers though, I don't think you can fully defuck something that is already fucked by adding something after the fact. Almost always there is some sort of tradeoff. Maybe refuckifiers is more accurate, or fuckredistributors. Something is always sacrificed, even if what is sacrificed is money that could bo to a better DAC or source or cabling or at least something that is moving forward and not sideways or at an oblique.

I think some people just like to tinker with stuff which is fine also as long as they have realistic expectations and some self awareness. I think some of the more dedicated DIY computer music server builders fall under this category - they recognise that the computer is a flawed source and just want to see how far they can improve the flawed system and learn something in the process. I mean maybe some people just enjoy messing around with the sort of oblique stuff for it's own sake. After a certain point in hifi almost any move will be more sideways than forwards anyway.

 

This last paragraph is what I would term a good definition of some of us in this hobby…

As for the admonishment of "realistic expectations", I see that as a necessary portion of the learning curve in terms of why 'tinker' and 'tinker' with what and then, most importantly, what happens next.

Of course some will certainly have expectations, which may or may not be realized.
But then that is also part of the learning curve, and could be construed as that line in the sand, or the test to see if they really should be 'tinkering' in the 1st place.


JJ

 

I reckon I have lost about $2500 buying and selling USB/SPDIF converters:

-AudioGD Digital Interface version 1
-JKSPDIF mkiii
-Audiophilleo 2 with Purepower
-Bel Canot Reflink
-Wireworld Platinum AES cable
-Oyaide BD510 BNC coax cable.

I'm still trying to sell the Reflink which is good converter in its own right but not the cool new thing.

I mean its easy to be self hating audiophile and consider that that money could have bought a Yggdrasil but I think that would discount the enjoyment in the process of wasting that money on those USB/SPDIF converters, as well as knowledge gained from that process. I don't think its possible to progress without making mistakes, at to be honest at each of those purchases the decision was rational given the knowledge I had at the time. How was I to know that Yggdrasil would turn up. All we were aware of up to that point was PCM1704 stuff which i tried and didn't like.

 

Yeah sometimes the learning curve can include expensive and time consuming side roads, especially in this hobby where doo-dads and the latest and greatest will or can become yesterdays news all to easily.

There will Always be something new that pops up, and many of us will fall into these exploratory investigations, and hopefully come out the other side with 'Better'.
But that's the nature of bleeding edge investigations they do tend to be, as Atomic Bob calls them, 'messy'.

This recent move into exploring AOIP and its strengths and weaknesses etc. is yet another example.
Only this time the results can be stellar, if it works for you and if your system can scale right along with this 'new' digital pathway to the dac.

It certainly has made improvements in my system and I'm not alone.
And this is but the first round of merging these varied, network based components together and it will morph and all of us 'early adopters' will or might get 'stuck' with yesterdays news.

Only the results attained thus far are more than worth it.

JJ

 

That's an interesting statement!

Thanks, JJ

 

So well put. Early adopters are a different group looking for the same end goal, the thrill of the sound. And they know what they are doing and that's how the hobby evolves, for sure. Successes built from trial and error is sweet, but the fun is also in the trying, as it should be, with eyes wide open. The wisdom is acquired from this, the understanding that most changes will be sideways.

People are pissed that people are misled that these devices make a big difference. The truth is, if they did make that much change, they would be bad, like tone controls that alter phase, or low pass filters that reduce room modes. But doing those can also help, in the right circumstances.

I am always skeptical of absolutists, of all kinds, the bits are bits crowd, the it can't be changing, the I can't hear it so it's not there, immeasurable so nonexistent ayatollahs. More recently, the rise of the it's the next greatest thing changed my world, which is the true noob nuclear garden of Eden Apple. Or fanboy sites misled by opinion leaders and their ranked choices.

Or the people self gratuitously posting after listening for an hour, switching back and forth, such a sad pointless lazy read.

The one thing you can't quite be sure of is how it will sound your system. And changes take time to settle. When you speak of a resolving system, that's like saying you have a fast car, which is so broad a term that it encompasses in their minds a BRZ (fun, not fast) to a Veyron (really fast) with infinite variations in between. And these have different use cases, some are family cars (Teslas) some are track cars (Formula E).

And in a lot cases they can be adjusted by changing suspension setups, tires, fuel, batteries (in Electric cars). But in the end those are secondary. It's still the same car, slightly different.

BUT if you already own the whole rig, you can't change one or two things easily, and not expect to really lose an arm or a leg. So you work to get closer to the sound you like. But remember the big picture, please. A 1k USB Spdif tweaking a 10k component's sound is a valid value proposition, but not a $400 Spdif helping a $400 dac. That's pushing it. It's your choice of course, even if I disagree.

A tweak is just that, a small screw in the engine of reproduction. Don't pay for it like it's an engine, it's not. Buy a better car, first. Seeking to understand is hard. Sparing the noobs is a great thing, but tweakers and Early adopters clearly bring value, also, if one follows the whole thread, not just one post.

 

@drez your point about progress and learning though mistakes is such a good one. All the great inventors made plenty of mistakes on their paths to the successes we celebrate and remember. We often forget the order of magnitude more mistakes logged in their notebooks.

I have my Reflink. It isn't the latest, greatest, shiny and FoTM, but that doesn't detract that it performs as intended very well and has for some time. Nor do the latest developments invalidate the hours of enjoyment we both have derived from it.

@johnjen Also well put. The hours experimenting to develop points of auditory reference to which other experiences may be compared is messy indeed. But once reliable, repeatable experiences may be demonstrated, I find it interesting to explore how far down the cost chain one can descend while maintaining the greatest level of performance for the various attributes of the reference system.

 

Useful toys for sure. I want to come over to your house.

Actually I was having a pipe dream that some day I'd find time to take a closer look as well, and I started reading the XCHI spec. It go a bit complicated and real world priorities took over... so this may well be a stupid question. At least on Linux, I think the audio driver sets up the interval for the isoc endpoint, provides a list of frame descriptors, and the host driver handles the scheduling from that point forward. When you say U12 was an excellent performer, I think you're saying that any XMOS endpoint with ASIO running on the host is going to be an excellent performer? If not, what factors would change the USB interface performance between different XMOS endpoints?

 

This reminds me of the saying about being a Master…
'A Master is one who has tried and failed more times than the student has even tried'…

JJ

 

Another similar and related aspect is the attitude that I am going to save those unsuspecting consumers of wasting their money.
Which while it can be true enough, is an example of one elevating themselves to the status of being the guardian of truth or some such.
Which is a close cousin to the end user who is looking for the quick fix which 'solves' all of the problems both real and imagined.
There are Always trade offs with any tweak, or mod or any other form of magic bullet…
And the more sophisticated and tighter the design of the device being modded, the greater these unanticipated secondary changes will be.
IOW there are no free lunches.

That is so true. Quick and dirty evaluations while possible are ofttimes fraught with a high degree of uncertainty, built in.
And the only true way to really evaluate any change or new fangled thing-a-ma-bob is at home in ones own system over weeks as the changes settle in.
At least that is what I have learned over decades of fussing and tweaking.

I agree $$ spent to upgrade and the cost of the original device should strike a balance between them, but not always as some gear can get 'uncorked' to a remarkable degree, but this is rare.

Yes some tweaks may be a small adjustment while others can be much more over the top, like rebuilding the entire circuit using better parts etc.
But there are some mods which can be low cost and can make significant improvements, but they usually involve a more hands on approach and a degree of technical expertise to begin with.

And for some buying an entire new device may not be feasable $$$$ wise, but throwing $$ at an experiment could be considered as a 'mad money' trial and error experiment.
BUT experience (or lack of same) along with the strength of the urge to 'make it better' can be a powerful influence and learning tool in and of itself.
So for me anyway, the trick is keeping the perspective of discovery, rather than the expectation of transformation, is the key.

This distinction (discovery vs transformation) also helps in keeping the $$ of the tweak and expectation biases in check, as in weighing the $$ and time and effort vs potential gains, but there are also the associated secondary consequences, which in some cases can be more pronounced than the primary desired outcome.

But as long as tweaking is kept in 'proper context' a great deal of experience and learning can be gained, and in some cases there is no $$$ equivalent to learning what happens when an experiment 'goes to far'.
And this understanding may only come into ones awareness at a much later time.

JJ

 

Mike Moffat didn't seem to like XMOS based usb decoders for high power draw issues:

http://www.head-fi.org/t/784471/what-a-long-strange-trip-its-been-robert-hunter/780#post_12713000


Also it seems new versions of Windows/MAC OS is getting more restrictive on USB output power current:

http://www.head-fi.org/t/784471/what-a-long-strange-trip-its-been-robert-hunter/780#post_12712479

 

Anyone who is reading this thread thinking that they may need to spend big $ to "fix usb", take a step back, breath into a paper bag, and count backwards from 10.

These new usb converters / reclockers should only be used for tweaking the last 0.05% - 4% of sound if you already have end game dac. Save your money (and sanity) start up a savings fund, and wait until you have Yggdrasil $, top of the line headphones and amps. I own a mutec 1.2, and while I like what it did, I would have been better off waiting for Yggdrasil, better turntable cart, or a different set of cans. I bet most could not blind a/b accurately (I couldn't), as DAC changes are hard enough to hear.

With all due respect to advanced audiophiles, i know you hear night and day differences. I am more trying to warn average sbaf reader that these are last avenue to go down, and not day 1 need to own or your sound sucks type products.

TLDR: be careful, differences are real but very hard to hear, only tweaking last 1%, save your money for better dac / analog system

PS more or less just hyjacked zero's usb/digital rant, sorry

 

I'm going to use a photograph analogy here to explain what I think are the subtle differences between Mutec Mc3+USB+Yggdrasil vs Yggdrasil Internal Gen3 USB:

RAW = Mutec Mc3+USB > Yggdrasil
TIF = Gen3 USB
Both are great photos, just that one has slightly better tonality, slightly better defined shadows/dark area. So you have to decide if these slight differences is worth spending ~USD$1000 or not.

 

Wise words. Though I think you did, anyway, lol. No worries I don't think anyone said anything about night and day being exclusive to more resolving systems.

USB is good.

For my Gungnir Multibit, SPDIF from my computer is better ( with the the Berkeley) and in my setup it's a better tonal match than an Yggdrasil (detail oriented system will sound too dry). I change sources, and I can dispense with the the Berkeley. But that's fed Spdif from a CD player that cost 4 times the tweak. Is the CD player worth it? I already owned it so..

Caveats: I actually chose the Gungnir Multibit because a) it's digital b) it will be obsoleted c) it totally balances better in one of my setups d) is better Spdif and e) the detail I get out of it with the Berkeley USB brings it very close to analog with most good material. It's harsh with bad digital.

Can I get something else? Of course, and I'm actually looking right now. Do I like listening this way? It's okay, my CD player is better by far, to me standalone ( and someone trolled me about that here, too). And my table is better to me, in my setup. Is it resolving as a setup? it's among the better ones I've had in literal (not figurative) decades. Are headphone setups more revealing? For certain things absolutely. But for depth, soundstaging, decay, lifelike cohesiveness, no. Is my 1% more than someone else's, who knows? I know I remove anything in the setup that causes me to not hear changes, asap.

Again, if anyone hears night and day differences for USB regens et al., I've never been in that planet, tell me where that is. Do I hear Axl Rose's sex sounds in Rocket Queen better? No, not that there's anything wrong with that.

Noobs should should listen and buy the basics, 100% no f*cking queetion. And if you like it as a hobby and want to seriously write something worth reading, slog through this.

http://www.stereophile.com/reference/50/#h12UV5ZSqmAqPoXD.97

Rant over.

 

All of my comments are for WinX systems, which is NOT a real-time pre-emptive OS. Well, at a particularly low level it is, but that is between the special print spooler, the mostly hidden DOS command prompt and the windows thread. For the Windows portion it is merely best effort. Audio driver implementations vary all over the map in ability to deliver their respective data streams. I cannot comment on linux not having used it for audio. In a past life I did *nix programming. So my fondness for a specific XMOS driver is also WinX specific.

 

The U12 to which I frequently refer has an in-board Linear Power Supply to address both the high power draw as well s SMPS power supply noise. Otherwise I don't think it would perform as well.

 

I would agree with the analogy if it actually cost $1000 to shoot RAW and convert it manually later. There's even free software for that.

 

Mutec = Digilloyd