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

This is my thought for the day. Actually, it's my thought for life. It's probably crap, but hey...

Decent engineering is for the real world. A power supply handles, or should do, the vagaries that arrive at it. And if some very expensive piece of kit is not able to handle the stuff that arrives at its various orifices, without the sticker of things therein having to worry and spend and worry and spend and worry and spend, then it could well be over-priced bad engineering.

Mind you, I love balanced cabling. I love it because it is engineered, whereas phono stuff was a historical accident. If I currently had balanced gear, I'd use it. Hats off to its developers. But I don't pretend that I need it for a ten inch interconnect, or a six-foot cable. There are extreme situations, but I don't live next door to a broadcasting station, power generating, or a medical lab running high-power magnetic and x-ray stuff. Some people do. (they probably need a Faraday Cage more than nervosa hifi equipment! Or lead-lined walls.)

Then, yes, I know this is SBAF, not Hydrogenaudio, but... What consenting adults and hifi equipment stick in each other's orifices in private is entirely up to them. I won't say I don't judge: I have my own opinions, prejudices, tastes, preferences, and jesus-that's-disgusting levels, but, still, it's their right to do that stuff. But if they start to tell me that it makes music sound better, then my disbelief soars unless they provide at least some level of objective testing.

I don't ask Hydrogenaud.io approved double blind. Just some sort of blind testing, even half-arsed.

Yes, it's just me, and I know... better I just stayed away from threads like this and kept my ellipses to myself...




(By the way, how did bears ever get into the thread title?)

 

It's a play on words, referencing this...

 

After using an ifi isilencer + on all four USB ports I regularly use, between 2 PC's and 2 different setups, for a few weeks without removing them, and then actually A/B'ing them on both setups, I've concluded that the change the "ReBalance" feature makes to the signal line is just bad -- it removes low-end and decay detail, but there's fake topping A90 type clarity added, and my TH900 chain I had those positive first impressions on didn't suffer for this as much as other setups I tried later.

The first unit I bought was also a return without being advertised that way, and had a lot of burn-in, because the other 3 all sounded worse when first installed.

I think the 5v power filtration they do is still legitimate, if you have some USB ports on a surge protector or outlet you want to add some filtration to, or if you have an adapter to inject power into a USB cable w/o using the mobo's, you could use a second port and the ifi isilencer+ that way to avoid using another outlet. These are poor value propositions, but might make sense if you got an isilencer + very cheap used or included with other used gear you're getting, idk.

I wound up sticking 2 onto USB ports on the surge protector for my bedside rig, to power the active USB cables to each PC, and I plugged the other 2 into the outputs of my ipower x -> ipower USB combo to create "ifi's Basilisk" powering the cables in my deskside rig. I really doubt I'll hear any improvement with this over the previous setup, especially as all the Basilisk powers are tiny USB signal repeater chips, but it doesn't sound worse for the addition, and I just subjectively like having more capacitors in the chain.

 

I’m currently testing an SMSL D-6 with HQPlayer via NAA/USB (with great results, BTW). I noticed that it sounds better connected to my Zen Stream than it does when connected to an RPi4 (which uses its stock PSU, which I heard is exceptionally noisy): On the Pi, there’s some treble fuckery going on, more sibilance, more fatiguing. So I guess the noise is spilling into the D-6 via the ground pane.

Unfortunately, the Zen Stream only supports rates up to 384 kHz/DSD256, so I’m looking for ways to silence the Pi that don’t cost an arm and a leg and support at least 768 kHz/DSD512 (USB). I’m more curious than determined, so I’d appreciate if I didn’t have to spend more than € 100,–.

Suggestions?

P.S.: I’m not looking for a signal reclocker (though it doesn’t hurt if there’s one included), as I don’t think that will make much of a difference, the signal is buffered on the receiving side, anyway. I mainly want to keep the noise out of the D-6.

 

iFi iPower if you want something cheap, LPSU for best performance

 

In addition to using a better power supply (which does seem to work, even though the Pi's various internal supplies are switch-mode anyway), you could try turning off a lot of the superfluous stuff which might be producing audible noise (this is a nervosa thread, after all ). Scattered through this thread are scripts to turn off the Pi's LEDs, wifi & bluetooth, and HDMI; and to under-clock it.

 

My friend went down this road with the Zen stream as a replacement for a pi4 with the entry level ifi power supply. The stream had some issue and it went back. He then tried the $100 and $300+ ifi supplies and found improvement in SQ with his pi4. And unfortunately the most expensive sounded the best.

 

Thanks for your answers so far.

What about stuff to put in between the Pi and the D-6? (If it it sounded above as if I didn’t want that, excuse my poor description, I just wanted to say that I didn’t expect reclocking specifically to make a difference, not that I didn’t want anything at all between the two.)

While I would expect a better PSU to make things better, I still wonder how much noise the Pi produces itself even with the best PSU on earth. Wouldn’t it make more sense to filter out the noise between Pi and D-6 than to give the Pi clean power? Or do you think that such filters are limited in what they can do, and a better PSU would be required, anyway?

Has anybody tested the effects of a PSU as opposed to those of various kinds of USB cleaners?

 

My friend does not use a re-clocker between his pi and Audio Research dac. He does use good ethernet switches (not crazy expensive ones btw) and upgraded power supplies on them along with cat 8 and hears a nice jump in SQ.

I have tested 4 or more reclockers from TOL Mutec on down in my system over the years. While the sound does sound different, I did not buy any of them even at a good discount. My stance is that if the dac has a very good usb input, they are not needed. Then again maybe I don't have the requisite high enough tens of thousands of dollars invested to be able to appreciate them.

 

Yeah, well, the problem in this case is that this is a very cheap DAC that only provides great sound because it allows to deliver HQPlayer upsampled data unaltered to the DAC chip (Direct DSD), without HQPlayer it doesn’t sound good at all.

So the USB input is stock XMOS without any isolation whatsoever—which probably is the reason why I hear a difference at all between Zen Stream and Pi. I guess I wouldn’t need any of this, with, say, a Holo DAC, e.g.

 

I've tried lots of LPS and re-clocker/decrapifier stuff over the years (everything from old ADUM4150 stuff to Wyrd to Uptone ISO Regen to now a SOtM hub), and I'd largely agree that you're better off with LPS stuff for bang for your buck. There's less logic in stuff like this ("galvanic isolation should sound better" does not lead to does sound better) and more just trial and error. It is true, though, IME, that better DAC's with better digital input circuitry etc. tend to yield less returns with upstream stuff than cheaper DAC's, sometimes. If you're set on decrapifiers, just get an old Wyrd/Uptone Regen/one of the iFi decrapifiers on HiFiShark and try it out. Don't spend more than $100-ish. You can get a Wyrd for like $50.

 

Here's a cheap Wyrd on eBay

 

Hi!

I have a couple of questions which are very much in the weeds, so feel free to say you aren't sure or whatever. I have just been embarking on this journey and now am at the stage of straightening out my digital chain. I am currently running the cheap usb card, like Purrin suggested. This is all with a view to using a PC as the source.

What's the consensus regarding Unison vs AES? I believe Mike suggests Unison might be preferable.
I am asking because I have the option of getting a Dante to usb converter with dedicated PSU, or, for close to the same price, the AVIO Dante-to-AES PoE converter. Or I could just do the more common thing and get a singxer su2 for USB to AES.

Any contributions area appreciated

 

your local Grumpy Old Man says...

Especially if you have an interest in this stuff as a hobby, my suggestion is that you listen and see. This is not to say don't ask for recommendations, but if you take this path, you might find yourself happy and richer.

Sure, you might also spend more money in total (the better-to-buy-once principle) but if one can keep that within reason, at least you will have made the journey.

I remember my early days of computer audio, nearly 25 years ago. I went from standard to GBP 5, to GBP 50, to GBP 100 (200, actually, discontinued card at half price) internal cards and every step was clearly worthwhile. In fact, I thought that the RME card which I kept for many years, sounded better than my CD player. (The next step, in those days, was Lynx at GBP1,000-plus --- and that I never took)

If we do not progress in that kind of way, we not only loose the thrill of discovery, but we allow the establishment of all sorts of nervosa unnecessaries to become standard equipment. Part of the shopping list that every beginner takes to the store.

I was struck, some time ago, by a member telling something like "I was very happy with the sound, and then I read that this or that could sound bad." He read it; he didn't hear it.

/GrumpyOldMan



Oh... and... I've meaning to ask for ages... How did bears get into all this?

 

Years later, I’m still preferring Dante/AES file delivery. For me the truest timbre of all instruments and sounds in the most real (perceived ) space. This despite upgrading YggyB to Unison (for 2 yrs additional warranty).

Be aware that Dante does not auto-change sample rate, if you have widely varying files.

 

I think I quoted this exact thing about 80 pages ago, but it's a play on words of this scene:

 

What should one do to address this? Upsample/downsample everything?

 

I don’t think that would be much fun. Dante not viable in that situation.

In my case, almost all the music I want to listen to is readily available in Redbook, so my rips are 44, as are Tidal files, and Yggdrasil does a superb job with them.

 

Thanks @Thad E Ginathom

Also thank you for this! I did not realise!!! Ijust cancelled a purchase.

 

@atomicbob can explain more, but Dante is hardware/software/transport solution that comes from the pro audio world. Since it came out, more agnostic software/transport solutions are available for the audiophile, for example Ethernet > Pi plus hats > AES devices, or Roon's RAAT over Ethernet > whatever. On top of this USB is of course dominant, and certain companies (e.g. Holo Audio & Schiit) are making it work at a very high level.

That said, in any given situation & investment Dante might work for you but you have to pay attention to the details.