advertisement


High end streamers, waste of money or not, please?

I've used both the Sonos and Blue OS app and there's little between them. Sonos probably edges it, but i'd sacrifice that for the superior sound quality that the Bluesound offers over Sonos.
I’d agree, we’ve had SONOS since it came into the UK and have to give them their due that all these years later we kept it in preference to the Bluesound Node. That’s in the context of whole house music for family and friends, the App is still top of the pile for general non enthusiast use. Into a HiFi system I would have chosen the Node over the SONOS, it sounds a bit better and can do Hi Res.
 
£1k altogether, £4k saved over the Melco and better performance to boot. What’s not to like?
I've never used a Melco so I've no idea and would never spend anything close to 5k on a piece of equipment.
The Innuos server that I have was a 2nd hand unit and cost slightly more than a new Stream Box S2.
 
Best sound I’ve gotten from USB (i.e. matching the optical input into my Hugo TT2 from a Meridian 200 CD transport) to date has actually been the simplest and the cheapest. A raspberry pi 4 (4gb model) housed in a flirc aluminium case, running Moode, powered by an Anker 26800mAh Powercore and attached to my TT2 via a Supra USB cable. No usb ‘cleaners’ or any other foo necessary. Couldn't tell the difference between the Anker and the Salas PSU (via www.diyaudio.com) I’d made from a ‘kit’ a while back, so the Anker must be very good indeed.

Extremely happy indeed, and has completely scotched any previous ideas I had about ‘upgrading’ to a ‘high end’ streamer. The beauty of it is the already very good streaming software is free and always being updated, any future Raspberry Pi upgrades can easily be purchased without breaking the bank. Contrast that to a pricey server whose manufacturer could stop updating the software at some point, or stop offering support etc.
 
Best sound I’ve gotten from USB (i.e. matching the optical input into my Hugo TT2 from a Meridian 200 CD transport) to date has actually been the simplest and the cheapest. A raspberry pi 4 (4gb model) housed in a flirc aluminium case, running Moode, powered by an Anker 26800mAh Powercore and attached to my TT2 via a Supra USB cable. No usb ‘cleaners’ or any other foo necessary.

Extremely happy indeed, and has completely scotched any previous ideas I had about ‘upgrading’ to a ‘high end’ streamer. The beauty of it is the already very good streaming software is free and always being updated, any future Raspberry Pi upgrades can easily be purchased without breaking the bank. Contrast that to a pricey server whose manufacturer could stop updating the software at some point, or stop offering support etc.
Totally agree. To be honest, a simple Pi 3 or 4 with usb into a decent Dac is all you really want.There may be noise but I doubt it is hearable. All this jitter angst needs to be confined to the audio dustbin now that well fettled dacs were on top of this years ago.I’m now a firm believer that high quality can be had for relative pennies and so called high end systems add very little really. I visited Audio Science Review review recently and wondered whether all this focus on measurement was a distraction from the real goal. If it engages you, who cares!
 
I think the supposed usb ‘noise’ issue on the Pi3 was just typical audiophool Chinese whispers BS taken as ‘fact’ at some point and spread as such. Some dislike the idea of anybody being able to afford top notch sound at budget prices I think, as it takes away the snobbery and bragging rights some feel the need to go on about etc. All too prevalent a pattern these days alas. Archimago measured the Pi3 and it was found to be perfectly up to the task and not electrically noisy on its USB output at all. Having said that the USB and ethernet lines were shared on Pi3, and apparently that’s not the case with Pi4. So I assume the Pi4 would offer an even cleaner usb output. Not that that is needed for any half decent dac.
 
The thing is, jitter has to be significant to hear an effect. Most is below this level and controlled by the new generation of dacs so best just to ignore any jitter angst, now a marketing tool to get listeners to feel paranoia and therefore pay sometimes ludicrous sums needlessly really.
 
Last edited:
The thing is, jitter has to be significant to hear an effect. Most is below this level and controlled by the new generation of dacs so best just to ignore any jitter angst, now a marketing tool to get listeners to feel paranoia and therefore pay sometimes ludicrous sums needlessly really.

Indeed.. async USB kills the jitter bogeyman which leaves the dreaded USB noise bogeyman.

I am using a low powered Gigabyte BRIX mini-pc running a custom ram-root linux I built that runs squeezelite and obviously usb into my DAC.

I use a HDPLEX LPS to provide the needed 12V.

The BRIX has no drives and music is sent via wifi from a fanless, noiseless server with SSD drives and LMS.

The server is at the back of my room to keep wifi strength high.

The endpoint also contains Kodi so I can play my ripped music DVD's and I use an nfs mount for Kodi to see the DVD rips stored on the server.

No need for fancy audiophile ethernet cables, routers etc or fancy femento USB cards etc.

To test USB noise, I took a pink noise WAV track, run high-pass filtering, low pass filtering and compression against it to create a test WAV file that produces the faintest 11khz signal.

I then played that test WAV via LMS and turned my volume up to "11" and what did I hear.... the faintest 11khz signal and the usual amp "rush" given the volume was at max.

By pausing and restarting the playback, I could objectively determine that there was no additional noise added via the PC playback chain.

In terms of subjective playback quality, I have benchmarked this setup against commercial products and it has never been found wanting.

Note I dont stream tidal/qobuz etc... only local playback of ripped CD's.. but LMS supports all that if I wanted to.

I am a great believer in less is more with regard to computer based playback and it boggles my mind the complexity that some computer audiophiles insist on having in their playback chain (from the router all the way to the DAC)

They obviously enjoy the tweaking side and will swear by "vails being lifted" multiple times in succession when changing PC memory, SSD drive types, finding a specific CPU clock frequency sweet spot, adding USB reclockers, noise blockers etc but surely only so many vails can be lifted in succession.

Many of the "all in one" commercial PC's have heroic levels of hardware stuck in them and from my point of view, the more "stuff" you need to power, the more noise they generate which means the probability of that noise getting into the DAC via USB increases dramatically.

If you want to upsample (as I do on my server) or convert PCM to DSD etc... do that on the server side but keep the endpoint as simple as possible.

A small, lower powered, diskless, ram-root linux end point running in a client/server topology with the endpoint running any flavour of playback software you choose (squeezelite, roon etc) will provide a high quality experience into any well engineered USB DAC.



Peter
 
They obviously enjoy the tweaking side and will swear by "vails being lifted" multiple times in succession when changing PC memory, SSD drive types, finding a specific CPU clock frequency sweet spot, adding USB reclockers, noise blockers etc but surely only so many vails can be lifted in succession.
I think you generally have problems after lifting the seventh! :D
 
I agree with complexity and I also feel that the more expensive stuff is over engineered in order to create a sound bite and selling point to justify sometimes extraordinary cost.
 
Indeed.. async USB kills the jitter bogeyman which leaves the dreaded USB noise bogeyman. ...

To test USB noise, I took a pink noise WAV track, run high-pass filtering, low pass filtering and compression against it to create a test WAV file that produces the faintest 11khz signal.

I then played that test WAV via LMS and turned my volume up to "11" and what did I hear.... the faintest 11khz signal and the usual amp "rush" given the volume was at max.

By pausing and restarting the playback, I could objectively determine that there was no additional noise added via the PC playback chain.

In terms of subjective playback quality, I have benchmarked this setup against commercial products and it has never been found wanting. ...
Yes - exactly right IMHO. That test is what I did to investigate the USB noise bogeyman with my own kit. And in my case too it's simply not an issue.

I generated a range of "signature" digital noise signals (including the 8 kHz USB micro-frame rate and its sub-harmonics specifically for USB, but others too). I leaked them onto the audio path at a low but audible level to learn how to recognize them. I then checked that at some lower level they did disappear audibly with the system volume turned up to maximum.

I am not saying that all streaming kit is immune to digitally-originated noise but I am happy that mine is sufficiently good to not worry. And I suspect a lot of modern kit is good enough too without needing to enter the world of high-end streamers.
 
IME, the stuff in front of the DAC does matter, but I would not spend more on it than on speakers, amp or DAC.
 
Archimago measured the Pi3 and it was found to be perfectly up to the task and not electrically noisy on its USB output at all. Having said that the USB and ethernet lines were shared on Pi3, and apparently that’s not the case with Pi4.

The Pi 4 does measure slightly better but I have a 4 and a 3b and it isn't audible on either.
 
IME, the stuff in front of the DAC does matter, but I would not spend more on it than on speakers, amp or DAC.
If it sounds great, it doesn’t really matter. A well implemented Dac will deal with jitter and then it’s a matter of the sound.I think the rest is audio neurosis.
 
I am a great believer in less is more with regard to computer based playback and it boggles my mind the complexity that some computer audiophiles insist on having in their playback chain (from the router all the way to the DAC)
In my professional engineering life this was the approach I always adopted. The quote I like to keep in mind is from the writer and aviator Antoine de Saint-Exupery: "It seems that perfection is achieved not when there is no more to add but when there is no more to take away". An approach I like to apply to assembling audio systems.
They obviously enjoy the tweaking side and will swear by "vails being lifted" multiple times in succession when changing PC memory, SSD drive types, finding a specific CPU clock frequency sweet spot, adding USB reclockers, noise blockers etc but surely only so many vails can be lifted in succession.
There's nothing wrong, of course, with enjoying tweaking as part of the hobby. But not everyone does. I prefer to identify the fundamentals and try and get them sufficiently right that tweaking becomes unnecessary.
Many of the "all in one" commercial PC's have heroic levels of hardware stuck in them and from my point of view, the more "stuff" you need to power, the more noise they generate which means the probability of that noise getting into the DAC via USB increases dramatically.

If you want to upsample (as I do on my server) or convert PCM to DSD etc... do that on the server side but keep the endpoint as simple as possible.

A small, lower powered, diskless, ram-root linux end point running in a client/server topology with the endpoint running any flavour of playback software you choose (squeezelite, roon etc) will provide a high quality experience into any well engineered USB DAC.
I agree that keeping the noise down at source is always a good thing but the fundamental in a streaming system, IMHO, is to find a DAC that has sufficiently good digital noise rejection over an asynchronous USB interface.

In experimenting with streaming systems over the years I have arrived at the same preference as you. A low-power end-point attached to the DAC via asynchronous USB, which acts just a network bridge between USB and Ethernet and/or WiFi. The server with its files and streaming service access gets hidden somewhere away from the audio system.

I continue to enjoy an inexpensive digital streaming system working into relatively much more expensive audio kit.
 
I was considering using Vortexbox again and using the Pi as the player. May experiment in the future.The ripping facility would be good as well.
 
There's nothing wrong, of course, with enjoying tweaking as part of the hobby.

Indeed and I never would criticise how anyone gets their kicks in this hobby (I personally do some kinky stuff!!!) ... just get a bit perplexed with the logic behind some of the changes executed and how they really can "lift veils".

the fundamental in a streaming system, IMHO, is to find a DAC that has sufficiently good digital noise rejection over an asynchronous USB interface.
.

The interesting thing is how many different USB receivers there are in reality (which is where the noise rejection is taking place...correct me if I am wrong).

You have XMOS and Tenor as the main vendors (I think) and then some lower level brands.

Any decent DAC is going to use a name brand USB receiver(typically XMOS from what I see) so I dont think its too hard to find a DAC that wouldnt have good noise rejection.

And a decent DAC doesn't mean expensive.... the flavour of the month Topping DAC's all use XMOS and they start at very reasonable $/£/€'s

Maybe there is some additional noise rejection circuitry after the USB receiver that is proprietary to each DAC vendor that means even when using the same "gold standard" XMOS receiver, there is still some difference between DAC's with regard to noise rejection?

Peter
 
The interesting thing is how many different USB receivers there are in reality (which is where the noise rejection is taking place...correct me if I am wrong).

You have XMOS and Tenor as the main vendors (I think) and then some lower level brands.

Any decent DAC is going to use a name brand USB receiver(typically XMOS from what I see) so I dont think its too hard to find a DAC that wouldnt have good noise rejection.

And a decent DAC doesn't mean expensive.... the flavour of the month Topping DAC's all use XMOS and they start at very reasonable $/£/€'s

Maybe there is some additional noise rejection circuitry after the USB receiver that is proprietary to each DAC vendor that means even when using the same "gold standard" XMOS receiver, there is still some difference between DAC's with regard to noise rejection?
Noise coupling between input and output may depend more on how the interface is integrated into the DAC than on the interface itself.

Looking at the list of coupling mechanisms is probably not for a general audio group but maybe just one. For the "common impedance coupling" mechanism, if a noise current on the USB ground line is allowed to flow through a stray impedance that is in series with an output circuit then it becomes output noise voltage. This is avoided by good layout practice regarding digital and analogue "grounds" (which is not an expensive matter unless it needs multiple board iterations to get right), or possibly by a ground isolator (which may come at some cost).
 
Jumping in the conversation late, but for years I have tried to buy a number of higherend streamers and I just couldn't get them to significantly outperform my reclocked sonos and DAC. More recently I finally made the move to Roon > USB > Dac (Qutest then H390). I really would love to hear a better front end - very curious when the Hegel H390 gets the Roon update if it outperforms vs USB.

My mind is still open, but just haven't heard the benefit of better streamer with no DAC.
 


advertisement


Back
Top