Article: The Value Proposition In Computer Audio: Nuts, Bolts, and Building Blocks - Building a Home for Your Player Software Part 2

[bluesman]

View full article

[The Computer Audiophile]

Hi @bluesman, thanks for the awesome article. I love this stuff. It isn't for everybody of course, but for those of us that are into it, your articles are terrific. This one is no exception. 

 

Thanks for putting all the work in, so many others don't have to :~)

[davide256]

Thanks! Haven't played with an RPI4 yet, did you find the USB out improved vs the RPI3b+?  Played around with a variety of these devices about 18 months

ago but could never get USB out from the RPI3b+ to be as clean as a microRendu.

[The Computer Audiophile]

6 minutes ago, davide256 said:

but could never get USB out from the RPI3b+ to be as clean as a microRendu.

You're going to be playing with them for a very long time if you want that end result :~)

[Mortsnets]

I've read that the RPi4 USB has solved drop out problems, but sound quality on mine is not nearly as good as using my Allo DigiOne.

[bluesman]

1 hour ago, davide256 said:

Thanks! Haven't played with an RPI4 yet, did you find the USB out improved vs the RPI3b+?  Played around with a variety of these devices about 18 months

ago but could never get USB out from the RPI3b+ to be as clean as a microRendu.

If you used the stock Pi SMPS and powered your DAC through USB with any other USB peripheral connected, you weren’t getting all the 3B+ can give.

 

On basic program material (eg 2ch DSD or lighter, little or no DSP) that doesn’t overstress the 3, I don’t hear a difference between 3B+ and either USB bus on the 4 through DACs that are getting enough of their own clean power.  But when feeding DSD files from a USB HDD without its own PS while playing them back through a USB DAC using USB power, I think the 3‘s grainy and suspect it’s from borderline low voltage plus slight data flow interruptions.
 

To Chris’s point, I’ve never owned a microRendu and have only heard them a few times in other people’s systems. From memory, I think SQ from Roon Bridge on a 4 gig Pi feeding files of any size and complexity from NAS by gigabit Ethernet into an externally powered DAC via USB3 with no other USB peripherals connected is mighty close (if not equal) to what I remember of the mR.  I don’t recall and can’t find the CPU core count, the Ethernet spec or the various internal bus and R/W speeds for a mR, any of which may have contributed partly to this. Also, many mR buyers went for seriously over-the-top PSs because it was sold without one.  
 

 

[davide256]

24 minutes ago, bluesman said:

If you used the stock Pi SMPS and powered your DAC through USB with any other USB peripheral connected, you weren’t getting all the 3B+ can give.

 

On basic program material (eg 2ch DSD or lighter, little or no DSP) that doesn’t overstress the 3, I don’t hear a difference between 3B+ and either USB bus on the 4 through DACs that are getting enough of their own clean power.  But when feeding DSD files from a USB HDD without its own PS while playing them back through a USB DAC using USB power, I think the 3‘s grainy and suspect it’s from borderline low voltage plus slight data flow interruptions.
 

To Chris’s point, I’ve never owned a microRendu and have only heard them a few times in other people’s systems. From memory, I think SQ from Roon Bridge on a 4 gig Pi feeding files of any size and complexity from NAS by gigabit Ethernet into an externally powered DAC via USB3 with no other USB peripherals connected is mighty close (if not equal) to what I remember of the mR.  I don’t recall and can’t find the CPU core count, the Ethernet spec or the various internal bus and R/W speeds for a mR, any of which may have contributed partly to this. Also, many mR buyers went for seriously over-the-top PSs because it was sold without one.  
 

 

Was using the LPS 1.2 for 5V PS in my experimentation. Oddly enough the reason I stopped using a microRendu was that it didn't scale up with improved

power supply whereas a Pentium NUC did. The criticism of RPI before the RPI4 was Ethernet and USB shared the same bus, guess its high time to go back

and dabble again since the RPI4 doesn't do that. My wife has just been asking if I could setup a music source in the living room...

[bluesman]

1 hour ago, Mortsnets said:

I've read that the RPi4 USB has solved drop out problems, but sound quality on mine is not nearly as good as using my Allo DigiOne.

I assume you compared them into the same DAC and audio system powered by the same PS, which would be the only way I know of to compare them directly.  And I assume you’re comparing USB3 from the Pi to SPDIF from the Allo.  If the DAC had its own power for both, the only variable other than the Pi’s USB is the DAC’s USB processor(s) - and the DAC side of the USB pairing is most often ignored in these optical-vs-USB comparisons.

 

If you prefer SPDIF to USB, as many do, that’s fine and what you should use.  I’m still on the fence because I haven’t heard any truly direct comparisons myself except my Emotiva Stealth and my SMSL SU-8, both of which have optical, coax (on the Stealth), and USB inputs. The Emotiva has an older style C-media USB chip that’s simply not as good as its other digital inputs, and the difference is audible.  But the SMSL (v2) has much better USB circuitry that sounds little if any different to me from optical - I doubt that I could identify one from the other blinded.

 

Stay safe and enjoy the music!

[davide256]

25 minutes ago, bluesman said:

I assume you compared them into the same DAC and audio system powered by the same PS, which would be the only way I know of to compare them directly.  And I assume you’re comparing USB3 from the Pi to SPDIF from the Allo.  If the DAC had its own power for both, the only variable other than the Pi’s USB is the DAC’s USB processor(s) - and the DAC side of the USB pairing is most often ignored in these optical-vs-USB comparisons.

 

If you prefer SPDIF to USB, as many do, that’s fine and what you should use.  I’m still on the fence because I haven’t heard any truly direct comparisons myself except my Emotiva Stealth and my SMSL SU-8, both of which have optical, coax (on the Stealth), and USB inputs. The Emotiva has an older style C-media USB chip that’s simply not as good as its other digital inputs, and the difference is audible.  But the SMSL (v2) has much better USB circuitry that sounds little if any different to me from optical - I doubt that I could identify one from the other blinded.

 

Stay safe and enjoy the music!

There is a trade off using the Digione vs using USB out for the RPI3b+. The Digione has a sweet sound, few digital nasties but it didn't resolve transients they way

USB out did... but the USB out had digital irritants. I even "hacked" the PS input to the Digione so that it could be independently powered but it made no difference.

It's tempting to try the Signature Digione , better clocks and circuit board.

[bluesman]

34 minutes ago, davide256 said:

Was using the LPS 1.2 for 5V PS in my experimentation. Oddly enough the reason I stopped using a microRendu was that it didn't scale up with improved

power supply whereas a Pentium NUC did. The criticism of RPI before the RPI4 was Ethernet and USB shared the same bus, guess its high time to go back

and dabble again since the RPI4 doesn't do that. My wife has just been asking if I could setup a music source in the living room...

Very interesting!  The microRendu was sold without PS, and they offered 9 recommended options (from a $50 iPower to 4 figure monsters) with glowing descriptions of the sonic improvements to be had. I find these and similar quotes about power supplies for the mR and many other devices to be more than a little amusing (bold italics added by me):

• From the 2018 TAS review: "If it seems goofy that a power supply should cost nearly twice the price of the player, remember that the power supply contains much heavier components, is housed in a larger, fancier case, and is usually the primary factor in how a component sounds."
• From the PositiveFeedback review: "[The Channel Island Audio PS is] for those who are looking for a serious step up in sound quality, without totally bankrupting themselves...the VDC7 Mk. II offers a tremendous improvement in the overall presentation."

The PI 4 performs wonderfully for me - I now have one driving my living room system and one in my recording "studio".  I just sent Chris my next article, which is about easy mods to the Pi (none destructive and all well within the ability of anyone who can get a new one to play music).  It's amazing how you can improve them with a little cooling, an inexpensive externally powered USB SSD, a better microSD card than the one recommended and sold by the Pi Foundation, and a few simple command line entries.

[firedog]

Great write up. Looking forward  to  reading about your tweaks for the Pi.

I've got a 3B+ that was running PiCore player, b/c it was intstalled in a case with a touchscreen and I used it as a SBT replacement (much larger display) - here's a picture stolen from Archimago's blog:

 

If you want a touchscreen controlled device, the Pi in case with a touchscreen is a great solution and PiCore Player and LMS are a great software for it. LMS is still great server software and does pretty much anything an audiophile could wish for with it's various plugins. This system has an added DAC board ("HAT" in Pi parlance), as I use the DAC analog outs to a set of powered speakers. It's a wonderful solution for that type of setup. Basically everything in one small box that attaches directly to the speakers. 
 

I've since switched that endpoint to running Ropieee, which is a PI OS to turn a Pi into a Roon Endpoint. Works great with Roon, and can also function as a Squeezebox endpoint or an HQP NAA.

 

I've  had various endpoints in my main system, including an mR and a Bricasti M5. But today,in my main system I have a Pi4 with just 2GB of RAM (as I knew I'd just be using it for fairly light work) that functions as Roon Bridge endpoint via dietpi OS (it can also run HQP NAA and Squeezelite, among other programs). IMO, it's a fantastic Roon endpoint. Once you set it up it's a no intervention device. Mine is also in a "nice" milled aluminum case with passive cooling that looks "audiophile". Cost about $20 for the fancy case, and since the device doesn't do anything except act as an endpoint, it runs between 50C and 55C even in a warm room. It"s runnning 24/7 and "just works". I've streamed DSD 256 over it to a compatible DAC with no issues.

 

I have a fairly standard PS for it that does have a bit of noise filtering built in. I think it sounds really good in my setup and perfectly audiophile. At some point I'll compare it with and without an expensive PS to see if that makes a difference, and may go for a more "audiophile" product as an endpoint. I personally like the Sonore  family and the OS on their endpoints, so I may replace it with one of those. But I'm fully enjoying the SQ it is giving me right now. Bottom line: it's cheap, works really well, looks nice, and SQ is more than adequate, even on a good system. 

[bluesman]

57 minutes ago, davide256 said:

It's tempting to try the Signature Digione , better clocks and circuit board.

I'm always happy to try anything.  But since I retired, I've stopped paying for the privilege unless I want or need the item anyway, or I'm getting it at a price that will let me resell it without taking a bath.  Now that we're in an apartment, the shelves are so full that I'm even contemplating selling some of my old favorites because they're not getting used much.

 

As for the Signature DigiOne, I'd suggest reading the very revealing thread following the ASR review [sorry - I can't get the link to be hot: https://www.audio “science” review/forum/index.php?threads/review-measurements-of-allo-digione-signature-and-diyinhk-pro3z.4660/] before putting out the money for both the device and the 2 power supplies it needs. The wide diversity of opinion suggests to me that it's not a slam dunk improvement - if it were, more people would hear it than not.   I'm absolutely not an ASR disciple, but I find some of the reviews and subsequent posts to be as entertaining and amusing as most current sitcoms.  A quote of note in the linked thread comes from ASR's leader Amir:  "if you have a high-quality DAC, there is no reason to anguish over how you are driving it."  I have no idea what he really means by that, since he doesn't say that he hears no difference among the many alternatives.  But I can't think of much in audio or most other areas that's truly worthy of anguish.  Some truly boneheaded actions deserve a little regret - but I strongly suggest reserving anguish for the seriously major stuff in life.

 

Even the Allo web page about it is confusing.  For example, it says clearly and specifically that "You can use anything on [the "clean" side], linear power supply or pure batteries power". But without including a SMPS on that list, they apparently don't really mean "anything".

[firedog]

30 minutes ago, bluesman said:

A quote of note in the linked thread comes from ASR's leader Amir:  "if you have a high-quality DAC, there is no reason to anguish over how you are driving it."  I have no idea what he really means by that, since he doesn't say that he hears no difference among the many alternatives.

Based on various comments he's made, he seems to think that with a "well engineered" (my phrase) DAC the cheaper and more expensive source devices will sound the same - providing that they - like the Signature DigiOne - measure well: low noise, low jitter. IOW, get a good DAC (which may only cost a $few hundred) and all properly engineered source devices will sound the same. 
 

Note that in another comment to the review you mentioned the Allo people themselves say that the signature DigiOne was engineered to measure as well as anything on the market. But also say this:

Quote

Now , test data , clearly shows that THD+N numbers with a good dac (with good jitter reduction) will not change with a poor/good digital source. On the other hand , you have multiple subjective opinions that a better source produce a better sound (I want to stress that no test data shows it)

 

[plissken]

1 hour ago, bluesman said:

A quote of note in the linked thread comes from ASR's leader Amir:  "if you have a high-quality DAC, there is no reason to anguish over how you are driving it."  I have no idea what he really means by that,

 

Unfortunately only partially quoting and dropping it here means that all context is lost.

 

Here's the full quote:

As we see here, there is no difference at all how you drive the S/PDIF input on the Exasound E32. It excellently cleans up whatever ills may be on S/PDIF and renders the same excellent output. So if you have a high-quality DAC, there is no reason to anguish over how you are driving it.

 

So Amir is using the Exasound 32 as an engineering example of what you get when designers take care to sanitize their inputs from typical upstream issues. Also Benchmark did this with one of their DAC's and I believe a 100 foot cable that was just ludicrously out of spec, also former member Mansr midwifed a USB setup that was out of spec but showed 100% reconstruction by the DAC regardless.

 

Great article BTW. And there is now an 8GB of RAM Pi but we have to wait for a 64bit OS for it.

 

 

 

[The Computer Audiophile]

1 minute ago, plissken said:

 

Unfortunately only partially quoting and dropping it here means that all context is lost.

 

Here's the full quote:

As we see here, there is no difference at all how you drive the S/PDIF input on the Exasound E32. It excellently cleans up whatever ills may be on S/PDIF and renders the same excellent output. So if you have a high-quality DAC, there is no reason to anguish over how you are driving it.

 

So Amir is using the Exasound 32 as an engineering example of what you get when designers take care to sanitize their inputs from typical upstream issues. Also Benchmark did this with one of their DAC's and I believe a 100 foot cable that was just ludicrously out of spec, also former member Mansr midwifed a USB setup that was out of spec but showed 100% reconstruction by the DAC regardless.

 

Great article BTW. And there is now an 8GB of RAM Pi but we have to wait for a 64bit OS for it.

Thankfully we were rescued by one of Amir's disciples. 

 

The context there added nothing and in fact was nothing but a side note because it involved a subset of his all encompassing statement, "So if you have a high-quality DAC, there is no reason to anguish over how you are driving it."

[plissken]

9 minutes ago, The Computer Audiophile said:

Thankfully we were rescued by one of Amir's disciples. 

 

The context there added nothing and in fact was nothing but a side note because it involved a subset of his all encompassing statement, "So if you have a high-quality DAC, there is no reason to anguish over how you are driving it."

 

If you would only grab the rope Chris.

 

Also figured out why your author couldn't get the hotlink to work

 

https://www.roonlabs.com works as hotlink

 

https://www.audio “science” review does work.   Well at least until it doesn't 😉 

 

www.audio “science” review doesn't work,

 

So the automation here is messing with the hot link to a specific domain.

[The Computer Audiophile]

6 minutes ago, plissken said:

 

If you would only grab the rope Chris.

 

Also figured out why your author couldn't get the hotlink to work

 

https://www.roonlabs.com works as hotlink

 

https://www.audio “science” review does work.   Well at least until it doesn't 😉 

 

www.audio “science” review doesn't work,

 

So the automation here is messing with the hot link to a specific domain.

Absolutely. One of Amir's other disciples created topics here specifically to bad mouth AS and attempt to get people to go to ASR. Thus, I created a rule to stop the links from working. I also put the word science in quotes because it's only fitting. 

[plissken]

22 minutes ago, The Computer Audiophile said:

Thankfully we were rescued by one of Amir's disciples.

 

Also let me know if the name calling only goes one way between you and me. AKA what kind of problem do we personally have?

[The Computer Audiophile]

Just now, plissken said:

 

Also let me know if the name calling only goes one way between you and me. AKA what kind of problem do we personally have?

Not a name, just a description of your behavior. 

[plissken]

4 minutes ago, The Computer Audiophile said:

Not a name, just a description of your behavior. 

 

Bluesman had a question about a quote. Both Firedog and I answered along the same lines. Failing to see the mountain made out of a molehill.

[The Computer Audiophile]

Just now, plissken said:

 

Bluesman had a question about a quote. Both Firedog and I answered along the same lines.

But you have a long history of disruption. 

[plissken]

Better this moves over and you PM me if you feel that strongly as to reduce the amount of injected noise.

[mfsoa]

Ahh, the old "If you think you think a cable or power supply made a difference then that proves that your equipment is poorly designed or malfunctioning..."

 

It must get old someday, hopefully soon.

 

Sorry for the OT...

[The Computer Audiophile]

13 minutes ago, plissken said:

Better this moves over and you PM me if you feel that strongly as to reduce the amount of injected noise.

Best thing I've read all day. 

 

 

[bluesman]

58 minutes ago, The Computer Audiophile said:

Absolutely. One of Amir's other disciples created topics here specifically to bad mouth AS and attempt to get people to go to ASR. Thus, I created a rule to stop the links from working. I also put the word science in quotes because it's only fitting. 

I was unaware of both this fact and everything associated with it, so I apologize for starting an errant topic.  But I suspect that no one will fail to understand my motivation for that link once you see the comments in the thread following the review. 

 

There are so many divergent opinions about the same device and the same issue (USB vs SPDIF) that I have no idea whether either alternative in either case (Signature vs "unsigned" microRendu and USB vs SPDIF) is actually different from the other in any way at all.  And so it is with the Raspberry Pi 4 vs other audio devices in the same role.  Either it's your slice of Pi or it's not.

 

I really got a kick out of the comments in that thread.  Here's just a tiny sample of the wacky world of internet audio comedy:

 

• "with a good clock coax is typically better than USB because of less noise"
• "SPDIF is just generally inferior to USB. So if you are hearing a difference it is just with additional distortion from the inferior SPDIF interface which is a clockless mess of light."
• "Thats wrong on so many levels its almost scary, I hardly know where to begin"
• "There is no way it can improve the sound on an rme adi. Likely it is making the sound much worse but you prefer the distortion."
• And there's this absolutely priceless gem:  "Ditch your shit motherboard instead of wasting money on nonsense audio products."

 

You could not make this stuff up if you tried!