What is the theory behind why using an NAA with HQP improves the sound?

[tboooe]

I've been using HQP for a few months now and love it. Now I am toying with the idea of adding an NAA to my system but I am struggling with trying to understand why it would improve my situation. Now, my pc running HQP and Roon server is connected to my DAC via USB. Since I am upsampling to DSD the NAA will also have to use USB. Am I right to assume that in order for me to get an improvement, the NAA must have a superior USB implementation than the one on my pc mobo? Or, is there some other benefit to using an NAA? I've been reading that some people (even Miska) are using a Raspberry Pi as an NAA which I cant believe has a better USB implementation than a pc mobo.

 

Anyway, I appreciate any insight on this, at least in theory.

[davide256]

The USB implementation on the Pi is known not to be good... I had to add a Hifiberry+ for SPDIF out before I could find the Pi 2 musical at all.

[Le Concombre Masqué]

theory is to disrupt transmission of noise to the DAC ; the NAA doing almost nothing is less noisy than a directly connected heavy working computer.

I've been using HQP for a few months now and love it. Now I am toying with the idea of adding an NAA to my system but I am struggling with trying to understand why it would improve my situation. Now, my pc running HQP and Roon server is connected to my DAC via USB. Since I am upsampling to DSD the NAA will also have to use USB. Am I right to assume that in order for me to get an improvement, the NAA must have a superior USB implementation than the one on my pc mobo? Or, is there some other benefit to using an NAA? I've been reading that some people (even Miska) are using a Raspberry Pi as an NAA which I cant believe has a better USB implementation than a pc mobo.

 

Anyway, I appreciate any insight on this, at least in theory.

[feelingears]

As the OP wrote, it's not exactly clear why an NAA (such as the Sonic Transporter) is needed or even helps. (Sure, I read the AudioStream review recently, but that doesn't quite answer the question since Langvora didn't use HQPlayer despite mentioning it in one paragraph.)

 

I understand we're trying to limit the noise (internal and external) our devices are generating and receiving, but I don't quite understand why an NAA is any better in a setup with HQPlayer if you need a "full" machine to do the number-crunching HQPlayer desires/requires. The OP's setup seems to limit the devices (and power supplies) and perhaps only needs a good USB noise-reducing cable to achieve its best sound.

 

Related to the OP's question is: If HQP is not present, then the music file source is the device sending data over the network, and could be NAS or a PC. If HQP is present, then it's a likely a PC. My (odd?) experience is that I used an Ethernet network isolator to isolate my source (NAS--see sig below) and it did some kind of magic. I'm pretty pleased and so far think it's "all good" although something is either not quite right or just different enough to make me second guess the overall improvement. So that said, I'm also wondering if a NAA would help, but what does it do better that my NAS (or a PC running HQP) doesn't do?

(And, I realize that an NAS isn't considered a high-quality sonic source, but in a brief test--really brief--I thought it sounded better than my PC running the same LMS server software.)

 

Any clarifications would be appreciated!

[davide256]

I suspect that when the OP does try a network attached renderer he will find quite an improvement in tone color richness at a possible trade off of transient detail. I've always found the midrange more correct for tone colors when file source and renderer are separate hardware. I don't see that HQP NAA offers any advantage over MPD/DLNA when I compare HQP and Foobar2000 through microRendu.

[mikey8811]

Aren't you already using an NAA in the form of the microRendu?

 

As the OP wrote, it's not exactly clear why an NAA (such as the Sonic Transporter) is needed or even helps. (Sure, I read the AudioStream review recently, but that doesn't quite answer the question since Langvora didn't use HQPlayer despite mentioning it in one paragraph.)

 

I understand we're trying to limit the noise (internal and external) our devices are generating and receiving, but I don't quite understand why an NAA is any better in a setup with HQPlayer if you need a "full" machine to do the number-crunching HQPlayer desires/requires. The OP's setup seems to limit the devices (and power supplies) and perhaps only needs a good USB noise-reducing cable to achieve its best sound.

 

Related to the OP's question is: If HQP is not present, then the music file source is the device sending data over the network, and could be NAS or a PC. If HQP is present, then it's a likely a PC. My (odd?) experience is that I used an Ethernet network isolator to isolate my source (NAS--see sig below) and it did some kind of magic. I'm pretty pleased and so far think it's "all good" although something is either not quite right or just different enough to make me second guess the overall improvement. So that said, I'm also wondering if a NAA would help, but what does it do better that my NAS (or a PC running HQP) doesn't do?

(And, I realize that an NAS isn't considered a high-quality sonic source, but in a brief test--really brief--I thought it sounded better than my PC running the same LMS server software.)

 

Any clarifications would be appreciated!

[tboooe]

I don't see that HQP NAA offers any advantage over MPD/DLNA when I compare HQP and Foobar2000 through microRendu.

So are you saying here that using an NAA running on perhaps something like a Pi3 sounds as good as using a microRendu (assuming HQP is running on a server PC)? If so that is great to hear since a Pi3 is much less expensive than a microrendu.

 

So as I understand, the benefit of an NAA is to eliminate the noise that is being generated by the server PC that is doing cpu intensive processes? But isnt this noise transmitted via ethernet to the NAA?

[Miska]

So as I understand, the benefit of an NAA is to eliminate the noise that is being generated by the server PC that is doing cpu intensive processes? But isnt this noise transmitted via ethernet to the NAA?

 

USB in most cases is not isolated at all and has direct galvanic connection between host and the device.

 

While ethernet by definition is transformer-isolated at all connecting ends. In addition, using ethernet or other IP capable networking gear also allows using optical interconnects further isolating any noise. It also allows longer cabling, so HQPlayer computer can reside outside of the listening room to easily deal with possible acoustic noises (fans) of powerful computers.

 

In addition IP networking provides one-to-many and many-to-one connectivity USB lacks. So you can either use same DAC from multiple computers, or multiple DACs (around the house) from the same computer.

[davide256]

So are you saying here that using an NAA running on perhaps something like a Pi3 sounds as good as using a microRendu (assuming HQP is running on a server PC)? If so that is great to hear since a Pi3 is much less expensive than a microrendu.

 

So as I understand, the benefit of an NAA is to eliminate the noise that is being generated by the server PC that is doing cpu intensive processes? But isnt this noise transmitted via ethernet to the NAA?

 

Mixing apples and oranges. The microRendu supports 4 main types of network software interface to it's renderer hardware, Squeezelite, MPD/DLNA, HQP NAA and Shairport (for apple devices).The primary function of the microRendu is to render; it takes a file transmitted over the network by solution of choice and processes it into digital signal output to USB. Comparing a Pi to the microRendu as renderer hardware is like comparing a Ford Escort to a BMW sedan for performance.

 

 

When data is sent over a network, source and renderer have no direct connection... there has to be a switch or router in between. So whatever is happening at the source can't "ghost" electrical interference directly to your USB output as happens in a PC providing USB output. You can still have electrical interference from the router itself but that will be random vs synchronized interference ( say someone sharing the router kicks off a 1000mbps file xfer in the middle of your listening and the router is less than electrically or bandwidth wise well behaved under the load)

[feelingears]

Aren't you already using an NAA in the form of the microRendu?

 

Oops, I may have terminology (and minimum requirements) mixed up here (in part due to the AudioStream article I referenced). I interpreted the OP's question in part as "would adding a device such as the microRendu or a Sonic Transporter, or both, significantly improve the sound quality."

 

I'm intrigued by @davide256 's comment about transient detail vs. tonal color. I'd say that in my short experience with computer networked audio, this is very much the case. Sometimes tones improve at the cost of other tones, sometimes at the cost of what I call acoustic resolution (the attack and decay of sound--is that transient detail?).

[YashN]

Sometimes tones improve at the cost of other tones, sometimes at the cost of what I call acoustic resolution (the attack and decay of sound--is that transient detail?).

 

Yep, except I'd say the decay can be a long one, so in that case, it's not that part which is 'transient', but perhaps the onset of it. The end of the decay is quite important too though: if reverb tails are cut off, you get an artificially smaller soundstage or ambiance.

 

As for the attack transient, I have come to the conclusion it is of supreme importance for a lot of things, including timbre resolution and recognition as well as spatial/localisation effects / soundstage / ambiance.

[YashN]

Mixing apples and oranges. The microRendu supports 4 main types of network software interface to it's renderer hardware, Squeezelite, MPD/DLNA, HQP NAA and Shairport (for apple devices).The primary function of the microRendu is to render; it takes a file transmitted over the network by solution of choice and processes it into digital signal output to USB. Comparing a Pi to the microRendu as renderer hardware is like comparing a Ford Escort to a BMW sedan for performance.

 

The primary point of comparison should be that the Pi is a general-purpose computer and that the microRendu is a purpose-built device (all hardware, all software designed specifically for maximum performance for audiophile use). Since the Raspi is a general purpose device, it should have a larger noise profile than the microRendu.