System architecture - "traditional" vs computer-centric

[klettermann]

Again, the more I learn the less I know. I’ve been at this since college ('70s) and taken a fairly conventional approach. Originally that meant turntable, preamp/integrated amp, speakers. Plug it in, set it up properly, enjoy music. Digital expanded that model somewhat, but still plug-and-play audio appliances along the lines of Case 1 below. To that we can add some kind of DSP in many cases.

1. Conventional component-style system
Discrete sources — streamer / CD transport / turntable
→ standard Ethernet or direct digital link (S/PDIF, AES/EBU, sometimes USB)
→ DAC (or integrated DAC/amp)
→ preamp / integrated amp
→ power amp (if separate)
→ speakers + subwoofer(s)

Recent curiosity has kind of broken my mold as I've become aware of a different and parallel approach that's computer-centric, along the lines of Case 2. This includes devices and optimization layers I didn't know existed. I honestly don’t know how representative this is here, but I see versions of it often enough that it seems common, or at least aspirational. What the hell is all this stuff and is it an improvement or just an exercise in complexity?

2. Computer-centric system
Music server / streaming service
→ Mac Mini / NUC / PC running Roon or HQPlayer
→ network optimization devices / specialty switch / etc
→ endpoint computer (NUC / Raspberry Pi / bridge)
→ optional linear PSU / optical isolation / USB accessories
→ DAC
→ preamp
→ power amp
→ speakers + subwoofer(s)

Maybe somebody familiar with both approaches could contrast them, or correct my framing if it’s off? What are the pros/cons of approach 2? Better sound quality? Greater tweakability? An endless path of refinement? Convenience tradeoffs? I'm sure there are things many I don’t know, and at the same time things I’ve never need to know. Comment welcome and thanks! Cheers.

 

[Oddball]

Looks like you are eager to explore a journey. Which is great if that is what will make you happy.

I can tell you with high level of certainty is that solutions are already there. It is not really PC based, but could be as well soon (when PC gets ART later this year).

Nothing like ART to control the multiple subs that a good system should have. Trinnov has some tricks, but much more expensive and esoteric, so not of interest to most.

After ART range (150hz), the most effective would be cardioid speakers as they tend to behave better in the room. There are of course room treatments as well, and some good behaving passive speakers.

Also, not every room is the same, so this is just a bit of general guidance for medium to large sized rooms.

I am amazed how little people care about ART on this forum, but I guess they just need time as currently mostly available on AVRs which most certified members don't have or want.

 

[Kal Rubinson]

2. Computer-centric system
Music server / streaming service
→ Mac Mini / NUC / PC running Roon or HQPlayer
→ network optimization devices / specialty switch / etc
→ endpoint computer (NUC / Raspberry Pi / bridge)
→ optional linear PSU / optical isolation / USB accessories
→ DAC
→ preamp
→ power amp
→ speakers + subwoofer(s)

My system is simpler.

→ Mac Mini / NUC / PC running Roon or HQPlayer and/or Jriver (running Dirac-ART as VST plug-in)
→ network optimization devices / specialty switch / etc *
→ endpoint computer (NUC / Raspberry Pi / bridge) *
→ optional linear PSU / optical isolation / USB accessories *
→ networked DAC / preamp
→ preamp
→ power amp + active speakers
→ speakers + subwoofer(s)

Items marked * are only necessary if there are local conditions that require them.

 

[john61ct]

Really there is no "A vs B" here, you pick and choose how to use modern tech to get the SQ and features you want at prices you can afford.

Unless you want to "go extreme" in either direction, which is silly - like being religious IMO.

User ad-hoc tone controls / loudness contour are not to be shunned by rational people, digital implementations are likely to be superior to the OG gear.

Speakers remain the most important factor, the most "matter of taste" out of all the rest. You buy clean amplification with enough power to match your speaker choice.

Digital sources are better than analog, but

playing from a general-purpose PC is just one choice, as is using Raspberry Pi or a purpose-built streamer, network player + NAS file library, CD player whatever.

Even playing from LPs, such devices can be used for "pass through" DSP, designing your own bass management crossovers, getting arrival times tuned adds great soundstage / transients clarity

Room correction and per-driver EQ requires DSP and makes a huge improvement, maybe more than speaker choice, depending on the effects of your ROOM.

Immersive multichannel surround, whether upsampling or decoding the proprietary Atmos / DTS etc formats, can bring a whole other dimension to music not just films.

These are the big decisions, tech details like gain stage matching, balanced vs SE, TOSlink vs Coax S/PDIF are details that can be figured out as you go rather than driving product selection.

Modern gear that is reco'd here has no problem getting converted between analog and digital, even several times back and forth will not cause any audible changes.

The hardest part is not wasting money on stuff that doesn't actually make a real audible difference.

 

[Roland68]

Again, the more I learn the less I know. I’ve been at this since college ('70s) and taken a fairly conventional approach. Originally that meant turntable, preamp/integrated amp, speakers. Plug it in, set it up properly, enjoy music. Digital expanded that model somewhat, but still plug-and-play audio appliances along the lines of Case 1 below. To that we can add some kind of DSP in many cases.

1. Conventional component-style system
Discrete sources — streamer / CD transport / turntable
→ standard Ethernet or direct digital link (S/PDIF, AES/EBU, sometimes USB)
→ DAC (or integrated DAC/amp)
→ preamp / integrated amp
→ power amp (if separate)
→ speakers + subwoofer(s)

Recent curiosity has kind of broken my mold as I've become aware of a different and parallel approach that's computer-centric, along the lines of Case 2. This includes devices and optimization layers I didn't know existed. I honestly don’t know how representative this is here, but I see versions of it often enough that it seems common, or at least aspirational. What the hell is all this stuff and is it an improvement or just an exercise in complexity?

2. Computer-centric system
Music server / streaming service
→ Mac Mini / NUC / PC running Roon or HQPlayer
→ network optimization devices / specialty switch / etc
→ endpoint computer (NUC / Raspberry Pi / bridge)
→ optional linear PSU / optical isolation / USB accessories
→ DAC
→ preamp
→ power amp
→ speakers + subwoofer(s)

Maybe somebody familiar with both approaches could contrast them, or correct my framing if it’s off? What are the pros/cons of approach 2? Better sound quality? Greater tweakability? An endless path of refinement? Convenience tradeoffs? I'm sure there are things many I don’t know, and at the same time things I’ve never need to know. Comment welcome and thanks! Cheers.

Based on numerous tests conducted over the last 25 years, including various network solutions and even those using Dante, I can tell you that you won't get better sound quality than a directly played CD or a high-resolution FLAC file.

None of these components can improve the sound, because your source file/medium is already at 100% quality. How could you possibly improve that data?
That's just marketing hype for snake oil.

So choose the components that are most convenient and practical for you to use for listening to music, and use as few of them as possible. Anything else is a waste of money and nonsense.

 

[john61ct]

High-res formats from streaming services assuming good underlying network connections should be as good as local CDs / files wrt audible differences.

OP note that some people enjoy the sound of "controlled distortion" of say tube (pre-) amplification, calling it "warmth" etc

But the ethos here is transparency / fidelity / accuracy, so "just don't screw up the signal" is indeed the fundamental guideline.

 

[klettermann]

To be clear, I think my system is pretty damn good and I've gone to pains with room treatment, sub integration, proper EQing, etc. I can't really think of how it could be reasonably improved so i just listen to the music.

At the same time there seem to be lots of folks out there doing optical Ethernet, USB purification, wild over-sampling and God knows what else. All that stuff costs real money that real people are willing to spend for some perceived gain. Or the hope of a gain? Or the belief that many small improvements add up to 1 big improvement? Or maybe it isn't really even about SQ at all, but rather the pursuit of endless micro-improvements? I have no idea. So my curiosity is making me nuts. Hope that made sense.

 

[john61ct]

To be clear, I think my system is pretty damn good and I've gone to pains with room treatment, sub integration, proper EQing, etc. I can't really think of how it could be reasonably improved so i just listen to the music.

Then you can rest easy, sorry I assumed your questions were sincere.

But educate yourself if you DO decide it's worth testing out some of the areas I mentioned.

Even if 95% of the "audio nervosa" some audiophools engage in, is a total waste of their money and time

there are topic domains I mentioned that can make for astounding improvements in listener enjoyment.

If that listener keeps an open mind.

And you can avoid the phoolery by reading threads here and posting questions.

 

[klettermann]

Then you can rest easy, sorry I assumed your questions were sincere.

But educate yourself if you DO decide it's worth testing out some of the areas I mentioned.

I absolutely am sincere in trying to understand the computer audio thing. I struggle to just dismiss it as audiophoolery when so many are apparently devoted to it. I'm coming to the view that it's really 2 different hobbies.

PS - I just realized that I already posted sometime along these lines. But in the meantime I keep getting astounded.

 

[Kal Rubinson]

To be clear, I think my system is pretty damn good and I've gone to pains with room treatment, sub integration, proper EQing, etc. I can't really think of how it could be reasonably improved so i just listen to the music.

At the same time there seem to be lots of folks out there doing optical Ethernet, USB purification, wild over-sampling and God knows what else. All that stuff costs real money that real people are willing to spend for some perceived gain. Or the hope of a gain? Or the belief that many small improvements add up to 1 big improvement? Or maybe it isn't really even about SQ at all, but rather the pursuit of endless micro-improvements? I have no idea. So my curiosity is making me nuts. Hope that made sense.

But you were not clear. You proposed distinguishing between "conventional" and computer-centric systems when, in fact, you seem to be wondering about the need for and potential value of tweaking and gadgetry. In practice, both your systems are "computer-centric" although type 1 embeds the computer in a streamer / CD transport and the other is explicit.

 

[john61ct]

No it's not a separate hobby, as I tried to explain it isn't any one category as you think, it's a greyscale continuum of preferences, that can 100% be based on objective rationality.

The recent decades' advances in pure delivery of high SQ, and especially in price / performance have been simply astounding.

For example, I have no interest in using general purpose computers as network renderer / streaming players.

However something like Wiim Ultra or the WU Amp sure deliver a great value, but few would lump them in (or streaming in general) as part of "the computer audio thing".

Using chips in components is a grey area for you I guess? but many functions now considered indispensable do require them.

What do you think of active analog crossovers as opposed to passive networks? Those have been around a while now 8-D

 

[Mr. Widget]

My system is simpler.

→ Mac Mini / NUC / PC running Roon or HQPlayer and/or Jriver (running Dirac-ART as VST plug-in)
→ network optimization devices / specialty switch / etc *
→ endpoint computer (NUC / Raspberry Pi / bridge) *
→ optional linear PSU / optical isolation / USB accessories *
→ networked DAC / preamp
→ preamp
→ power amp + active speakers
→ speakers + subwoofer(s)

Items marked * are only necessary if there are local conditions that require them.

My current system is even simpler and no Dirac-ART necessary.

Home network with a NAS.
Roon Nucleus+ (essentially a NUC) → D&D8c speakers.

 

[CD2]

My current system is even simpler and no Dirac-ART necessary.

Home network with a NAS.
Roon Nucleus+ (essentially a NUC) → D&D8c speakers.

Yep. At this point, you can get a pair of good active speakers, get them talking to your home network, and you're done. The remaining tweaking is gonna be room correction.

 

[levimax]

PC based systems are hard to beat but have one major issue which is there is really no practical way to play Dolby Atmos music without a Dolby receiver at which point why bother with a PC.

 

[JiiPee]

The OPs first scenario of conventional system contained already a significant evolutionary step: Introduction of CD format, which meant that the distribution of music could be done with a medium where the information about music was coded digitally. This raised the limit of sound quality that a consumer could reasonably get, because both the vinyl record and magnetic tape analogue formats have physical limitations of what is possible to achieve in practice.

While the CD brought digitally coded music information, the conventional system architecture converted it immediately to analogue domain so that the rest of the music reproduction system remained just the same as before with analogue sources. This meant that many of the advantages of handling music reproduction in digital domain were still unused.

The move from conventional system to "computer-centric system" is really about distributing the music information via commercial data-network instead of using physical media (like CD), and keeping the music information in digital domain for as long as possible. On this road we have already come so far that the current state of the art systems do even the speaker crossovers in digital domain before the conversion to analogue domain needed for speaker drivers. There are several significant advantages in this approach:
- Distribution via network means that consumers can have cheap access to vast amount of music performances.
- Handling the music information in digital domain can be done with general purpose micro computers which are available at relatively cheap prices.
- Digitally coded music information is much more robust against electric interference that could cause audible distortion in analogue signal.
- Adjusting the sound to listening room acoustics and individual preferences is both cheaper and easier, which means that a larger proportion of end users can benefit from it in practise.
- Having digital crossovers and dedicated amplifiers for each driver gives new opportunities in speaker design for better performance
- The user interface is easier for many end users (especially compared to the delicate handling of vinyl players and records)
- etc...

 

[PristineSound]

This question is kind of like asking should I eat Pizza or Burgers for lunch. Or which one is better, pizza or burger. The answer is, it depends.

 

[Keith_W]

PC based systems are hard to beat but have one major issue which is there is really no practical way to play Dolby Atmos music without a Dolby receiver at which point why bother with a PC.

I look at it the other way round. If Dolby can't be bothered to cater to us, why should I go out of my way to consume their product.

 

[levimax]

I look at it the other way round. If Dolby can't be bothered to cater to us, why should I go out of my way to consume their product.

I am no fan of Dolby, just a successful version of MQA collecting a tax on a lot of music and hardware without doing much. If there was ever a place for an open standard it seems like it should be multi channel music but Dolby has apparently won the war.

 

[Keith_W]

I am no fan of Dolby, just a successful version of MQA collecting a tax on a lot of music and hardware without doing much.

I agree with the second half of your statement, but Atmos is a much more impressive technology than MQA. MQA was quite obviously a scam. Atmos is something I might be mildly interested in for music, and very interested in if I was into movies. Which I fortunately am not.

If there was ever a place for an open standard it seems like it should be multi channel music but Dolby has apparently won the war.

There is! There's IAMF. ASR thread here and here.

 

[antcollinet]

At the same time there seem to be lots of folks out there doing optical Ethernet, USB purification, wild over-sampling and God knows what else.

None of those things make any improvement to the sound.

One exception is if you get a noisy ground loop - can sometimes happen especially with PC's in the system. Then you need isolation between PC (if PC is the cause) and DAC. This can be done easily and cheaply- the cheapest being an optical TOSLINK connection. (Around $£EUR 25 if you don't have an optical output from your PC already)

Special clocks, audio network switches, linear power supplies - etc etc and so on ad-infinitum - are all useless snake oil if sold on the basis of improved sound.

Even Hi res audio (Higher res than CD) will not make any audible difference from the format - though high res files are sometimes made from a different/better master with less "loudness" compression.