Saturday, 4 March 2017

MUSINGS: Computer Audio Part I: Demystifying "Computer Audio Demystified"


It's nice to see that some of the seminars at the Rocky Mountain Audio Fest (RMAF) 2016 have been posted on YouTube. It gives all of us who may have wanted to visit a chance to review the "latest and greatest" tips and updates...

Although the RMAF is hosted by the Colorado Audio Society, remember that much of the material comes from Industry; with of course the potential for vested interests influencing the content. I think many visitors to this blog would be interested in computer audio, so let's spend some time looking at the information disseminated... For your consideration, let us explore the one presented by Steve Silberman of AudioQuest - "Computer Audio Demystified" (RMAF 2016):



Feel free to watch the full 1.5 hours. For today's segment, we'll look at a bit of reality testing vs. myth vs. mysticism. You'll notice that the comments section for this video on YouTube has been turned off. Interesting. How else to discuss the contents then than in a blog post like this, right? :-)

Over the years, I have attended a number of audio store "meet and greets", demos, openings and a couple of the recent Vancouver Audio Shows (at least those are the ones I've posted on). Inevitably one is treated to talks like this from the folks in the industry... In fact, back in 2015, I attended one of these AudioQuest talks and mentioned in the show report their Jitterbug device being demonstrated along with similar suggestions and comments shown in the video.

A number of times while watching the video above, the word "infomercial" did go through my mind, partly because that word was used in the video in a self-referential way (as in "I hope this doesn't sound like an infomercial"). Furthermore, throughout the presentation, it's good to see that Mr. Silberman was acutely aware of the controversial statements being made and able to anticipate "flames" and "hear forums in the distance exploding" from his recommendations (32:30). Apparently there was "blowback" about his presentation last year on YouTube; I have not seen that one but if the content is similar, it would not be surprising.

Realize that I'm not saying the content in this presentation is all bad of course. There are good recommendations around backing up one's music collection. Good discussions for folks who want to stream audio (TIDAL), mention of playback options (JRiver software, various hardware options like Aurender, even microRendu), and discussions around the potential value of metadata services (like Roon). Even some nice tracks for demoing - Rob Wasserman & Jennifer Warnes' "Ballad of The Runaway Horse" (from Duets, 1988, DR13) and John Mayer's "Born and Raised" (from Born and Raised, 2012, DR9) being the first two.

However, I must admit that it does get a bit frustrating seeing that in the supposed process of "demystifying" computer audio, some of the the recommendations and discussions actually seem to create confusion and mystery!

So then, for the purpose of hopefully demystifying the "demystification", let us go over some of the material presented and see if we can discuss some alternate points-of-view; aiming to base claims on reason and point to experiments to provide evidence for these alternate views if they exist. Without further ado, in a semi-chronological order, my "top 10" rebuttal of observations and comments I thought were presented in a biased fashion, perpetuation of myth, erroneous or just plain mystical:



1. DEMOS SUPPOSEDLY SHOWING ALL KINDS OF AUDIBLE DIFFERENCES ARE DONE IN THESE TALKS.

Do these sighted "demos" in audio shows prove anything about audibility? No, not IMO... At the start of this presentation (around 7:00), we see the "test" being done between a presumably old Core 2 Duo CPU computer with 1GB RAM, spinning HD, Windows-JRiver laptop vs. new quad core MacBook OS X-Roon with 16GB RAM and SSD.

Just imagine the test set-up here for the attendees. Likely unfamiliar music, speakers suboptimally placed (tweeter level way too elevated on the table!), unfamiliar and acoustically suboptimal room, a bunch of people (around 50?) sitting around making noise, probably the HVAC going in the room, and most folks sitting outside the sweet spot... It also took 40 seconds to switch between inputs. The second playback was much shorter in duration. Someone says it sounded "very different". We're told it's the same DAC afterwards, later on an astute audience member mentions the second playback seemed louder (11:50) and the answer was "when you lower the noise floor there is the perception of loudness" ("someone in the forum will flame me for that" :-).

You know, the way I see it is this. Folks who pay money to attend these audio shows are not there to start a fight. We go because it's nice to check out the new stuff from manufacturers, enjoy the music, hang out with some fellow hobbyists, and marvel at the toys. And I believe that for the most part, attendees know that there's a lot of pseudoscience and fluff being pushed. The people I've met at Industry-sponsored events seem to be agreeable folks. Even the most "hard core" audiophile I've met seem self-conscious and in a deeper discussion typically will agree that they have doubts about much of what is being presented by the Industry and press. As such I don't believe you'll ever have a demo in these shows where audiophiles will not be agreeable with the presenter and anyone with a dissenting view likely will just keep his/her mouth shut and allow the presentation to move along. That's certainly the way I felt in 2015 at the AudioQuest presentation and looking around during demos that day, it's not like there was ever unanimous consent that cable switches and tweaks resulted in such enthusiasm that the audience was about to give a standing ovation to some "revolutionary" change in sound quality nor folks lining up with credit cards in hand.

Notice what we are not told nor did anyone ask in this demo. Were the computers set up in a bit-perfect fashion? This could explain the volume differences. Imagine if the JRiver set-up wasn't using ASIO/WASAPI and the Windows volume control was still activated (<100%) and going through the Windows Mixer compared to the MacBook using Roon and streaming bit-perfectly to the DAC. Folks should keep this in mind the next time attending a show like this and wondering about volume differences.

As for the comment about lower noise floor and "perception of loudness". All I can say is - prove it. This is easily demonstrated by measuring the noise floor and showing us the difference especially if it's that obvious to be audible in such a suboptimal playback environment! Notice how none of these presenters ever provide proof of theories like this. I can say that I have never experienced such a thing over the years using a good quality DAC neither in subjective listening nor when measuring the output. Where there is interference from an extremely electrically noisy machine (purposely demonstrated here where I used an unshielded RCA cable instead of a proper coaxial), the anomaly in the noise floor can be easily shown even if not obvious with listening.

Related to these kind of demos, notice the poll question used at 3:45 "Who thinks all computers sound the same?" And then he expects people to raise their hands... Look, if I were in the audience, I wouldn't raise my hand either! The real question is: "Who thinks all computers sending the same bit-perfect digital audio data to a good DAC through an asynchronous interface sounds the same?" To that, I (and I suspect many other audiophiles) would probably raise my hand without too much reservation assuming the computers are quiet themselves and didn't sound like a hair dryer in the room. The question then becomes what is a "good DAC" but I don't think it's tough these days to point to many examples - DACs capable of very low noise floor, good jitter suppression with asynchronous interface, flat frequency response, minimal distortion, etc. All of this can be demonstrated objectively.

By the way, I don't think it would be that difficult to do a demo to show a difference between computers if there is one:
1. Get 2 of the same DACs. It doesn't have to be ultra expensive. 
2. Connect the old/supposedly poor quality Windows computer to one DAC, the new MacBook to the other. Keep the basic setup the same - same cables, same type of balanced/unbalanced connection, etc. Bit-perfect settings like ASIO output with foobar or JRiver would be fine.
3. Get an amp that can switch inputs quickly. Connect the two DACs to separate inputs. Don't tell the audience which DAC is connected to which input (just Input 1 vs. Input 2). 
4. Play the same song on both computers to the DAC at the same time. Switch instantaneously between Input 1 & 2 every 10 seconds to the audience asking them to note whether there is a difference and which they prefer. The presenter should of course try not to do anything that would bias the listeners since he/she would know what is what. 
5. Poll the results without raising hands to avoid a crowd effect - maybe hand out a piece of paper where people can "vote". Two simple questions: "Rate the audibility of the difference between the 2 computers - use the scale 0 [NONE] to 10 [OBVIOUS]." and "Which did you prefer - Input 1 or Input 2?". Get audience members to hand the results in to be counted.
6. Do a big, fun reveal at the end - which is which and what was the audience's preference?
There you go, a single blind test... Do this a few times among the various audio shows. Before long, you'll get enough data to run some statistical analysis and publish a research paper for academic interest that will be quoted, discussed, and argued over for decades :-).



2. MORE PROCESSING POWER IS GOOD...

"Justification" for a high specification computer system supposedly exists. "It takes a lot more processing power to do audio than people think" (3:13). Hmmmm... No.

At around 12:30, the presenter mentions that "at idling", his computer already uses up 4GB RAM, thus we should go for 16GB RAM. Realize that memory use in the activity monitor is also reflective of cache and buffers, not a reflection of what the computer and OS "needs". These days with all the RAM available, more of the hard drive is cached, programs are more liberal with buffers, etc... If this were not the case, then how else is the 1GB Windows 10 machine running and playing the song off JRiver earlier in the demo? Surely it wasn't swapping between RAM and another 3GB of virtual memory on the HD was it? He also recommends a quad core CPU ("seems to give you a bit of a leg up")... Yeah, right.

So if all this is so important, in another 10 years, we'll need something like a 12-core, 64+GB RAM computer, right? And that will make a difference? For what kind of music format exactly - are we still talking stereo, primarily 16/44, 24/96, etc. file formats!?

Remember that this generalization of needing a powerful computer doesn't quite fit with the experience for many of us going towards low power, cheaper, slower, 1-2GB RAM computers like the Raspberry Pi devices as per a few weeks back. Nor does this general recommendation make sense if we agree with the Computer Audiophile award of the microRendu as the "product of the year" (a slower machine than even the Raspberry Pi 3). So why the disconnect?

Look guys, unless we're purposely doing DSP on the computer (like say some heavy use of HQPlayer, or applying room correction DSP), streaming bit-perfect CD audio, hi-res and even DSD does not take much processing power at all (remember, you're more likely to be straining the network system streaming hi-res if you're on WiFi, but that's not about processing speed). Unless you're surfing the web, editing music, viewing Netflix while also streaming the music, a typical Windows or Mac doesn't need to be that powerful! And so long as we're not punishing the CPU and I/O terribly, you're just not going to hear sonic degradation. If you do, it's likely going to be buffer under-runs meaning the sound will stop intermittently as the machine tries to keep up with feeding the DAC.

On the plus side, I can agree somewhat with the presenter that it's probably a good idea not to put your HD storage on the same USB bus as the DAC (at 20:00). If multiple processes are trying to access a slow USB HD, you could have problems with buffer under-runs playing a song off the drive. I have no reason to think it's going to cause any more data errors to be sent to the USB DAC though, nor would I worry about any subtle sound differences - the problems will be obvious. Avoiding data transfer overload on a USB port is good "hygienic practice" so as to avoid problems. (To hear what actual USB errors sound like, go here.)

Personally, I do not like having a powerful computer in the same room as the sound system because speedy CPUs tend to create heat, and a higher chance to rev up internal fans. I cannot tolerate the noise pollution from such a fast computer which is why slower machines like the Raspberry Pi make a wonderful streaming device from a multi-terabyte NAS or server computer located elsewhere.



3. MORE JITTER TALK...

Sure, over the last few years, USB eye patterns have been published. Remember the 2013 test of USB cables in Hi-Fi News? Already then, I wondered what relevance this has to the output from a DAC. Remember guys, the eye pattern is a reflection of the quality of the USB digital data transfer at a microscopic level. Just because a USB transmission line may have a little more timing fluctuation doesn't mean the audio signal would actually change because the DAC's analogue output is many steps down from the data transfer. It's all about tolerance levels and whether the jitter is going to result in data error and whether the DAC is temporally linked to the timing irregularities. Data error is an easy one to figure out because a problem with that is going to cause all kinds of audible pops and crackles; not to mention issues like hard drive corruption hooked up to the USB port. As for jitter timing issues, modern asynchronous DACs deal with this already as I have demonstrated many times and evident in published Stereophile measurements.

As you can see, the presenter doesn't demonstrate anything here and goes on faith that this jitter he speaks of "corroborates with what we're hearing" (17:59). So, has anyone ever shown audible differences in unsighted listening given the minimum levels of jitter we see these days with decent USB DACs? (Have a look at a post like this to examine measurable jitter differences.)

Seriously folks, it's time to stop worrying about jitter with modern USB, ethernet, and even HDMI DAC's unless there are actual objective results to go by. You might want to still pay attention if using an S/PDIF connection like TosLink and coaxial digital inputs, but even then, I doubt it's audible unless truly awful. Nothing wrong with manufacturers advertising that they've dealt with jitter. What's more important is that audiophiles develop a realistic mindset that this is just another mental tick-box and not to obsess over it thinking that jitter is somehow a big deal these days and obsess over a "femtoclock" making any major difference. Jitter has been "fixed" for a long time now and inaudible in modern asynchronous DACs IMO.

For those who want to dig deeper into the John Westlake measurements mentioned in the video, here's his original post where he uploaded the images, and more discussions here.

BTW there was a question from the audience at 23:00 about buffering. While the presenter is correct that if errors are sent to the DAC, then buffering will not help, he did not clarify though that asynchronous buffering is a solution to jitter which might be what the audience member was referring to.




4. USE OF OPTICAL ISOLATION (@ 23:45 question asked, then discussed 47:50)...

Readers here on the blog will know I talked about this already recently. Take a look at that network layout (around 47:15) in the video where optical isolation is used after the ASUS RP-AC68U repeater which is then wirelessly communicating with an ASUS RT-AC87U router connected to the high-speed internet modem. Clearly there is much fear about WiFi implied in this type of layout where he believes the optical isolation can "cure" whatever ills might be emanating from the RP-AC68U repeater! Seems a wee bit irrational doesn't it? And how can he know if the little StarTech Fiber Media Converter might not be introducing its own noise and "pollute" everything downstream (come on, let's see the noise floor measurements). I would say it's just as likely the media converter might make things worse. Notice on the diagram how there's another network switch beyond the fibre optics along with computer, DAC/streamer, plus Thunderbolt hard drives (all AudioQuest cabling of course). Again, what makes anyone think that this ethernet switch plus the optical device isn't noisier than just a direct connection from the RT-AC87U router? (Unless somebody bothered to do some measurements of course!)

IMO the optical isolation is literally money down the drain unless proven otherwise objectively. There's hysteria around the ills of "RF broadcasts" from WiFi at 57:30, he tells us he has 3 Apple Airport Expresses in his home, each with an optical bridge in a most remarkably serious fashion!

The demo at 53:00 with and without optical isolation is just as questionable as above in point 1 (this time using a John Mayer track). This time switching from one set-up to the other took even longer and was more clumsy (remember the limits of echoic memory, folks). Again, why not try something cool and convincing - have 2 DACs, one being fed from a computer that had optical isolation and another without, then switch inputs back and forth? More fantastic material for a research experiment!



5. MIND THE POWER SUPPLY & OF COURSE HARD DRIVES MATTER... ('Cuz everything matters?!)

Someone asked about the use of bus power vs. external power supply for a portable HD (32:20). "This will be another I'm going to get flamed on... I can hear forums in the distance exploding." The answer he gives is to prefer external power supply because apparently this implies a better, higher quality storage solution. As usual, is there any evidence that this makes any shred of difference in the real world?

He believes faster drives, bigger cache/buffer "correlate with better sound". Also as expected, there's a general belief that SSD sounds better (sure, SSDs are silent and use less power, but storage ability and "sound" would be the same).

Basically, we can say that so long as it's more expensive and faster, it's better for sound quality, right!? Buy, buy, buy... Upgrade, upgrade, upgrade!!!

By this logic, we can say that interfaces like SATA III is better than SATA I, hybrid HD like the Seagate Firecuda should result in sound quality between a true SSD and standard spinning HD. Perhaps we can even extend this and say a fibre optic internet connection sounds better than cable high-speed internet for TIDAL streaming? Great theories, right guys? And this can go on endlessly without actual objective evidence for any of it; yet could just as well be included in recommendations on a presentation like this.

The idea that "everything matters" for sound quality is problematic because it also implies that there are meaningful answers for everything (and that these gurus somehow "know" the answer). It puts the consumer in a position where naturally they will ask the "guru" whether "this hard drive sounds better than that SSD?". Notice on subjective review sites how at times obviously the reviewer cannot possibly answer these comparison questions often asked in the comments section. On more than one occasion, I've seen the subjective reviewer appears frustrated that these comparison questions are being asked by the readership. Back in the late 80's, J. Gordon Holt got it right; we must be reasonable about what makes a difference and what is essentially irrelevant... I've always considered that a sign of wisdom from the late Mr. Holt (especially later in his years).

In this video, we see a very nice example of that - "Any recommendation if I have both a NAS and Thunderbolt which way I might go?" (31:30). The answer when it comes to sound quality is of course there's no difference assuming you're using reputable devices. But notice the presenter says the "Thunderbolt is a lot quicker, a much more direct path" and suggests "give it both a listen and see what you think". Yes, Thunderbolt is faster, but so what? Is a NAS over a good home network not fast enough? Nothing wrong with "giving it a listen" as the standard answer, but the real opportunities for further discussion have been missed. Why not spend time discussing how much each setup might cost? How difficult is it to expand storage with Thunderbolt vs. a NAS enclosure? Anyone curious if the Thunderbolt drive makes noise in the audio room? Any need to share the data with other machines on the network thus the NAS might be a better overall solution? These and other questions could be discussed intelligently and provide a level of understanding beyond ridiculous and meaningless "sounds better" declarations.



6. OHHHH, THE NOISE...

An irrational fear of noise continues to be perpetuated (54:50). Supposedly, the effect of noise on an ethernet network is "not subtle". Supposedly "make or break" effect at trade shows suggesting that the noise must be so bad that it would overwhelm improvements in their gear (is he referring to things like AudioQuest's Jitterbug and cable switches?).

Lemme get this straight... If hotels and trade show rooms are so bad for network noise, why then are we routinely hearing demos of equipment using TIDAL as suggested by show reports in Stereophile, TAS, and elsewhere!? Didn't seem to bother anyone most of the time. And on at least one occasion recently here at the Vancouver Audio Show, I heard an attendee comment that TIDAL "sounds great" over the hotel network and the person was expecting the music to be sourced from a CD player in the room.

Bottom line for me is that I don't deny there could be noise in unique and rare circumstances. But I have yet to "hear" an ethernet network system so bad that it deserves this kind of rhetoric as if many people should pay any attention to it. (A reference on this for further reading.)



7. ON SPECIALIZED HARDWARE...

There was a question asked by an attendee at 1:19:30 about "specialized computer hardware" like cards for USB ports. The presenter doesn't want to talk about cards such as the "low noise" SOtM tX-USBexp.

Apparently Mr. Silberman sees this as "hot rodding" and prefers to let other companies deal with "tearing the computer apart". Even more funny was the comment "that's off in the deeper end of the pool". Who's calling what the "deep end" of the pool!?

The attendee asking the question brought up an important point. If we're really talking about best bang-for-the-buck optimization and we actually believe that there's so much noise in the system from the network and from the computer so the DAC should be shielded from this, why not let something like the SOtM USB card deal with it at the USB output just before the DAC? In fact, if the USB card is believed to work well, then don't bother with any network optical isolation and let's also not freak out about WiFi and "RF interference". The card is sold to consumers after all and it doesn't take a rocket scientist to know how to install a PCI Express card. Given the proximity of the USB port to the DAC, doing this would have way more effect than network tweaks!

Gee... Where exactly is the "deep" end of this pool? (Oooohhh, I see, AudioQuest doesn't make a low noise USB card, just the Jitterbug...)



8. FILE FORMATS MATTER!!!

"What I prefer to use is uncompressed" (1:20:40) - AIFF, WAV. "There's great debate" about whether people can hear these uncompressed versus lossless compressed files.

What "great debate"? Only in the minds of these extreme subjectivist audiophiles with magical thinking.

Seriously folks, it is 2017 (late 2016 when video recorded) - not 1998 before we had multicore, gigahertz processors as commodity items. I'm really quite sure computers these days have no issue with FLAC, ALAC, WV, APE, etc. for realtime playback. (Already I demonstrated <5% CPU usage streaming 24/384 to an underclocked Pi 3 a few weeks back!) He again promotes the myth that WAV/AIFF can sound better. He knows "it's not supposed to" but still feels "there's a little bit of a glare", "fatigue" with lossless compression.

In the old days (like early 2010's), people would at least try to explain this position - stuff like how the CPU works harder to decompress, creates noise in doing so, maybe increases jitter. Nope, no explanation needed these days I guess.

"For me, I just like the idea of fully uncompressed and storage space is inexpensive enough so I have the peace of mind of a perfect file..." I think that is a nice expression of the insecurity and psychological Band-Aid needed to justify the position despite the absolute lack of rational or empirical explanation.

A reminder: I had a look/listen to this "debate" between WAV / AIFF / FLAC / ALAC / WV / APE a few years back. Don't worry, be happy.



9. CD RIPPER - H.A.???

Apparently HydrogenAudio is a bitperfect ripper :-). (1:23:35) Okay folks, I can forgive him for this slip of the tongue. I'm sure many of the Hydrogenaudio folks on the forums will be very happy to hear that Industry audiophiles are thinking of them too...

I'm guessing that he was referring to EAC. Unless there is such a thing as Hydrogen Audio CD ripper I'm unaware of?




10. POWER CONDITIONING...

Oh those noisy power lines! (1:25:00) Using Bob Marley's "Jamming" (off Exodus), he demonstrates the system with and without an $8000 "reference, state of the art" power conditioner (Niagara 7000). Again we're treated to an inefficient 1 minute cable switch, completely sighted listening, and the inevitable rationale of needing to "suppress that noise [ed. from computers, HDs, switching power supplies...] from getting into the stereo system." "Ton of noise..."

Good grief, as if the typical audiophile in the developed world is living under a radio tower or next door to an automobile assembly line resulting in such severe noise on the power grid.

It's too bad we didn't get a shot of the attendee reactions. Perhaps I missed it, but again no standing ovations, heard no euphoric comments, just someone asking in a rather pedestrian fashion what the power conditioner device was after the music demo. I'm not saying that a noise filter and power conditioner is worthless (remember, I have a Belkin PureAV PF60, examined awhile back). However, US$8000 could be the last audiophile DAC you ever buy, great speakers, and most importantly, represents many many great albums... I'm sure there's also a nice cruise or vacation to an exotic location hidden in those dollars if one desires. Any of those uses will produce a greater effect than what was heard I bet.

To end off the "demystification" talk, here's a nugget from the presenter: "Luxury once tasted becomes a necessity."

First, one has to consider whether the attendees truly "tasted" any difference. I wasn't there so maybe they did, maybe they didn't... Whether I agree with the "necessity" part is up for some discussion, but I do agree that this is very much about luxury and what it represents on a psychological basis rather than reality of the "utilitarian function" of producing excellent sound quality.

CONCLUSIONS:

I imagine myself as a computer audio new-comer trying to get into this side of the audiophile hobby, sitting among the 50 or so attendees at that talk. What would I walk away from this seminar thinking? Taken at face value, completely unskeptically, I might walk out with the following mental notes:
- Computer audio is complicated if I really want to do it right - everybody seems to hear a big difference! 
- Computer audio seems expensive to "truly" get good sound! 
- There's something nasty called jitter and I better not mess up with how I plug in my hard drives, or else "the jitter" with get bad! 
- I really need a fast modern computer! At least 16GB RAM, quad core processor, and SSD! Because I will hear the difference.
- Man, computer networks are really complicated and noisy! I might needs a wireless bridge, and a few StarTech optical isolators so that the noise from WiFi doesn't ruin my audio. Maybe I should upgrade to a Synology router. Probably need to call up the local Cisco network expert to make sure I'm doing it right.
- AIFF or WAV files only - no compression (lossless or otherwise). Got it! I need to get it done right from the start when I rip those CD's. No lossless compressed formats like FLAC and ALAC for me because they don't sound as good.

Demystified indeed! Or did I just attend a talk that increased my anxiety with questionable claims, and promoted spending money on various products "in an informative and supposedly objective way" (a.k.a. an infomercial)?

--------------------------

Postscript...

Remember that computers are engineered devices. They are the result of the application of physical principles, mathematics, and logic. The folks at Intel and AMD and IBM and Apple and Samsung and HP and Lenovo and countless other high-tech firms did not consult the Ouija board, a medium, the daily horoscope, attend a seance, ask the local priest, consult a spiritual elder every time they developed the next generation of computer technology. Neither did the guys who developed your DAC chips and intelligently designed that DAC board and programmed its firmware. There is no mysticism in any of this to "demystify". For those not versed in the world of technology, sure, it might seem "magical" and overwhelming at first, but it's simply knowledge which can be learned given interest and time.

Not surprisingly, the mystification of computer audio is to be found within the audiophile Industry itself and related mouthpieces in the audiophile press who seem to lack basic critical thinking skills and lack sound research methodology (such as these people). This results in faith in pseudoscience rather than adhering to scientific principles which are the foundation to the computers they so desire to "optimize". Mystification includes not just believing in things that have no evidence but also magnifying the impact of real but otherwise minor factors like "jitter" and "noise" in the vast majority of circumstances to "trump" up benefits for various products typically with no evidence presented for their claims. They seem to accept and perpetuate purely subjective testimonies without the insight to recognize whether these internal experiences are actually related to an external reality as opposed to their own psychological biases, or simply acknowledge that like every sensory system, hearing/listening has its own acuity and cognitive limitations. As I said in a recent previous post, just because we perhaps want to believe in our own "attribution theory" of how things work doesn't mean such theories necessarily hold any truth especially in the engineering world of audio devices built upon mature understanding of the technology.

Recently, we had CES 2017 and is it any surprise that the "high end audio" industry laments the fact that interest in their products has been dropping at CES over the years?

Consider the advancements made with the new robots, "wall paper" OLED displays, new computer designs, VR gear, increased availability of UHD Blu-Ray with HDR, perhaps even cool new home appliances on display at this year's CES. Now consider what you might see at the "high-end audio" part of the show. Some cool new speakers, more headphone amps, power amp, yet another audio streamer, vacuum tube amps are always "hot", fancy headphones maybe. And along side them another "top of the line USB cable"? A fancy power cable? Another RCA interconnect? And "magnetic conduction" interconnects? More cables "worth" >$10k (by non-engineering folks)? And some even using lithium ion batteries?

Let's just say that obviously some of the products above seem more innovative than others... And as a whole, it's perhaps not surprising that compared to other areas in consumer technology, innovation which drives much of the excitement in CES seems a bit lacking in the "high-end audio" world.

Coming back to the main topic today then. This RMAF 2016 video by Mr. Silberman is interesting as a document on the current nature of computer audiophilia. It's interesting that in the video he makes all kinds of reference to being "flamed", or that some audio forums might "explode"... Instead of anger which he's alluding to, I think many "more objective" audiophiles like myself would look at this with mild amusement mixed with embarrassment for the presenter because he knows that he's treading into uncomfortably unorthodox territory. It's amazing that presumably an otherwise intelligent individual can keep a straight face giving this talk. I would not be surprised if his presentation "job" gets tougher each year as audiophiles become more used to the technology, experienced and educated about the power of digital computer audio, leaving these myths and mystical beliefs behind.

There are certainly many things in this world to be concerned about and countless mysteries yet to be unlocked. Computer audio and basic home networking need not be on this list.

Stay sane, stay real. It's only computer audio! Fear, Uncertainty and Doubt (FUD) are easy to disseminate in the perfectionistic subjective audiophile subgroup; especially if one is motivated with something to sell to "fix" these doubts for those receptive to the message.

As usual, I don't want to really just end here on a more negative note without concrete suggestions. But given the length of this post already, let's defer some practical discussions for next week where we can look at the "lay of the land" in 2017 and consider a few options for setting up a computer audio playback system that aims to be pragmatic, affordable, and sound fantastic of course.

Until next time... I hope you're all enjoying the music!

PS: A big shout out to ryanred for the amazing work on Titanfall 2 - beautiful graphics and amazing fluidity at 4K resolution using my nVidia GTX 1080 graphics board with visual quality check boxes maxxed out! Impressive game...

24 comments:

  1. again, great post from you.. I started watching the video and felt embarrased for the presenter already at 25 minutes. at 32 minutes I turned it off. So many wrongings in what he said.

    Though, what I have found to be with computer Audio is this:
    - high end oscillator retiming the packets even on usb (dac is the sync here) Uptone regen has a a point with their product
    - galvanized usb
    - 5v cutoff from the general purpose computer

    For AES/EBU a high end oscillator is important as the host is responsible for the clock not the dac.! Hence why I ordered the super cheap Gustard u12 with high end oscillator to help time the packets. (My tascam uh7000 has a messy soundstage with sending aes/ebu to my dac).

    I have the Intona galvanized USB separation unit, and there is audible difference in soundstage with and without it (my dac doesnt have galvanized usb input). I can also measure and see the fft spectrum is more "tight" in the way noise and sound plays back with the Intona plugged in. at Deep bass, I can see a 5 decibel difference when I play a 1000hz tone where with the Intona its lower noise on deep bass.

    One last thing. I read the specifications for some of the crystal oscillators being used and guess what, they degrade over time. I read that a 1ppm oscillator ages 3ppm first year, and 1ppm every year thereafter.

    The Intona oscillator according to Intona doesnt degrade as thats its purposely design of that oscillator.!

    Take a look at Uptone regen's oscillator:
    http://www.crystek.com/crystal/spec-sheets/clock/CCHD-575.pdf

    Check the aging part.. 3ppm first year, then 1ppm each year thereafter!


    Now back to my Intona, with and without it, looking at that spec sheet and watching the noise floor for that Regen oscillator, I now understand why there is a 5db difference in deep bass with and without the Intona.

    ReplyDelete
    Replies
    1. Thanks for the note Svein. No surprise that the AES/EBU will require high quality oscillator! Same goes for the TosLink and coaxial S/PDIF in general.

      I'd love to see the results from the Intona especially over time how this degrades the audio signal from a DAC! That would be fascinating...

      Delete
  2. Steve Silberman is a well known charlatan in audio circles. Best avoided. He purposely mystifies and tries to confuse and scare the uneducated to further peddle his warez. A charlatan. And not even a very good one.

    ReplyDelete
    Replies
    1. Can't say I know much about Mr. Silberman other than this video... I do hope this guy "changes his tune" over time.

      If Industry members like this can't find a better, honest way to present their products and a coherent message about the value of high-fidelity audio, then they should just go away IMO. Better to have no presentation than one filled with inaccuracies; likely to scare away reasonable people interested in the hobby.

      Delete
    2. You know, I've never really ever figured out if people like Mr Silberman are purposely misleading people, or they actually believe what they say? Either way they create lots of damage, by way of further perpetuating mistruths, or by turning people away from the audio hobby. This is the same guy that tells us different brand HDD's sound different, different RAM manufacturers sound different etc etc... For real? Never backed up by any measurements, ofcourse. I think its in their best interests to have their "brethren" confused. They get to sell more stuff.....

      Delete
  3. While I agree with everything here, this

    "Personally, I do not like having a powerful computer in the same room as the sound system because speedy CPUs tend to create heat, and a higher chance to rev up internal fans. I cannot tolerate the noise pollution from such a fast computer which is why slower machines like the Raspberry Pi make a wonderful streaming device from a multi-terabyte NAS or server computer located elsewhere."

    I have to take issue with.

    Based on my i5 4690k W10 machine which has 5 fans and is silent because 3 are 120mm 1 is 140 and the smallest is built into the 980 reference, oh and the i5 is overclocked to 4gHz and uses a huge heat sink.

    ReplyDelete
    Replies
    1. Hi Stephen,
      Yeah, this is why I used the "Personally" qualifier :-). Absolutely, if one knows what one is doing in building the machine, there's nothing wrong with a powerful computer.

      For many years I was doing this even back in the day of AMD Athlons and Core 2 Duos. We'll touch on this next week as well...

      Delete
    2. Fair enough, I had an 1.2gHz Althlon and it did indeed sound like a hairdryer due to the lack of a heat spreader on the cpu, making it impossible to fit a large heatsink. Talking of hairdryers, have you ever owned an Xbox? My neighbours kid has one and it is quite possibly the loudest thing I've ever heard!

      Looking forward to the next part, this nonsense cannot be de-bunked enough!




      Delete
    3. Hi Stephen,
      Yes I have both the original XBOX and XBox 360. The original XBOX was truly awful. Thankfully I got a later model 360 after they fixed the red ring issue which wasn't as bad.

      The gradual reduction in CPU TDP and more quiet machine these days have been truly a blessing! :-)

      Delete
  4. The bottom line is that due to human expectation bias, any demo that is not level and FR matched, time synched, and Double Blind can easily yield positive perceptions for audible differences, even a comparison of a very high quality system to itself. At this point we know enough brain and ear science that we know a highly convincing list of the reasons why.

    ReplyDelete
    Replies
    1. Yup.

      Alas, some members of the extreme audiophile club (cult?) seem to believe they have "golden ears" capable of perceiving the ultimate level of detail (dynamic range, soundstage, time domain, etc...) and minds completely capable of transcending psychological biases plaguing mere mortals.

      Delete
  5. I saw this demonstration about 3 years ago in Montreal,
    What is disturbing is to find out that this guy repeat almost word by word the same presentation,
    like a well-rehearsed sales pitch...
    The most disturbing affirmation for me was the claim that you need a MacBook with 16G of RAM on SSD.
    After the demonstration I asked this guy if he was totally convinced that you need as much cpu and ram to playback a 44.1KHz.
    His only answer was...theirs’s an audible difference, that's enough for me to suggest this as the ultimate configuration.
    Obscurantisme in Audiophilia...

    ReplyDelete
    Replies
    1. Hey there Blogue,
      Ha! I bet that 4th generation "Haswell" quad core CPU in the 2013 MacBook absolutely sucks compared to a 2016 "Skylake". Soon to be superseded in sound quality by the forthcoming 2017 "Kaby Lake".

      "Ultimate" configuration eh? Obviously can never exists. Liquidity must flow!

      Delete
  6. It is possible to build an I7-basxpensiveed system with no fans at all. Been there, done that! The resulting system will be large and relatively expensive, that is all.

    ReplyDelete
    Replies
    1. Hey there Arny. Nice to see you at these parts...

      Yeah, although I have not tried an i7 fanless, I have done modern i5's fanless without too much difficulty or expense. Certainly with modern generation CPUs with medium-power parts at 65W and below, it's coming along nicely at reasonable price points!

      Delete
  7. The 2016 YouTube video may have comments disabled but the same 2015 YouTube video does not. I guess they learned their lesson.

    According to AQ Kingston RAM is the worst sounding module out there also.

    As to the networking side of things. I ordered in a 12 foot $330 WireWorld CAT8 cable and made a 315 foot CAT5e cable. I captured two tracks into my ADC with swaps made out randomly. I made them available for download and so far no one can tell when I made changes.

    The generic cable was $0.30 / foot. The WW? $27.50/foot.

    ReplyDelete
    Replies
    1. Thanks Matt.

      I see the 2015 version - here's the YouTube:
      https://youtu.be/oOrjQXyqZvw

      Damn, I knew there was something wrong... Now I know it's because of the 16GB of Kingston RAM :-). LOL.

      I watched that video for 20 minutes and couldn't tolerate it any more. Clearly from the way he talks about this, his technical knowledge is limited. How can anyone who's supposed to be an *expert* in computer audio make rookie mistakes that hard drives are "optical" or that 128kbps means "kilobytes". Painful.

      Matt, if you have that download available still (CAT8 vs. CAT5e), feel free to post the link! I'm sure folks would love to give it a try.

      Delete
    2. Here you go:

      https://www.amazon.com/clouddrive/share/5NUQOsUEqZKVgon85pYhxnlNYlAhoHhT3FF9cP0NKfr?ref_=cd_ph_share_link_copy

      https://www.amazon.com/clouddrive/share/bD0Z87vfscWzA3PhyE42JuVGagNXmvcKTxqWhQpLIKO?_encoding=UTF8&mgh=1&ref_=cd_ph_share_link_copy

      I like the tidbit in the video where he mentions setting up a static address. It was 192.168.1.256 (well something .256). I got a good chuckle out of that since you can't have a 256 IPV4 address.

      Delete
  8. Thanks for the great review of that YouTube video, It sure had me Mystified, you resolved all my Questions that I assumed were BS. Cheers.
    Razz

    ReplyDelete
  9. Having sold all the cables insulated by "snake skin" now those "gurus" are trying to sell those gigantic computer to the non adverted audiophiles !
    Me, I have found in my garbage a very old Centrino Duo, 2GB RAM.
    It runs Qobuz Desktop, JRiver Media Center 22 (as DSP), TeamViewer 12 (to communicate with it remotely).
    CPU utilization is (in the mean) 27%, memory used is 1.6GB.
    The bottleneck is the network: Qobuz Desktop asks for 9Mbps for approx. 20 seconds, when a new track is loaded, and since my network provides a maximum of 10Mbps (hour dependent) I cannot do anything else (like downloading a movie, even on my "main" i7) with the network.

    ReplyDelete
    Replies
    1. A bit same, recycled OEM pentium, E series processor with 55W consumption, passive cooling GPU and it can run full HD, stream to another computer in the network, browse FB and run room correction and multiband comp/limiter at the same time.. Audio alone takes about 5% max since i am using "heavy" processing. All default buffers as tweaking windows buffers is stupid, it makes the whole system unstable.

      Network is gigabit ethernet, wired. With WiFi, i can choke it quite easily. When i had 100mbit i can't remember having even once streaming stop or stutter because of network, audio is one of the lightest data streams.. CD could run with full quality using 90s modems.

      Delete
  10. Unbelievable how calmly he stands there and lies. I can say that i would've been thrown out of that room the moment he said that the obvious loudness difference is caused by noise floor. That is so demonstrably false, i can hear bloody truck in the background at one point in the video so we know it is very close to some loading dock and street noises are clearly coming thru... We are talking about 40dB background noises, and that is conservative estimation. Expected noise floor should be way below that in both devices.

    Absolute rubbish.

    ReplyDelete
  11. I note from a recent Darko article that Steve Silberman, formerly of AudioQuest and Ayre Acoustics, was recently appointed as Roon’s VP of Sales and Marketing. As a, up until now, happy user of Roon, I'm not sure how I feel about this.

    ReplyDelete