Ben Robinson

The thread is rather long-winded so in short: I asked Sandyk what test he'd prefer as he did not like my straw man test. He stated would I be satisfied with double-blind ABX tests, and I said yes if he sent me details of such tests showing repeated selection of audiophile SATA cable more than expected by chance, then I'd say that was something to look at closely. Sandyk said he\she would send this to me but in fact did not send such a thing. I gave him the respect of accepting his proposed test data, sadly he\she did not give me the respect of sending me what was agreed. So I entered the thread to point out it the agreement was not followed, the above noise ensued. And please remember it's illogical to prove a negative. Common sense, politeness and indeed the scientific method demand that those making the claims perform tests and make test data available for examination. Obviously to critique test data it must be forensic. It is very peculiar that only in the world of the audiophile do we find people who staunchly resist providing empirical data to support their - in this case extraordinary - claims. If I turned to you and said I had a family of purple pixies living under my bed you would ask me to prove it, if I then said, "prove they do not live there!", you would rightly shake your head and mutter the word, "idiot", under your breath as you walk off. That's what's happening here. Extraordinary claims flippantly made (despite the peta-bytes of data transferred over SATA, worth billions of ££££, every day, music for some reason arrives in a greater or lesser state depending on cable used) but as soon as you ask for data suddenly everyone forgets logical thinking and the general way the world works and demands a special case is made for music. And, it is absolutely cool within the audiophile fraternity to talk about un-falsifiable differences in sound perception but when you move into an area of science that is subject to massive rigour - computer science - people like me will shake our heads, laugh and walk off, unless you can play by the rules of that discipline. Music is merely science. Things are in key for scientific reasons of frequency. Musical intervals are mathematical rules. Musical harmonics are there for scientific reasons. Music is not a special case that sits outside of science. People just can't be arsed to make the effort to think logically about it and they are then ripe for duping from the snake-oil salesman, SATA leads and all. I have just over £10k of hi-fi & home cinema gear myself, selected mainly on functional and budget merits (by the time you reach £10k solutions you've moved way beyond jack of all trades integrated solutions and any sonic difference is down to crappy hi-fi shop listening tests administered by eager salesman, or simple fact of some kit integrating better with others), so it's not like I don't respect good quality audio, I just know that some things can make a difference and some cannot. Good vs crap speaker = big difference. Industry standard SATA cable vs audiophile cable = no difference until someone proves otherwise with measurable data that stands up to challenge. I am actually not going to discuss it anymore, it's like talking to religionists. You say, "prove your god exists, or at least give me some falsifiable evidence he does". All of a sudden it doesn't work like that, He's a special case and you have to believe in faith alone and how dare you say He doesn't exist you terrible blasphemer. yadayadayada. Your faith is all cool in the audiophile world, but don't bring this baloney into my world of computer technology, because I'll just kick it out. And pretty much anyone I know who operates and is qualified in the computer sphere says the same, and yet we all have reasonably expensive hi-fi investments and an interest in music & technology. It's because we love music and have bothered to apply some science to it that we can tell a blagger a mile off. They say things like this: "studios, with their particular preferences for monitoring speakers and muscle amps etc, may not be the best place to detect SQ differences" Studios are the very best place to measure such differences because the entire thing is setup to be utterly neutral in order to achieve a totally neutral mix-down. the last thing you want when mixing a track for playback on any number of permutations of home system, car hi-fi, phone, tablet, DJ mixer etc, is any colour at all. The very problem with using studio monitoring solutions for recreational listening is they are too analytical, revealing and dry. They absolutely show up any deficiency in the process, that's their job. All your records would sound pretty awful if they were NOT any good at that. But as I say, I also have a reasonable hi-fi at home, so I have experience of both worlds for listening.

I did read your earlier post but did not consider it worth a response. However, as you have asked again I am responding. As stated many times, I have not only updated\changed SATA cables, I have swapped out entire studios full of top of the line DAWs and noticed no difference in audio quality, merely performance and functionality. But my experience is subjective and I therefore relegated it to inclusion purely for a bit of fun. Vested interests try to poo-poo DBT ABX tests but that is mere nonsense. They are essentially saying that the industry standard test for drug safety, efficacy, psychological treatment testing etc is not good enough for a few people with self-professed 'golden ears' to test a bit of music. Well, excuse me while I believe the educated and highly qualified scientists in the world over the arm-chair testers with vested interests. I am sure you would all love it if you were prescribed medication by your GP where the only test rigour was a load of unqualified arm-chair testers saying, "I feel OK after taking it for a while". Honestly, having opinion about music and trying to find audio purity is one thing, talking nonsense about test methodologies good enough for life critical applications just makes people look silly. That said, it's true to say that even the best of those tests cannot account for the fact people listen to music and therefore are highly susceptible to all sorts of bias; I have already stated we each experience the world through the virtual reality of our senses. And, if a placebo effect makes Joe Bloggs think his music sound better, then hey, maybe he should consider spending money on it because you quickly get to a technical point with digital where you cannot make anything better. But let's not kid ourselves about the psychosomatic effect Joe Bloggs (thinks) he is hearing. Crucially, let's not forget it was another poster - SandyK - who recommended the ABX DBT test. And this is one of the problems, even the proponents of such things as better SATA cables cannot agree how to test things between themselves. And yet the hi-fi industry then walks into a domain which requires very high rigour of its practitioners - I have a BSc Computer Science plus nigh on every professional accreditation going, and I am hardly a rarity, plus all our testers are ISO accredited - and wonders why we don't take it seriously. Without any evidence I will no more take seriously a lay-person telling me that SATA cables affect sound quality than Colloms would me telling him the little plastic insulators on transistor legs affect it.

Most likely you are right and it won't change but I think that because dB is such a crucial component of a listening test it should be baselined at the start of each iteration to prove that it's not changed. A power fluctuation either externally or internal to the analog stages of the playback solution could in theory change dB by 0.1. So it's really a case of if something is worth doing it's worth doing properly and you baseline the important metrics at each test run. Sorry, not read any new PMs today, perhaps missed the advert points. At the end of the day, it really comes down to, "do they actually want to get to the bottom of the matter?" If so, get on with it and do the job properly and provide irrefutable proof, not some half-assed job that is easily refuted on basic test principles. And from a magazine perspective they clearly MUST own a dB meter of sufficient resolution or all of their listening tests are discredited anyway, because clearly any testing they have done or will do of 2 or more amps is going to involve a relatively large change in volume between them, so they are obviously measuring that. And they of course must be used to running double-blind ABX tests otherwise, again, they are simply selling an opinion. So they absolutely have the wherewithal to deliver the test, if they do not, why are they even talking about it with any sliver of authority? If they are operating in a sphere and stating themselves as an authority, then saying, "it's too hard to validate what we are saying", is just not going to cut it in the real world. Don't get me wrong, discussion, creative thinking, theorising, searching for truth, all great stuff but at some point you have to 'shit or get off the put' (as the Canadians say)!

Now you are saying that 2 SATA cables meeting their specifications need to be decibel-normalised ??? Both replays have to be at the same volume because it is shown that a very small volume increase (I think it's +0.3dB?) will make a listener prefer the louder source and say it is better quality. So a proper objective test is ABX, double-blind and dB matched because clearly there is scope for error in replay of each test iteration (volume controls etc). We could lock volume controls but you could find small volume increases in each iteration of test due to other environmental factors (power mainly), and as it's so crucial we would measure it IMO given the ease with which we could do it. You are demanding of non engineers what they are unable to give you. No, absolutely not. I have read everything you have to say and I salute you as a listener in trying to find audio purity. However, as another poster pointed out, this Colloms chap has an engineering degree and a history of audio testing, therefore whilst it would be unreasonable of me to expect you personally to carry out the above test, I would expect Colloms could carry it out. If he cannot, the industry - certainly a magazine - can do it. Why have they not? If people cannot prove their extraordinary claims in life, they really should not make them. Otherwise consumers, who can tend to be rather sheeplike in the hi-fi world as in lots of other spheres, can spend money on snake-oil. So it is in fact their moral duty to back up their claims IMO. I would never recommend a given solution for a client based upon my personal feelings, in fact I'd be out of a job if I ever did! - my claims are tested by independent authorities and I absolutely mandate we are not part of that process, in order to remove any possibility of bias. The hi-fi world is no different and, now it's becoming based on the IT world, it should play by our best practise. Instead all I am hearing is, "it's too difficult". Not really good enough. These are normal test demands that can be met by relatively low-skilled people. You and I could make this test with only a small bill of materials and a few resources: Audio server, amp, speakers - Multiple SATA cables (test products) dB meter (ensure each test iteration has same volume) 2 non-blind resources (swap each resource around) for test setup 2 blind resources for test administration (swap around resource as above) ABX samples Say 10 test subjects, no doubt willingly found in the hi-fi listening world 1 ISO accredited test arbitrator (very important so both 'sides' see him as neutral but skilled) 1 test approach document 1 test script document This test is cheap and would be done in two days at most. Why has no one in the industry done it? EDIT: I am told 0.1dB is enough of an increase to make someone perceive an increase in sound quality. So a key expense is the dB meter of sufficient resolution\accuracy but any hi-fi mag will have one already otherwise all their lab tests would be invalid anyway.

Don't get me wrong, he near definitely has no axe to grind and is probably dedicated to his search for hi-fi truth. He's just unwilling to realise that his subjective truth is not necessarily the objective reality. He is as prone to the human condition as the next man, it's just he's not willing to put himself outside of that condition by removing himself as a test component. And he may be well qualified in electronics, but he absolutely does not understand computer technology, not one bit (pun intended). As stated. I have done my homework. A 20 year career as an infrastructure architect (specialising in storage and directory service architecture) on some of the largest technology projects the world has known, and now director of two UK technology companies, plus 10 years in the music industry creating, producing and selling electronic music has given me a lot of homework to do. Much to my wife's chagrin!

Hi Finally found some time to read the items SandyK PM'ed to me, or try to read them! The Hificritic links do not resolve to any threads, so they are a no go. I read this article however: Audio Networking: A potpourri of computer network audio findings, including updates on the Naim UnitiServe, gigabit switches, CAT 5/6 cable and ripping issues. Review By Martin Colloms And it is merely an article discussing some listening tests, which are not double-blind, dB-normalised, ABX tests, so are as prone to error as your standard listening test administered by a hi-fi salesman in any given demo room. i.e. utterly biased and error-prone. This article however is full of the misunderstandings typical of someone who does not understand computer technology, "Cited further advantages for the SSD are lower power consumption, and hence lowered electrical noise; no moving parts, so lower acoustical noise and vibration, and therefore potentially cleaner CD rips and therefore the promise of better sound quality. " At the level of the rip, it will ALWAYS be the same due to a verify and error correction of any half-decent ripper. It should be noted there is no technical need for an SSD in a music serving device because the bandwidth and latency requirements of even lossless music, are so low that even the cheapest of modern HDDs can meet them. I actually use a RAID10 NAS to serve my entire Brighton office of 100 employees with over 5 TB (i.e. non-SSD) of data over trunked ethernet, and it's more than fast enough. However, installing SSD does of course raise the sales price... I nearly choked on my coffee at this sentence, "while the SSD sounds cool, sophisticated, pure, smooth, vital, almost crystal clear, it ultimately lacks the full quotient of musical drive and rhythm that we know is possible from our test programme, due, we believe, to a touch of 'processed sound' detachment. " SSD and HDD are exactly the same architectural components, conforming to precisely the same standards with identical partition and format. "Processed sound detachment" is just gobbledegook in this context and, with the above sentence, the writer destroyed any shred of technical credibility I was willing to offer him. He clearly should no more enter the arena of computer technology than I should start reviewing different methods for gene splicing in the Lancet. I then tried to read this article: http://www.hificritic.com/Forum/yaf_...s-works.aspx?= Most recent.See replies 86 to 91 But either the link or database it depends upon is dead, or you have incorrectly formed the link in your PM. I tried various forms of the link to no avail. The other two PMs you sent me are merely people's subjective anecdotal experience and not what we discussed on the forum: double-blind, ABX, db-normalised tests. I respect people's subjective opinion but remind you how fallible it is. After a serious and life-threatening motorbike accident resulting in massive multiple internal injuries and breaks\dislocations, I was able to manage a fair level of my pain via positive attitude - mind over matter - throughout roadside stabilisation and a 20 minute ambulance ride (collapsed lung and non-paramedic attended, so no pain relief medication possible). If I can do that I have no doubt someone can convince themself X bit of music sounds better over Y cable. So whilst I thank you for the time taken to send it to me, there is nothing here that represents testable data and certainly nothing that can be falsified (i.e. scientific theory data) so it's pretty meaningless to anyone interested in seeing an un-obscured and objective view of the truth. If the broken links do contain such data, then please resolve and send on because I absolutely would be interested in real test data, as I said previously. Remember, I stated I'd be interested to read about repeatable double-blind, ABX, decibel-normalised listening tests of SATA cables, and I genuinely would be. But that is not what I have been sent.

Thanks for sending them on. I'm back to work with a 2 day meeting bonanza so won't be able to look until Friday. And if people didn't like getting pissed off they'd never sign up to public forums! ;-)

Hi Quick one before bed. Yes, there would need to be a series of double-blind (if the test is not double-blind then the researcher\tester can of course inadvertently introduce bias) ABX test with verified data, such as you postulate. So yes with improvements (double-blind, ABX, dB normalised. etc) that is a valid listening test and the results have to be above that which is found randomly. If those results are consistently above the statistical results for chance then of course that's something to start investigating.

Which part of the statement, "my own subjective experience, therefore unreliable and included only for fun", strikes you as asking to be taken as fact? My original statement is below, please show me where I ask it to be taken as fact, and not as the subjective noise that it is. I have actually stated that different SATA cables, drives and even controllers made no subjective difference, when I described the subjective experience I had when updating a production studio year in year out. I also stated it's subjective and therefore not reliable and I merely included it for fun. This statement was made at your behest as you kept asking me if I had tried it (different cables). I posted that in the course of my daily business of owning a production studio it was tried many times but the output of such a subjective experience is valueless. It is there in black and white, so please don't lie about me, that's very uncouth of you indeed. As you have yet again had to resort to ad-hominens and also in this case a blatant lie, I can only assume that for whatever reason, you feel you cannot enter the debate on objective and impersonal terms. Whilst I am prepared to discuss the topic, I don't think we'll get anywhere until you stop making this conversation personal. Certainly all the time you are lying about what is previously said we shall get nowhere because we'll keep retreading old ground. I note you are also still to answer the question of what makes audio data so special that it does not arrive intact and must have higher than spec SATA cables? Why are you avoiding this question over several posts? No, that is exactly what you are talking about when you tell me that certain SATA cables are an improvement over spec. SATA cables are a part of storage architecture in the same way bricks are a part of building architecture. This is what you fail to grasp. You cannot divorce them. However, I have already cleared the decks of my experience and merely asked you to collaborate on a simple test with me. I am afraid that argument is only dead in the eyes of those that wish it to be dead. The entire world is transferring data right now using parity and CRC systems that rely upon bits being bits. Everything seems to be working just fine. This message will appear correctly on all reader's screens for example, purely down to IP network error checking and indeed the disk parity checks on RAID arrays (connected with SAS and SATA cables) the world over. I am afraid you could not be more wrong, you really should exercise Occam's Razor a tad more often. Well, it's lucky in the case of bits they can be measured very easily and of course, testing good practise (any ISO accredited test managers here will agree) deems the simplest test the best, so long as the test case can be resolved. In this case it can and is in fact the very test that is exercised to test data integrity when enabling the drive write cache (DWC) on a drive before committing high value production data to that drive or set of drives. As I say, if it suffices for billions of dollars of data transmittal, it suffices for you to hear your music. This very test was used a few weeks ago to provide a go\no-go flag for enabling DWC on RAID 1 disks (mirrored disks) in 66 servers in an enterprise solution for a large UK airport. It worked for them. Clearly your PC\NAS demands a different, perhaps more stringent test, to that used for this sealed and DfT approved life critical system. The logical explanation if something can be heard with your ears and others confirm but test results do not, is actually that the test was somehow administered incorrectly: namely participants were inadvertently given cues (test not double-blind), subjects suffer confirmation bias or placebo affect due to having already spent the money, subjects have expectations already set by perhaps reading an article or sales pamphlet. Why do you think clinical trials are able to force placebo effect at will when they desire to? Your lack of understanding of methodology, how the human mind works, good test practise and how digital technology functions does not mean that empirical test data is rendered invalid. Every human experiences virtual reality, filtered through their own senses. Hence why, to use your terminology, such subjective testing in this day and age is considered 'pathetic'. Let's administer this new drug to 10,000 people and instead of measuring the effects objectively via double-blind testing and blood samples etc, we'll just ask them if they feel ok. Sounds like an amazing idea, I really think you should run for chair of the BMC. The above sentence is the only true thing you have said this entire conversation. You maybe able to hear it, as stated I respect your human experience. But knowing how fallible is the human experience, I think you heard it for placebo or confirmation bias or any number of physiological reasons. And until you can provide testable data, or indeed answer the lingering question of what makes audio 0s and 1s so special out of all humanity's data, I won't change my mind. Shame you can't improve the test, after all if it's that pathetic I'd have thought it would not take a lot of acumen to improve it. Don't give up your day job ;-)

OK, then I have taken your words a step too far. I respect your human experience. But I remind you just how fallible and fragile the human experience is. I would also say 'debunk' was too strong a term on my part, I challenged the forum to provide evidence, which has not been forthcoming, in order I could start to debunk it. Without an evidence based theory there is nothing to debunk, so yes, my wording was incorrect given the reticence in providing evidence. You merely have an idea at this point. I am questioning this idea and how you arrived at it. However, I actually stated it's not possible to debunk your ideas without evidence because they are not falsifiable, hence not fit for scientific discussion for that very reason. Much the same as religion, you will fall back on faith in subjective experience when faced with facts. I have actually stated that different SATA cables, drives and even controllers made no subjective difference, when I described the subjective experience I had when updating a production studio year in year out. I also stated it's subjective and therefore not reliable and I merely included it for fun. Please read earlier posts in the thread if you have missed this personal example. If you cannot see how 20 years storage architecture experience is entirely relevant when considering how different storage cables may affect data transmission, then I have to facepalm. That would be the same as somebody coming on here with 20 years experience designing transistors and you telling them their experience is invalid when it comes to discussing amplification. My main hobby is to collect performance cars and I therefore frequent car forums. If someone with 20 years experience of internal combustion joins a discussion (sadly a rare occurrence) I am very grateful for the insights they bring to my hobby, I don't poo poo them and tell them they are irrelevant. However, if you wish to consider my experience in storage architecture and 10 years in music production irrelevant, then I clear the decks of it. So, to move forward to the debunk or perhaps even a validation, do you accept my test design of copying a WAV from an authoritative source to a hard drive using 2 different SATA cables, and doing a resulting bit comparison as a good test? If not, please explain why it does not suffice and improve the test. You have also dodged my question, why do you feel music is a special case when it comes to data transmission over SATA? As in, why today alone has petabytes of various data been transferred with 100% integrity over a wide range of SATA cables, and yet musically this can be questioned? I have transferred 3TB myself today with not a single integrity issue, there are billions of such transfers happening right now. If SATA cabling was error prone, why do we not hear about it? Or rather, why do we only hear about from the one tiny vector of the audiophile? Thank you.

Then why are you doing exactly that? You are saying, "accept my statement as fact, that I can hear the difference between two SATA cables. I cannot present you with any testing data but please accept it as fact". And yet above you seem to acknowledge that is not how the scientific method works. Why are you making this contradiction please? You are the one making the extraordinary claim without the testing. I am merely refuting it based upon the decades of testing that the computer industry has in fact carried out on storage transfer and data integrity, and is carrying out daily in the transfer of billions of ££ of information far more valuable than the latest album from whoever. I actually don't need to do any testing (I am after all merely referring you to the vast library of storage architecture standards that are tested and used in production on a daily basis) but I could easily prove it by copying a file from an authoritative source to another drive using two different SATA cables. The file would of course arrive bit for bit perfect. However, I suspect you would be one of these people who would introduce snake oil at this point, saying that just because the 1s and 0s arrived intact, the file when listened to may sound different for some reason across either SATA cable. If I have misguidedly thought of you as a snake-oil type, then I apologise, please let me know, and - once you have explained why you are making the contradiction in the first paragraph above - I will conduct a test using bit comparison and two different SATA cables.

Thank you, that is what I am trying to say and you were far more concise about it! I then lay down the challenge of: If readers think music binary data is somehow a special case outside of these standards, please give me the evidence to show how that special case arises.

I have no theory, this is the point that seems to escape you. I am upholding the status quo, that digital data transferred over a cable of the correct specification will arrive intact and unblemished. This is a proven industry standard. I am merely asking you guys to show me the evidence of yours, so I may attempt to disprove that evidence. Rather than give that evidence, you would rather attempt to draw me into an argument. And this is where you show how wrong you are. It is entirely relevant. 1 and 0 carried over an enterprise SAN containing banking data are no different to audio data carried as 1 and 0 over your NAS or PC bus. They are identical and it is the gross ignorance of this in the audio world that is the issue. Audio data is no more special than if those 1001010 streams are banking data, this post, a technical drawing of a new car, a TIFF of a party I was at this weekend. If those 0s and 1s were subject to so much 'impurity' from poor SATA cables then next time I buy a new car I'd be at risk of it being off-spec, my friend having red and not blue hair at the party and ending up with a few extra thousand in the bank (nice idea, but it never happens). If you disagree please explain why audio data is a special case out of all the near infinite uses SATA storage is used for in the world. Thank you. I rest my case. 'Higher spec' SATA cables would indeed needlessly decrease profits in a SAN and therefore are never used. Data arrives intact without them. But you can be sure if it did not, the profits would dive and the lawsuits rise due to data integrity issues, and therefore a new spec would be designed by the relevant standards authorities. In your sentence above you prove my point concisely. The industry standard DAW that everyone expects to see when they walk in a high-end studio is Logic Pro and it only runs on OSX. So all professional level studios are kitted with Macs to be commercially viable in terms of meeting client expectation and having a common toolset understood by professional engineers the world over. That is the rationale. As stated I have 20 years hard experience in the storage world from the days of MFM drives\interfaces up to the latest fibre channel enterprise solutions, and everything in between. This is experience gained and measured using empirical data repeated in test after test over decades. First hand experience does not get a lot better than that I am afraid. And it's not an opinion, it's a tested, proven set of architectures that have helped develop and govern the very standards you use every time you listen to music. What is more likely to be true? Storage solution architecture standards, designs and tests governed by a multi-billion £ industry staffed by some of the finest computer minds the world has known - or some bloke sat on his sofa doing a non-double blind, non-dB normalised test between a couple of cables which he has bought\sponsored and therefore is highly susceptible to confirmation bias. And here again you show a lack of understanding of the domains you are now operating in; every time you listen to a CD, WAV, FLAC, MP3, streamed\played from a CD player, phone, cloud solution, NAS, PC, Mac, dedicated streamer or anything using a DAC, you are interfacing and depending upon the computer science industry. For it was us computer scientists who helped develop every single one of the technologies, protocols and standards that are used by any music listener who uses a digital medium at any point in the process. Your music is an application running on a storage architecture. Whilst the application changes, the architecture (frameworks, standards, protocols) never does. First hand music experience does not get a lot better than actually creating and producing some of the very music people are listening to today and have done over the last ten years. So I really don't see how you can imply I've made no effort to garner first hand experience, when I have spent my adult life not just gathering the experience, but being actively involved two industries that actually shape the experience. Human experience is highly fallible, just ask any policeman who is trying to gather witness evidence of a crime. Hence why we develop objective testing models to show us what is the real truth. What is your objective test model and your objective evidence?

As I am sure you know that is highly disingenuous of you. I am not the one making extraordinary claims and therefore have no need to present evidence. It is well understood scientific practise that it is the people who present theories that must justify them with evidence. If I were to then wish to challenge the theory it would be incumbent upon me to disprove the evidence and\or its results or you could of course call me out on that. That is part of the scientific method and we are ultimately talking about a matter of computer science here, therefore I would expect people attempting to practise that discipline follow the precepts laid down by my industry. One of the reasons *some* audiophiles are treated as idiots by my industry is because they do not understand and apply the scientific method to their testing. Go into any hi-fi showroom (and I have been to many, I have a reasonable hi-fi collection) and we can see this first hand with very poorly administered listening tests. As the public becomes more aware and more tech savvy, this starts to become more and more seen as unreliable in approach. You guys are merely perpetuating this with subjective views. Remember, I am saying something makes no difference, so it is a logical fallacy I need to present evidence of no existing difference. That said, in 2 decades of designing, governing and architecting some truly enormous storage solutions, I can confirm there is no drop in data integrity caused by different cables, controllers or even different hard drives. The only time a drop in data integrity arises is when cable or drive is failed, i.e. not to specification (either out of the box or later in life). But even the cheapest cables that are bundled with a bottom of the range PC are to designed to specification. This is indeed tested regularly and scientifically, when test teams collect data used to select enterprise storage solutions against architecture principles and specifications someone like myself draws up. And of course an enterprise storage solution is merely a very large bundle of cables, drives and controllers. Surprising then that HP do not specify audiophile SATA cables in their 3PAR solutions that serve the largest businesses in the world eh? And yes, as a subjective (and therefore not trustworthy, merely added for fun) measure, changing and upgrading Macs (and therefore the cables within them) in a fully kitted studio on a yearly basis yielded no audible improvements, merely the ability to process more and more samples\jobs in parallel as the years rolled on. Music produced in our studio was merely acclaimed (or not!) on its creative merit, because quality and integrity-wise, DAW behaviour plateaued a long time ago, internal cables and all. Just remember, we (wearing my audiophile hat) are the guests\newbies in my industry - computer science - and if the accepted good practises are not followed - measurable test data provided to support theories - then we (computer scientist hat on) shall think you are talking twaddle.

I was already conversant with the JS one and he is STILL applying analog ways of thinking to digital solutions. Make no mistake, some people I mostly have a lot of respect for, simply do not understand what they are talking about when the come head-to-head with the computer domain. They take their fervently held (and mostly accurate) beliefs from the analog sphere and simply port them to the digital world. Noise absolutely matters in analog land and one of the first things we'd do in the studio is add off-board sound cards to all our Macs. Why? Because the internals of Macs\PCs are hotbeds of noise that can affect the analog stages of on-board sound. Hard disc to SATA to SATA controller is purely digital in realm. Introduction of noise sufficient to 'lower quality' (pollute the digital stream) will result in errors that are noticeable. End of story. It takes as little as one incorrect register to cause a program crash. And you could not hear that one incorrect bit if it were occurring in audio world. It really comes down to the same thing, when JS provides us his empirical and falsifiable data we can have a look. Until then he's just another person putting forward some ideas. I've not read the other article but I shall when I get time at the close of business. I once used to speed everywhere in my car, because it felt that I was getting to meetings more quickly. Eventually I got sent on a Speed Awareness course and the point was made that we save very little time by speeding. I decided to test this and asked my PA to time two journeys to the same office, one with me hitting the regular 100mph and another with me observing the motorway speed limit. Net result (of course you can guess), negligible difference in arrival time. Over a 70 mile distance volume of traffic assured me the average speed was the same on both runs. Point being just because we think something makes sense doesn't mean it does. It has to be tested. And JS's theory still needs testing. Reading his posts he simply does not understand how computers work, but reading his biog he most definitely will be swayed by confirmation bias. So, he clearly has the engineering know how to test his ideas, let's see him get on with it. But, you only have to read these three sentences of JS' to see he struggles with the subject matter: "Digital is indeed sending fully analogue electrons over a wire. And indeed, the block pulse degrades with the length. A good digital cable is one who minimizes this degradation"