|
Post by soundandmotion on Oct 24, 2019 23:49:02 GMT 10
I assume we all like listening to music. No one needs to tell us why we should (already know why!), or even how we should (already know how!), although maybe some tips pop out... a glass of wine, dim lighting, adding a SuperMegaXYZ widget to our system... Well, if the widget costs more than a couple bucks, I want to know why you recommend it. And whether it's the SuperMegaXYZ or just an AcmeAmp 1, I may want to listen myself. So a listening "test", done by you and/or me. Whether it's a "test" or an "experiment" or just a "comparison", it's all done to provide further useful information to someone. In many cases, it is done just for oneself, ...and perhaps a couple friends, ...and maybe some strangers who trust you, or maybe a group that doesn't trust you, e.g. an editor and 2-3 reviewers, or some forum members. Why do a listening test? The only reason I can think of to do a listening test is to convince someone of the result, or at least stir their interest. Can you think of other reasons? And the first step to figuring out the how, is to first answer the why: whom will you want to convince? That guides you to the best way to execute the test. If you will only want to convince yourself, you already know what you need to do... unless you want to do the test to a higher standard than you already know. If you want to convince others, you need to understand what their standards are. If you don't find that, you may be disappointed that all your effort was unsuccessful. And if you don't want to waste your effort, you should find the conditions needed to convince before you start. There are no rules. There are no laws. If you find "rules" somewhere and you violate them, there are no ABX police who will lock you in the DBT prison. But sometimes there are standards or standard methods, but it depends very much on what exact question you intend to answer. You must also take into account a common feature of human nature: if your results contradict someone's beliefs or "knowledge", they will pick apart your methods. Whether they are objectivists, subjectivists, audiophiles, scientists or engineers, whether it is you or I, you can expect resistance to getting someone to admit they were wrong. Some listening tests are very easy to perform and some require quite a bit of effort. I'd love to discuss this further, if there's any interest. As for DBTs, they are often a waste of time, as the majority of E.E.s appear to reject the results when the results don't go the way that they expected. This is based on a series of 6 separate DBT sessions with 8 repeats in each, performed by E.E Martin Colloms and a friend , with comparison.wav files that I suppled . The results were a total of 48 out of 48 positive results and the DBTs were correctly performed, yet almost all the qualified E.Es in a large USA based forum refuse to accept the validity of the results. My reports have also been confirmed by 2 NSW based E.Es in the home of one of them. Before they became E.Es quite a while ago, they both worked behind the counter at a Sydney Jaycar store where I first met them back in the David Tillbrook days,(ETI5000, and AEM6000 amplifiers) where he visited the store on one occasion and I got to meet him. Hi Alex, You have mentioned these sessions many times on several forums. I was interested, so I PM-ed you on AS/CA. You were kind enough to send me further info. But I remained unconvinced. I didn't follow up because I knew you were quite passionate, and I didn't want to start a battle. It is important to realize that I'm not convinced you were wrong or that your methods were wrong or that the results lacked meaning. I'm simply unconvinced one way or the other for one simple reason: I need details about the exact methods used... not each time you mention the sessions, but somewhere to find. The people involved, and their degrees and/or employment history, carries very little power to convince me. I've met M.D.'s, Ph.D.'s and engineers who were frankly not so bright, and I used to know a handyman, a former felon, who was brilliant. I've been lucky enough to work with several brilliant people. You say "the DBTs were correctly performed", but to be honest, that is up to the people you want to convince to judge for themselves, whether they are E.E.'s, Ph.D.'s or handymen. I couldn't find the details I needed. ...and i don't want to fight about it. Cheers, SAM
|
|
sandyk (RIP Alex, 1939 - 2021)
Global Moderator
Posts: 226
About Me: Retired ex Principal Telecommunications Technical Officer with 43 years at Telstra (Australia)
I am also a Moderator in Hi Fi Critic Forum
Electronics hobbyist for >65 years with DIY projects including Loudspeakers, Stereo FM tuner, S/W Regen Receiver, Superhet AM ,
Synchrodyne PLL AM tuner (Phase Lock Loop),Stereo Tape Deck, Amplifiers including I.C. types, Class A, Class AB 100W/Ch. (ETI5000) 240W/Ch. Mosfet (AEM6000) ,several DACs , numerous PSUs including VERY low noise (<4uV) types etc.for myself and friends
Audio Industry Affiliation: NIL
|
Post by sandyk (RIP Alex, 1939 - 2021) on Oct 25, 2019 7:06:10 GMT 10
Sam I did not realise that you were the member that I was discussing this with at the time that I was banned with no warning in the other forum. If they had the decency to warn me first, I would have self moderated my posts there. The full details were published in several different threads in HiFi Critic Forum, which unfortunately is accessible to members only. Previous to that, I also performed a series of tests with both John Kenny and Marcin Ostapowicz from JPlay who also confirmed my reports, which caused me to contact Martin Colloms after I saw a post from him in Computer Audiophile. If you are interested, I can PM you a copy of the emails from one of the NSW E.E.s that I mentioned in my post, who later went on to demonstrate the same differences to his E.E. friend without me present. Incidentally, Anthony (acg) who is a member here has also confirmed my reports, along with David. Anthony compared a pair of Music Videos side by side using his 4K monitor and said this " Wow Alex. I looked at the first two links...big quality difference in the video. How are they different? Or, what did you change when recording them?"
Basically, one of the videos had 2 passes through a >3M long generic USB cable with a USB memory stick plugged into it. The other was generated from the same original music video after being saved to the same USB memory stick via an Uptone USB Regen plugged directly into a USB-A to USB Adaptor without +5V and shield connected through. The USB Regen was powered from a 12V Li Ion battery supplying a J.L.H PSU add-on where the 2 parallel 4,700uF electros in the capacitance multiplier area were quite low ESR types which accentuated HF detail both with A and V. No 1s and 0s were harmed in the process . Kind Regards Alex
P.S. This isn't the first time that I have reported hearing differences that others had not noticed. My findings in the DIY Audio thread at the attached link focussed more attention to the Current Mirror area in amplifiers. Nelson Pass was also a participant in the thread.
www.diyaudio.com/forums/solid-state/133018-current-mirror-discussion-15.html
|
|
jkenny
Full Member
Posts: 83
About Me: Audio equipment designer forever in pursuit of more realistic & engaging music reproduction purely because of the extra enjoyment of music created by such reproduction.
http://Ciunas.biz
|
Post by jkenny on Oct 25, 2019 8:12:59 GMT 10
My take on DBTs is that I find that if I can't hear an obvious difference in something, I'm not really interested in chasing it using a very difficult process in the form of DBT. The degree of difficulty, training & doggedness needed even for private DBTs just doesn't make sense to me.
Now don't get me wrong - some small audible differences that could only be verified by DBTs, I don't consider are worthless, I just feel that a number of these need to be accumulated together in order to result in some undeniably audible difference. So even though I might not be that interested in doing a DBT for what might be a small incremental change, I keep it in my mind & later might find that it can be added to something else I'm working on & provide a worthwhile incremental change. I also find that the more exposure we get to the sound of 'better' systems, the more capable we are of judging the sound. The corollary of this is that the less experience & exposure to these better sounding systems, the more likely we are to be stuck in a blinkered view that our playback system is as good as it gets.
Another aspect is important to bear in mind - some things that are marginally audible at one point in time may become obviously audible when other improvements have been made to a playback system - more correctly when hindrances to this audibility have been removed/reduced. One eample was mentioned above by Alex - I & others found the original Regen provided a marginal improvement to the sound - a bit more improvement when an Intona was used in front of it & a very noticeable improvement when the internal voltage regulator in the Regen was bypassed & 3.3V battery power used directly to power it.
In essence I consider DBTs as used on forums simply a debating point used by some people to win an argument - it generally has little to do with getting to the truth of matters
|
|
sandyk (RIP Alex, 1939 - 2021)
Global Moderator
Posts: 226
About Me: Retired ex Principal Telecommunications Technical Officer with 43 years at Telstra (Australia)
I am also a Moderator in Hi Fi Critic Forum
Electronics hobbyist for >65 years with DIY projects including Loudspeakers, Stereo FM tuner, S/W Regen Receiver, Superhet AM ,
Synchrodyne PLL AM tuner (Phase Lock Loop),Stereo Tape Deck, Amplifiers including I.C. types, Class A, Class AB 100W/Ch. (ETI5000) 240W/Ch. Mosfet (AEM6000) ,several DACs , numerous PSUs including VERY low noise (<4uV) types etc.for myself and friends
Audio Industry Affiliation: NIL
|
Post by sandyk (RIP Alex, 1939 - 2021) on Oct 25, 2019 8:39:56 GMT 10
My take on DBTs is that I find that if I can't hear an obvious difference in something, I'm not really interested in chasing it using a very difficult process in the form of DBT. The degree of difficulty, training & doggedness needed even for private DBTs just doesn't make sense to me. Now don't get me wrong - some small audible differences that could only be verified by DBTs, I don't consider are worthless, I just feel that a number of these need to be accumulated together in order to result in some undeniably audible difference. So even though I might not be that interested in doing a DBT for what might be a small incremental change, I keep it in my mind & later might find that it can be added to something else I'm working on & provide a worthwhile incremental change. I also find that the more exposure we get to the sound of 'better' systems, the more capable we are of judging the sound. The corollary of this is that the less experience & exposure to these better sounding systems, the more likely we are to be stuck in a blinkered view that our playback system is as good as it gets. Another aspect is important to bear in mind - some things that are marginally audible at one point in time may become obviously audible when other improvements have been made to a playback system - more correctly when hindrances to this audibility have been removed/reduced. One eample was mentioned above by Alex - I & others found the original Regen provided a marginal improvement to the sound - a bit more improvement when an Intona was used in front of it & a very noticeable improvement when the internal voltage regulator in the Regen was bypassed & 3.3V battery power used directly to power it. In essence I consider DBTs as used on forums simply a debating point used by some people to win an argument - it generally has little to do with getting to the truth of matters I agree with John. Private DBTs are a waste of time, and the results of this would never be accepted by most people suitably qualified in these areas. That's why I sought out Martin Colloms who is both qualified as an E.E. and experienced in these areas. I also agree with John about the blinkered view that our playback system is as good as it gets. Every time that I think that I am close to achieving this, something else comes along to help lift the performance even higher, especially as a result of new techniques, and way lower noise voltage regulation in the PSU areas, which also have vastly improved performance at much higher bandwidths. Ultracaps when properly implemented have also further improved results . I don't agree though that they are being used to their full potential when used BEFORE a voltage regulator though as most manufacturers are currently doing.
|
|
jkenny
Full Member
Posts: 83
About Me: Audio equipment designer forever in pursuit of more realistic & engaging music reproduction purely because of the extra enjoyment of music created by such reproduction.
http://Ciunas.biz
|
Post by jkenny on Oct 25, 2019 8:42:59 GMT 10
... It may be that this statistical aspect of auditory perception is of more importance when we are judging the 'naturalness' of our playback system? We are assimilating a 'feel' or 'gist' about the sound of our playback system over time which is stored in memory & it is this 'gist' that is compared when a new device is incorporated in our system? I reckon it could explain why we want to 'live with' a system for a while before we pass judgement on it. If this is correct, it would explain what is the biggest oversight of DBT & why long term listening can often be at odds with DBTs? Hi John, Now John, you know blind testing has no duration requirement. There are plenty of interesting things to bring up (like the paper you linked - you always find interesting papers), without creating straw men... Cheers, SAM Just to continue this aspect of the discussion from another thread - as you'll note I said "worth considering" as a possible weakness to DBTs? Just one of the many difficulties with DBTs is that, on forums, the focus tends to be on ABX testing. In other words some people insist that someone has to be able to identify a difference between A & B before they will accept a DBT preference test. This tends to force people to try to identify specific differences between A & B versions i.e some specific cue in a specific part of the music track. Forum DBTs are where I have the main issues with blind tests as regards to how they have become a tool for bullying others
|
|
sandyk (RIP Alex, 1939 - 2021)
Global Moderator
Posts: 226
About Me: Retired ex Principal Telecommunications Technical Officer with 43 years at Telstra (Australia)
I am also a Moderator in Hi Fi Critic Forum
Electronics hobbyist for >65 years with DIY projects including Loudspeakers, Stereo FM tuner, S/W Regen Receiver, Superhet AM ,
Synchrodyne PLL AM tuner (Phase Lock Loop),Stereo Tape Deck, Amplifiers including I.C. types, Class A, Class AB 100W/Ch. (ETI5000) 240W/Ch. Mosfet (AEM6000) ,several DACs , numerous PSUs including VERY low noise (<4uV) types etc.for myself and friends
Audio Industry Affiliation: NIL
|
Post by sandyk (RIP Alex, 1939 - 2021) on Oct 25, 2019 9:05:22 GMT 10
Hi John, Now John, you know blind testing has no duration requirement. There are plenty of interesting things to bring up (like the paper you linked - you always find interesting papers), without creating straw men... Cheers, SAM Forum DBTs are where I have the main issues with blind tests as regards to how they have become a tool for bullying others Spot on ! As well as demanding measurements from members who have neither the training or suitable equipment to do so. Interestingly, although some of the members demanding this do have the necessary training to do this, they don't themselves have access to suitable test equipment either.
|
|
jkenny
Full Member
Posts: 83
About Me: Audio equipment designer forever in pursuit of more realistic & engaging music reproduction purely because of the extra enjoyment of music created by such reproduction.
http://Ciunas.biz
|
Post by jkenny on Oct 25, 2019 9:16:58 GMT 10
especially as a result of new techniques, and way lower noise voltage regulation in the PSU areas, which also have vastly improved performance at much higher bandwidths. Ultracaps when properly implemented have also further improved results . I don't agree though that they are being used to their full potential when used BEFORE a voltage regulator though as most manufacturers are currently doing.
I agree that electrical noise is a neglected area as far as audio electronics & EE mindset is concerned. I find a lot of views expressed about electrical noise is very simplistic & people look no further than a low noise voltage regulator as the solution. Alex, as you know, my quest in audio electronics over many years now has been in the power supply area through battery power & lately supercapacitors. A lot of the improvements in other areas cannot be heard until the power supply is working optimally. But there are other ways that electrical noise can intrude into a system besides its PS - often the electrical connection to other devices can be a source of this intrusion - Intona, Regen & my own USB isolation demonstrated this to me. Back to DBT evaluation - based on the above, the quality of the playback system is a large determinant of the audibility of small differences & another reason why results from DBTS hold so many issues. I believe that using positive controls within every DBT would go some way towards providing an objective evaluation of the suitability of the setup & subject for the DBT in question Like a lot of other people, I pay attention to the anecdotal experience of others whom I have grown to trust or to any mass agreement of an audible improvement. The Intona was one such device that I looked into as a result of the number of reports about audible improvements. One has to be open to listening to & possibly trying new ideas in order to improve one's listening experience with their playback system.
|
|
|
Post by soundandmotion on Oct 25, 2019 16:44:30 GMT 10
Hi John and Alex,
Thanks for responding to this thread, and thanks for making clear that my OP was unclear on some critical points.
All double-blind listening tests (abbr.: DBT) are listening tests (abbr.: LT)(duh!), but not all LTs are DBTs. In my description above, I intended to include "living with" new caps, an Intona or Regen to see if there is an improvement, as an LT. My goal was to tie the procedure for the LT to the target audience to be convinced, even a target audience of one, yourself. Whether sighted, blind or double-blind, short or long listening times, short or long switchover times, if the goal is to listen to the system/equipment rather than (or perhaps in addition to) the music, I would call that a "test". If I bought a Regen, left it unconnected on top of my computer, and reported my LT results, would you be convinced, or would you complain that my procedure was lacking?
Between a publication-quality DBT in a purpose-built lab and a causal "just give it a listen at home" LT, there is a continuum of effort to control variables. It is not either/or. It is not black/white. The level of effort depends on whom you intend to convince and what the exact question is. I should mention that for every paper I've written or co-written, and for every experiment for which I was a subject, none were double-blind and no reviewer or editor has ever even mentioned double-blind, let alone rejected the work for that reason. Why? The "D" in DBT means double, i.e. 2. You blind both the experimenter and test subject when both can influence the results, even just possibly. When the experimenter is in a separate room and gives no cues on performance or test condition to the subject, "D" is not relevant. Even for publication, some tests are easy to setup and take very little time; some tests are difficult and take months.
The location is unimportant to me; the test procedure is important. For me, you can do a test along a busy highway, IF! the subject is sitting in a sound-proof booth. So I don't distinguish between home, lab, or forum LTs based on where, I care about the methods used, the how. Trust is also, unfortunately, part of the process (as mentioned by you above). I challenged a member at head-fi, who wants everyone to do controlled tests and claims he has done so with all his equipment, on the exact procedures he uses, and sadly I have to say I just don't believe him. But if you ask me what it will take to convince me, and I tell you, please don't call that a demand. I've simply answered you. If someone is not part of the target group that you hope to convince, you need not care if they are convinced.
I am an open-minded skeptic.
There is certainly more to say and I want to answer some individual points above, but I'll need to do that later.
Happy listening (test or pleasure or both)!!!
Cheers, SAM
|
|
sandyk (RIP Alex, 1939 - 2021)
Global Moderator
Posts: 226
About Me: Retired ex Principal Telecommunications Technical Officer with 43 years at Telstra (Australia)
I am also a Moderator in Hi Fi Critic Forum
Electronics hobbyist for >65 years with DIY projects including Loudspeakers, Stereo FM tuner, S/W Regen Receiver, Superhet AM ,
Synchrodyne PLL AM tuner (Phase Lock Loop),Stereo Tape Deck, Amplifiers including I.C. types, Class A, Class AB 100W/Ch. (ETI5000) 240W/Ch. Mosfet (AEM6000) ,several DACs , numerous PSUs including VERY low noise (<4uV) types etc.for myself and friends
Audio Industry Affiliation: NIL
|
Post by sandyk (RIP Alex, 1939 - 2021) on Oct 25, 2019 17:26:17 GMT 10
Hi John and Alex, Thanks for responding to this thread, and thanks for making clear that my OP was unclear on some critical points. All double-blind listening tests (abbr.: DBT) are listening tests (abbr.: LT)(duh!), but not all LTs are DBTs. In my description above, I intended to include "living with" new caps, an Intona or Regen to see if there is an improvement, as an LT. My goal was to tie the procedure for the LT to the target audience to be convinced, even a target audience of one, yourself. Whether sighted, blind or double-blind, short or long listening times, short or long switchover times, if the goal is to listen to the system/equipment rather than (or perhaps in addition to) the music, I would call that a "test". If I bought a Regen, left it unconnected on top of my computer, and reported my LT results, would you be convinced, or would you complain that my procedure was lacking? Between a publication-quality DBT in a purpose-built lab and a causal "just give it a listen at home" LT, there is a continuum of effort to control variables. It is not either/or. It is not black/white. The level of effort depends on whom you intend to convince and what the exact question is. I should mention that for every paper I've written or co-written, and for every experiment for which I was a subject, none were double-blind and no reviewer or editor has ever even mentioned double-blind, let alone rejected the work for that reason. Why? The "D" in DBT means double, i.e. 2. You blind both the experimenter and test subject when both can influence the results, even just possibly. When the experimenter is in a separate room and gives no cues on performance or test condition to the subject, "D" is not relevant. Even for publication, some tests are easy to setup and take very little time; some tests are difficult and take months. The location is unimportant to me; the test procedure is important. For me, you can do a test along a busy highway, IF! the subject is sitting in a sound-proof booth. So I don't distinguish between home, lab, or forum LTs based on where, I care about the methods used, the how. Trust is also, unfortunately, part of the process (as mentioned by you above). I challenged a member at head-fi, who wants everyone to do controlled tests and claims he has done so with all his equipment, on the exact procedures he uses, and sadly I have to say I just don't believe him. But if you ask me what it will take to convince me, and I tell you, please don't call that a demand. I've simply answered you. If someone is not part of the target group that you hope to convince, you need not care if they are convinced. I am an open-minded skeptic. There is certainly more to say and I want to answer some individual points above, but I'll need to do that later. Happy listening (test or pleasure or both)!!! Cheers, SAM Sam Briefly, because it's stinking hot in my room at present, I will mainly address the Regen issue. NOT all systems will benefit from it's use provided that the USB input of the DAC is well implemented. Earlier USB Input implementations were generic and not well designed apparently In fact, I can get almost identical results using a low noise external supply to the USB device provided that I tweak the PSU itself. However, the Regen is a convenient place to inject the quieter and better isolated PSU. Incidentally , at all listening sessions that David and myself have participated in , they have ALWAYS involved non sighted listening, with all components powered up to achieve thermal stability, as well as not giving visual cues. Although I am currently assisting John Dyson with his Dolby-A correction project, and have recently helped Barry D with the selection of the best sounding S/W for the conversion from.aiff to .wav of his new 24/192 album " Kay Sa", I no longer feel the need to convince others of what I am able to hear. Recently, a Recording Engineer (?) in A.S. forum (FrederickV) tried to make me look stupid by virtually forcing me to participate in his test that most participants failed. The test involve a 24/96 .wav file vs. the same .wav file converted to 16/44.1 and back again. I was correctly able to identify the original high res version and describe the differences. Initially, he implied that I must have been cheating. I did not need to as I heard clear differences. These days , I prefer others to hear AND see the differences between A or V files with identical check sums for themselves. If I was unable to SEE and HEAR these differences myself, I would not be able to prepare the comparison files Besides which, my hearing is quite variable from day to day, depending mainly on my BP, as I have a 21mm Diameter benign Acoustic Neuroma ( courtesy of Telstra) pressing on my right ear canal Regards Alex P.S. Perhaps more later tomorrow evening when hopefully this stifling heat goes
|
|
jkenny
Full Member
Posts: 83
About Me: Audio equipment designer forever in pursuit of more realistic & engaging music reproduction purely because of the extra enjoyment of music created by such reproduction.
http://Ciunas.biz
|
Post by jkenny on Oct 25, 2019 19:40:49 GMT 10
SAM, I consider myself experienced enough to do sighted listening & as I said if there isn't a sufficient audible difference during this listening, I'm not that interested in spending the time to do blind listening tests for very little payback in my listening pleasure.
I also engage in sighted listening with a core group of listeners from the TirnaHiFi forum & again if it's too close to call then we call it no difference. We are all experienced listeners who have seen the systems we listen to greatly improve over the years - both the DIY systems & the commercial systems. I consider this group my listening panel
I'm in a slightly different position as I manufacture audio devices & so I get feedback from many users about my devices. If they were telling me that they heard no audible improvement with my device & were returning them, I would be reviewing my listening abilities but gladly, this is not the case. So I get reports from others about what they hear when using a device of mine & they pretty much report exactly what I have heard myself. This is better than a blind test where people are not usually not listening for enjoyment in a relaxed manner. So I consider my own evaluation of the sound is validated by these reports & I don't need to convince others by presenting blind test results.
It's only on forums that I find this to be an issue & in fact, any whiff of commercial enterprise is immediately dismissed & demeaned as snake-oil & not to be trusted (IMO, it was one of the reasons behind the starting of this forum - to get away from the incessant tribalism inherent in most of the other audio forums)
In essence I now don't do blind tests of any sort - I consider them an inefficient way to progress
|
|
sandyk (RIP Alex, 1939 - 2021)
Global Moderator
Posts: 226
About Me: Retired ex Principal Telecommunications Technical Officer with 43 years at Telstra (Australia)
I am also a Moderator in Hi Fi Critic Forum
Electronics hobbyist for >65 years with DIY projects including Loudspeakers, Stereo FM tuner, S/W Regen Receiver, Superhet AM ,
Synchrodyne PLL AM tuner (Phase Lock Loop),Stereo Tape Deck, Amplifiers including I.C. types, Class A, Class AB 100W/Ch. (ETI5000) 240W/Ch. Mosfet (AEM6000) ,several DACs , numerous PSUs including VERY low noise (<4uV) types etc.for myself and friends
Audio Industry Affiliation: NIL
|
Post by sandyk (RIP Alex, 1939 - 2021) on Oct 25, 2019 20:34:43 GMT 10
|
|
|
Post by soundandmotion on Oct 25, 2019 21:57:50 GMT 10
Alex, I dug up our PM exchange from last year and reread it. I've also read your posts here and many on AS about what I file in my brain under "same bits, different sound". The reason I communicated with you is that I find it interesting. I play around with ideas for how it's possible, even if the ideas seem unlikely. But I have some sticking points. I often think that way, especially at work. I used PM because I didn't want to start a fight arising from some questions I have, which may sound confrontational, even though that's not the intent. Using pointed questions is also how we communicate at work. One issue, which seems like a simple difference of definitions, pops up often. Some people seem to use ABX, DBT, blind test and controlled test interchangeably, when they are all different in how I understand them. Take DBT, which is used very often, here and elsewhere. I was in a long battle at Hydrogen a few years ago when I said a self-administered Foobar ABX test is not a DBT. It was not a criticism at all. It was a terminology clarification. Double means 2. I need to be able to count 2 people or groups: subject(s) and experimenter(s). If there is no experimenter, NOT being a DBT is NOT a flaw or weakness. Even if a Quarterpounder with Cheese has MORE meat than a Double Cheeseburger, take off the bun and count the patties: 1 big one for the former, 2 little ones for the latter. Double doesn't mean more meat or better meat. A "blind" vs. "sighted" test has nothing to do with the eyes. It has to do with knowledge/information. You can do a "sighted" test with subjects who are congenitally blind or blindfolded, and you can do a "blind" test with eyes open and lights on. So if someone claims a DBT, I always want to know how was the switching done. If we're unclear on the meaning, then I don't get the meaning. I have more question/issues that I'll explore with you later... if you're willing. Thanks for the demo files. I can't look at the files now, but I downloaded them for later. Is that how you intended their use, or does the DropBox player need to be used? Cheers, SAM ------------------------------------------------------------------ John, It seems that you are very clear, both about whom you want to convince (primarily yourself, but also your listening panel group) and what it takes to convince you (known listening experience using sighted tests). That makes me happy. I don't think I've ever told you to do a blind test over all these years. Have I? If I ever do, smack me! Where you and I disagree is when you (from my perspective) or I (from your perspective) make "incorrect" statements about listening tests. For example, can lay people do a valid blind test at home? You say no and I say yes (depending on the question to answer), and I'm right! Just kidding. I respect your opinion. If I take issue with something you say about blind testing, I'm not telling you that you should do blind testing. And if I say that fully convincing me would include blind testing, keep in mind, first that you don't have to convince me; it's not a demand! Second, I'm not a black and white person; a compelling report of a sighted test will definitely pique my interest and might partially convince me that further exploration (by me or someone else) is warranted. Experience, to me, has a very limited value. You, and others, pointed out in the color perception thread that perception is an illusion. Stereo sound seeming like more than 2 speakers is an illusion. This is an illusion: Illusion, delusion, confusion... No matter how much experience I have (lots!) looking at these types of illusions, including confirming with a ruler, I can't see straight lines. I don't doubt it's possible to train oneself to determine if the lines are straight or slightly curved (if I created such slightly curved lines as tests), but I think it would take quite some training with feedback about the truth (using a ruler). Simply looking at the image many, many times, even knowing they are straight has never made them so for me. Cheers, SAM
|
|
jkenny
Full Member
Posts: 83
About Me: Audio equipment designer forever in pursuit of more realistic & engaging music reproduction purely because of the extra enjoyment of music created by such reproduction.
http://Ciunas.biz
|
Post by jkenny on Oct 25, 2019 23:32:50 GMT 10
John, It seems that you are very clear, both about whom you want to convince (primarily yourself, but also your listening panel group) and what it takes to convince you (known listening experience using sighted tests). That makes me happy. I don't think I've ever told you to do a blind test over all these years. Have I? If I ever do, smack me! I don't believe you have ever told me to do a blind test but we have communicated about blind tests both in forums & privately. Just to be clear I have done blind tests in the past both Foobar ABX & more casual blind A/B listening. One other group that convinces me that I am working in the right direction are the many customers that give me feedback & identify the same audible aspects I note. I enjoy our interactions & disagreements a sit helps to question my opinion & maybe refine my meaning. Your quote of my position "can lay people do a valid blind test at home? You say no" needs some refinement, I guess - I hinted at this already - I need positive controls to evaluate null results - how sensitive the test is for evaluating the audible differences. I find that this is the bit lacking in such tests. For instance, it seldom happens but if a new user of one of my devices reports that s(he) doesn't find an audible improvement, I usually investigate the grounding of their playback system & 99% of the time resolve the ground noise issue that is masking the audible improvement I know they should be hearing. In the case of Foobar ABX test null results (the usual home tests we are talking about), I have no way of evaluating their system's ability or their ability to discern the differences under test. BTW, I do consider Foobar ABX as a DBT - there are two agents involved in this test - the computer administering the test & the listener/subject doing the test - the computer doesn't tell & the subject has no knowledge if X is A or B during each trial. Lots to tease out here When I talk about the illusion of stereo, I mean that enough psychoacoustic cues are being satisfied in the listener to make for a believable version of the reality of an audio performance. This is difficult to tease out as few recordings are straight single mic recordings of a song performance. So there is an art involved in the whole recording chain i.e it's an attempt to provide a version of the performance agreed between artists & engineers/producers which will sound the way they have agreed it should sound. Usually this involves the use of psychoacoustic techniques to portray soundstage, ambience, room acoustics, etc even though these might use different tracks & have been recorded using different mics at different times. So this pleasing & realistic illusion is hopefully the final result of this process as long as it appeals to how our auditory perception expectations of real world acoustics behave. This gets complicated & we step outside of this limited definition when we are listening to computer/electronically generated sounds & effects which aren't found in nature but I would contend still follow the basic structures of sound behaviours The illusion you cite is an example of how we can probe how perception works by finding these anomalies where perception is incorrect in its analysis. When we accept that perception isn't about truth, it's about usefulness - we realise that it's design gives us a useful interface to the world that allows us to interact relatively successfully with it in order to survive & procreate. So the flaws of perception tend to be inconsequential to this goal i.e. the illusions are usually inconsequential to us & can actually be pleasing. The stereo illusion is achieved to varying degrees of success by its ability to adequately portray the cues necessary to satisfy auditory perception that this sound is not coming from two point sources (speakers) but rather is a useful portrayal of an audio performance happening beyond & between the speakers. This isn't an all or nothing illusion, it is successful to varying degrees depending on the quality of the playback system Where experience comes into play is that by experiencing the system that do this better then the direction is revealed which determine where/how your own systems should progress. By this experience you are more open to the fact that auditory perception, like all perceptions are relative - giving varying degrees of an experience rather than black & white (this last bit is somewhat rushed as I have to go out so I may return to it)
|
|
sandyk (RIP Alex, 1939 - 2021)
Global Moderator
Posts: 226
About Me: Retired ex Principal Telecommunications Technical Officer with 43 years at Telstra (Australia)
I am also a Moderator in Hi Fi Critic Forum
Electronics hobbyist for >65 years with DIY projects including Loudspeakers, Stereo FM tuner, S/W Regen Receiver, Superhet AM ,
Synchrodyne PLL AM tuner (Phase Lock Loop),Stereo Tape Deck, Amplifiers including I.C. types, Class A, Class AB 100W/Ch. (ETI5000) 240W/Ch. Mosfet (AEM6000) ,several DACs , numerous PSUs including VERY low noise (<4uV) types etc.for myself and friends
Audio Industry Affiliation: NIL
|
Post by sandyk (RIP Alex, 1939 - 2021) on Oct 26, 2019 7:12:14 GMT 10
Alex, I dug up our PM exchange from last year and reread it. I've also read your posts here and many on AS about what I file in my brain under "same bits, different sound". The reason I communicated with you is that I find it interesting. I play around with ideas for how it's possible, even if the ideas seem unlikely. But I have some sticking points. I often think that way, especially at work. I used PM because I didn't want to start a fight arising from some questions I have, which may sound confrontational, even though that's not the intent. Using pointed questions is also how we communicate at work. One issue, which seems like a simple difference of definitions, pops up often. Some people seem to use ABX, DBT, blind test and controlled test interchangeably, when they are all different in how I understand them. Take DBT, which is used very often, here and elsewhere. I was in a long battle at Hydrogen a few years ago when I said a self-administered Foobar ABX test is not a DBT. It was not a criticism at all. It was a terminology clarification. Double means 2. I need to be able to count 2 people or groups: subject(s) and experimenter(s). If there is no experimenter, NOT being a DBT is NOT a flaw or weakness. Even if a Quarterpounder with Cheese has MORE meat than a Double Cheeseburger, take off the bun and count the patties: 1 big one for the former, 2 little ones for the latter. Double doesn't mean more meat or better meat. A "blind" vs. "sighted" test has nothing to do with the eyes. It has to do with knowledge/information. You can do a "sighted" test with subjects who are congenitally blind or blindfolded, and you can do a "blind" test with eyes open and lights on. So if someone claims a DBT, I always want to know how was the switching done. If we're unclear on the meaning, then I don't get the meaning. I have more question/issues that I'll explore with you later... if you're willing. Thanks for the demo files. I can't look at the files now, but I downloaded them for later. Is that how you intended their use, or does the DropBox player need to be used? Cheers, SAM ------------------------------------------------------------------ Hi Sam We are in basic agreement here. The problem is that the phrase DBT is normally accepted, even by most E.Es as meaning that someone else controls the test procedure including the switching , but not necessarily from another location, without any cues to the participant. That is also why I mentioned all items being powered up during a non sighted test. As far as Foobar ABX goes, give me a break, Foobar ABX is a bloody joke, as the SQ from the Jack of all trades Foobar 2K is sub standard to start with, in comparison with the simplistic CPlay from System memory, and even more so with JRiver when ASIO is selected. My E.E. friend that I mentioned previously uses his own hardware design using relays. BTW, if you remember DEQX, he was employed there in earlier days and still has a DEQX processor from back then that he was involved in the design of.
Incidentally, WRT the WiFi experiment , one version looked like the TV brightness had been turned up close to maximum. Obviously, these differences are more noticed on a large screen monitor such as the 4K monitor that Anthony used , or a decent size quality TV. You will also see some references in a thread here to results from David, Chris and STC. David was also able to see the differences directly using his own 4K TV when I visited him a while back. We even spent several hours on a separate occasion trying to get good captures of his TV screen using his Pro Camera gear with Tripod and Remote control, but we had a problem with " banding" in the pictures due in part to the 200HZ refresh rate of his TV, and getting suitable exposure settings. Several people have been able to see the differences even using the lacklustre Dropbox player, but for best results they should be DL to an electrically quiet PC (not a typical Laptop) and played using a good S/W player such as JRiver 25, or even recent versions of GOM player or VLC, although the sound side of things is best compared using the superior JRiver 25 with the ASIO option enabled.
Kind Regards Alex
|
|
sandyk (RIP Alex, 1939 - 2021)
Global Moderator
Posts: 226
About Me: Retired ex Principal Telecommunications Technical Officer with 43 years at Telstra (Australia)
I am also a Moderator in Hi Fi Critic Forum
Electronics hobbyist for >65 years with DIY projects including Loudspeakers, Stereo FM tuner, S/W Regen Receiver, Superhet AM ,
Synchrodyne PLL AM tuner (Phase Lock Loop),Stereo Tape Deck, Amplifiers including I.C. types, Class A, Class AB 100W/Ch. (ETI5000) 240W/Ch. Mosfet (AEM6000) ,several DACs , numerous PSUs including VERY low noise (<4uV) types etc.for myself and friends
Audio Industry Affiliation: NIL
|
Post by sandyk (RIP Alex, 1939 - 2021) on Oct 27, 2019 8:29:40 GMT 10
Alex, I dug up our PM exchange from last year and reread it. I've also read your posts here and many on AS about what I file in my brain under "same bits, different sound". The reason I communicated with you is that I find it interesting. I play around with ideas for how it's possible, even if the ideas seem unlikely...…... Cheers, SAM ------------------------------------------------------------------ Sam I am not able to access previous PMs with you, so some of this may be repetitive. What I am about to describe here was based around the results of earlier investigations which were a combined effort of both Rock Grotto Audio Forum and Computer Audiophile members . There are also some measurements from Jim who is a USA Amateur Radio guy. This was well before John Swenson designed his USB Regen. www.computeraudiophile.com/forums/topic/13905-continuing-pursuit-of-power-supply-improvements-and-improved-dac-performance/ Just over a year ago the Newcastle HDTV channels changed over from mpeg2 to .mp4 transmissions. As my Samsung series 5000 HDTV exhibited Motion Artifacts on the new .mp4 transmissions, I started using a Teac STB for proper reception, especially from the WinTV and NBN networks. Unfortunately, they started off using way too low a bit rate, and the results were very little different to the SDTV channels, although they did try to use a little obvious EQ to try and improve this. As I also used the Teac STB to record some HDTV programs I found the results more than a little disappointing when compared with the original mpeg 2 HD transmissions . To overcome this, I made up the box in the photo for powering the USB memory stick, where the internal JLH PSU Add-on had a lower output impedance to 100kHz and higher due to capacitor selection in the capacitance multiplier section. This resulted in increased HF detail when played back, and the video quality was then almost identical to that of the original mpeg 2 transmissions.
A very interesting result, was that the recorded material looked markedly superior to the original direct .mp4 transmission before they were recorded !!! In ensuing months however, Win TV slowly increased the bit rate of their HD transmissions, undoubtedly in response to viewer complaints, and I no longer needed to use this box.
Kind Regards Alex
|
|