Mbare skrev:
Utgangspunktet mitt er at jeg (bombe!) er enig med Harley i mye.
...
Men han fører en argumentasjon som går langt utover en snever debatt om blindtester og som, slik jeg leser den, handler mer om hva det vil si å lytte til musikk, og som en følge av det, hi-fi.
Jeg synes ærlig talt at den debatten er et sidespor, at den er et forsøk på å skjule den reelle debatten med mange store ord, noe som veldig ofte skjer i dette miljøet når realitetene blir for vanskelige for enkelte.
Jeg overlater til komponist og utøver å skape kunst. De døde elektroniske og akustiske komponentene som står i stua mi skal ikke finne på å legge til et ekstra vers.
Most such tests, including this new CD vs. high-res comparison, are performed not by disinterested experimenters on a quest for the truth but by partisan hacks on a mission to discredit audiophiles.
Dokumentasjon? Den ene testen jeg har henvist til på tech-hjørnet var gjort av forskere ved Philips, og den tok for seg momenter relevant for SACD. Philips har vel hatt en viss interesse i SACD. Alt er peer-reviewet i AES. Partisan hacks?
I contend that such tests are an indictment of blind listening tests in general because of the patently absurd conclusions to which they lead.
En logisk brist. Dersom blindtester skal benyttes til å undersøke virkeligheten, og man legger til grunn seende tester som "fasit" så har det ingen hensikt å drive med blindtester. Blindtester som bekrefter den seende virkeligheten vil da være av begrenset interesse, mens motstridende blindtester vil bli forkastet.
Fra paperet:
Contrary to the objectivists misconceptions, much observational listening as practiced by magazine reviewers is conducted under carefully controlled conditionsmore controlled, in fact, than the conditions present during many other stages of the music recording and reproduction processes. Levels between components are matched to within 0.1dB or less. Linear differences, such as whether the product is polarity-inverting or not, are accounted for.
Vet noen i Fidelity hvordan de skal måte utgangen på en CD-spiller?
Nothing irks the scientific audio community more than reading or hearing about some new audio device or technique that countermands the laws of physics. The world of high-end audio abounds in this nonsense.
Unfortunately, objectivists attempt to discredit all observational listening by linking the entire high-end audio industry and responsible critical listeners to absurd pseudoscientific claims. This guilt by association??? technique is an affront to all serious listeners who repudiate the mumbo-jumbo and pseudoscientific elements of audio. The responsible high-end press, whose allegiance is to its readers, has a duty to expose such fraud for what it is, both in the interests of truth and to protect its readers from buying worthless devices.
Når en redaksjon slipper igjennom en artikkel som sier at en HDMI-kabel har bedre kontrast og mindre støy enn en annen så kan jeg ikke konkludere noe annet enn at de enten er uærlige eller befinner seg på en annen planet. Hvordan kan jeg da lese magasinets andre artikkel med noen tillit?
If the listener cannot identify the particular stimulus under these conditionswith statistical certainty the phenomenon is considered nonexistent.
Feil. Testen er inkonklusiv.
Unfortunately, objectivists attempt to discredit all observational listening by linking the entire high-end audio industry and responsible critical listeners to absurd pseudoscientific claims. This guilt by association??? technique is an affront to all serious listeners who repudiate the mumbo-jumbo and pseudoscientific elements of audio.
Legg merke til hvordan han først kritiserer det å avskrive seende tester "guilt by association", og så gjør akkurat det samme med blindtester:
Blind listening tests are frequently conducted under the following conditions:
1) The experimenters agenda is often to prove that no audible differences exist rather than to discover if differences do exist. (24)
2) There is an adversarial relationship between listener and experimenter, and the listener is aware that he will be exposed to ridicule if he fails.???
3) The playback system, music, room, and other conditions are foreign to the listener.
4) The experimenter controls all aspects of the test, including the music used, playback level, how long the listener can hear each presentation, how many times the listener can hear each presentation, the rapidity of switching between presentations, and in which musical passage the switching occurs.
5) The experimenter controls the number of successive trials without regard for the listeners fatigue, increasing the number of trials if a trend indicating reliable identification appears.
6) The number of successive trials is excessively high, in an attempt to get a greater statistical sample size.
Jeg er fristet til å henvise til den famøse kabeltesten av
Mike Lavigne. Pkt 1-6 var såvidt jeg kan se besvart i den testen.
To begin with the mechanical and procedural flaws of blind testing, the process is a gross distortion of reality in that the conditions present during blind testing are the antitheses of the conditions present during normal music listening. This is another example of the observation cited earlier that audio is the only field that tests its products in ways that bear absolutely no relationship to the ways in which those products are used. An individuals sensitivity to subtle differences is diminished under the stress and completely artificial conditions inherent in blind testing.
Mike Lavigne rapporterte såvidt jeg vet ikke noe stress. Deltakerne i Ivar Løkkens test burde kunne være ganske avslappet. I guidelines til blindtester er det å skape en avslappet atmosfære anført som et viktig punkt.
Alt han sier kan forøvrig sies om seende tester. Altså kan konklusjonen enten bli: "det er umulig å teste hifi (seende eller blindt)", eller "det er mulig å teste hifi".
Observational listeners employ the single presentation method??? of listening, in which a piece of music is auditioned in its entirety before listening to the same piece again through another component. This technique is in sharp contrast to the rapid back and forth switching employed in blind listening tests. Because of these procedural problems, the interaction between the listener and the test introduces unknown and therefore unquantifiable variables into the experiment. In a discipline that prides itself on recognizing and controlling all experimental variables, it is surprising that the myriad characteristics of human perception have been so studiously ignored.
Grunnen til at korte klipp ofte benyttes er fordi designeren av eksperimentet gjerne vil ha et positivt resultat. Jeg har selv utført blindtester hvor lytteren får velge fritt kuttlengde. Hvis du har tid og tålmodighet nok kan du bruke måneder og år.
During blind testing, the listeners tendency is to focus on a specific aspect of the presentation to aid him in identifying a particular component. This natural inclination to try hard??? is a fatal mistake because it is diametrically opposed to normal music-listening practice. By concentrating on the specific, the listener becomes oblivious to communication of musical expression the true indicator of audio equipment quality.
Kan det ikke like gjerne slå andre veien? Yter ikke sjakkspillere og atleter ofte best når de fokuserer? Jeg snakker ikke om "musical expression". Jeg snakker om "er det forskjell", eller "er det ikke forskjell". Noe proffe vinsmakere greier selv om det er vanskelig. De som har noe å fare med klarer det blindt og er høyt priset.
The mental activity forced on the listener by blind-testing methodology comparing, judging, calculating, trying to retrieve a memory of the previous sound, focusing on the specific, anxiety, fear of failing and of being judged, thinking ahead to the outcome, and thinking of the consequences of success or failure are all left-brain functions. These mental activitieswhich are never called into play during normal music listeningleave little room for sensitivity to how well the component conveys the musics meaning and valuethe true indicator of audio-equipment quality.
Det er fritt opp til forfatteren å designe en blindtest hvor han slipper å være redd for å feile. Heorfor greier vinsmakere å bruke sansene til å (ofte) kjenne igjen vin i en helt kunstig situasjon? Har vinsmakere bedre psyke enn audiofile
When blind listening tests, despite their effect of obscuring audible differences,
Dette har ikke blitt underbygget. Fri synsing.
When blind listening tests, despite their effect of obscuring audible differences, indicate that an audible phenomenon does exist (a phenomenon denied by objectivists), the results are either incorrectly reported as a null, (5) or judged not statistically significant.??? The disinterested??? experimenter often chooses to believe that certain listeners enjoyed an amazing run of luck rather than that they could discriminate a difference the experimenter had previously concluded to be inaudible
Jeg lurer på om forfatteren har hatt statistikk på skolebenken?
One of the strongest indictments of blind listening tests is, ironically, the very test cited by objectivists that proved??? all power amplifiers sound alike. This test revealed??? that power amplifiers as diverse as an output-transformerless tube design, an expensive solid-state unit, and a $220 Japanese receiver were sonically identical. (30) These amplifiers were as different from each other in design and measured performance as one could assemble. The amplifiers differed dramatically in their ability to deliver current to the loudspeaker, their distortion spectra, output impedance (which, with the OTL tube amplifier, would react with the load to introduce frequency response errors of 2dB), dynamic headroom, slew rate, transient intermodulation distortion, and virtually every other measurable grading of amplifier performance. Despite the large measurable differences in these amplifiers, the listeners could not distinguish among them. (31)
Tøv. Blindtester kan påvise en forskjell eller svikte i å påvise en forskjell. Det kan være mange grunner til at man svikter i å påvise en forskjell, f.eks utvalget av lyttere. Det skal ideelt sett bare være en grunn til at man påviser en forskjell: at den "virkelig" er der.
The objectivists cant have it both ways. If blind testing shows up differences very sensitively,??? yet the same methodology led to the absurd conclusion that an output-transformerless tube amplifier, a high-end solid-state design, and a $220 Japanese receiver, all having very different linear performances, were sonically identical, then the inescapable conclusion is that blind listening tests are fundamentally flawed.
Hva menes med "very different" og "very sensitive"? Jeg ser stadig konklusjoner uten argumentasjon.
How is it possible that a single listener, using non-blind observational listening techniques, was able to discover in less than ten minutes a distortion that escaped the scrutiny of 60 expert listeners, 20,000 trials, and elaborate double-blind, triple-stimulus, hidden-reference??? methodology?
Jeg har ingen problem med at enkelte lyttere er bedre enn andre, og at seende lytting kan være utmerket til å antyde problemer.
Blind listening tests subvert the entire reason we listen to music. Music isnt merely an arbitrary collection of pitches of varying amplitude and duration; it is filled with meaning, expression, and feeling
Ettersom de fleste av oss nå lytter til digitale kilder så er det snakk om en kjølig strøm av 1-ere og 0-ere.
Dersom jeg noen gang får høyttalere med følelser så kommer jeg til å sette dem på gata. Musikk er for meg store følelser. Teknikken er dumme bokser som prøver å videreformidle den "best mulig".
If the listener cares, the difference between components will be audible and musically meaningful; if the listener doesnt care, the difference may not be recognized because it has no value. This is why blind listening tests obscure differences between components; the musics meaning has no significance under such conditions.
I den grad musikken mister sin mening i blindtester så mister den også sin mening i seende tester. Betyr det at Harley ser seg om etter en ny jobb?
-k