1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

FLAC vs. 320 Mp3

Discussion in 'Sound Science' started by icedup, Sep 7, 2011.
22 23 24 25 26 27 28 29 30 31
33 34 35 36 37 38 39 40
  1. sonitus mirus
    An ABX is not purely subjective.  Its purpose is to remove the subjective part.
    No, of course nobody should care about the competency of a failed test, as this is simple to do.  The person that passes the test should be able to let others look at the data that was used.  If I were able to successfully pass a similar test, you can bet that I would be passing around these files for people to verify that what I did was legitimate.  I hear claims of science and facts, but when a scientific method is asked to be used, nobody wants to participate.  I'm only suggesting this as in the past I've seen similar results only to find that the files were not being volume matched correctly, or they used an older version of the encoder, or had a setting used that would not create the best quality.  Let us check, what could it hurt?
    My ultimate goal is not to try and discredit anyone.  I'm trying to remain objective and to discover the truth for myself, with as much confidence and accuracy as I can.
  2. chewy4
    To be considered as actual evidence towards anything it does need to be. No need to verify if a 5/10 or 12/30 test was done properly because it doesn't prove anything either way.
    Yes, I think it was an honest test. But there are many things people can do wrong without meaning to skew the results. I know that the default for my converter is at medium quality for LAME mp3's regardless of the bitrate, for example. You need to change the quality slider in order to get the best out of it.
  3. julian67

    No, the purpose of the blinding is to remove bias.

    Bias and subjectivity are not the same thing!

    It's a subjective test because the result depends on the senses/consciousness/perception of the tester. If he perfectly executes the abx test and produces a result, it may be a different result obtained by you or me also perfectly performing the exact same abx test. The same abx test perfectly executed by different persons may produce identical results, or different results, or no results at all. In each case the test is always subjective but is not biased.

    An objective test is one where the result is not variable according to the senses/consciousness/perception of the person performing it. The perfectly executed objective test will produce the same result irrespective of the person performing it. That's what makes it suitable for peer review.

    What is biased is declining to assume the same good faith in others that one expects to be accepted in oneself. Also biased is impugning the competence of others when there is no good reason to do so. Another bias is to make unreasonable conditions when encountering evidence which doesn't correlate to one's own expectation (expectation can also be read as....bias!).

    I have a suggestion: if you doubt deadlylover's competence, execution, integrity or anything else then send him a pm and ask him if he will post some short clips of the lossless files and the lossy.
  4. sonitus mirus
    I may have fallen victim to the improper usage of the term objective with regards to ABX testing.  I'm certainly no English major, and I am not the only one that has made this connection.
    If someone claims to have found evidence of life on Mars, even if they seemed to have followed specific, accepted procedures to come up with their findings, the next logical step is to have other, capable people attempt to repeat the results and analyze the processes used.   I don't see where bias comes into play.  Skepticism is an important part of the modern scientific method.  The process is not meant to be taken personally.
  5. julian67
    This thread contains an awful lot of posts that run, in essence, something like "I didn't hear a difference in my abx test(s), therefore the files will also sound the same to any other person in any other circumstances."

    Assuming good faith(!) then that kind of conclusion/assertion seems to arise from misconstruing bias and subjectivity as one and the same. This leads to believing the test results meaning something other than is the case.

    In this context this is especially important when considering null results. If one mistakenly assumes a single user abx test to be an objective test and not just an unbiased test then the null result assumes significance beyond one's own circumstances and perception. The empty space acquires an illusory weight and mass which appears universal. Why can't everyone see it? They're wilfully blind! (sorry for the pun).

    Someone sincerely believing they have demonstrated or established an objective truth may in good conscience feel very frustrated with people who don't agree, and might ascribe that dissent to defects in intelligence, character, competence, experience and so on. They might also feel entirely justified in trying to persuade people that they have some good news which will benefit everyone.

    Let us pray.
  6. julian67

    No, it isn't. There is no requirement or expectation that a different person should obtain the same results! That is absolutely fundamental. It's a subjective test that depends on the tester's senses and consciousness. If file a and file b are the same file then everyone ought to produce no result, but outside of that circumstance different results say nothing about the integrity of the test.

    The process can be examined but in this case the tester described his process. It's the same process known and accepted by anyone who is interested in these kinds of abx tests. The tester also describes his files and even which minor version of lame he used.

    If you read the thread it is obvious there is no evidence of anything except a competent tester honestly performing a known process and finding it extremely difficult to distinguish file a from file b. On some comparisons he produces no result, which is to say he cannot find a difference. On others he finds a difference but states it is so tiny that he doesn't find it makes any difference to his normal listening and he will continue to enjoy his mp3 files. There is absolutely nothing to suggest any user error or bad faith.

    To read that thread and then decline to accept the author's tests unless they are held to a logically absurd standard is not a reasonable position to adopt.

    But the tester is a current member. You could put your doubts to him.
  7. sonitus mirus
    Perhaps we need a process created?  
    1. ABX and get 13 of 15 correct. (Hey folks, look what I have found!)
    2. Make files available for others to analyze and test.  (Does everything look ok, can anyone else see what I see?)
    3. Someone else properly encode the same files and have the initial tester perform the test with the new files. (Let me try this again with your data.)
    4. Can this tester repeat this test using a different platform/computer? (Lets take my computer system out of the equation)
    Even if nobody else in the world is able to pass the test in step 2, step 3 and 4 should remove any doubt that the initial tester is capable of identifying the files correctly.
    It's not perfect, and anyone could probably cheat, but at least this is a step beyond simply posting an ABX log.  Usually step 2 proves something is incorrect with the process.  I've never seen any testing beyond this point.
  8. chewy4
    Why are you taking this so personally Julian?
    The request for evidence is pretty normal in science. It's nothing personal. It's just about wanting to know something rather than just believe it. Nobody is calling anybody an "idiot liar".
  9. xnor
    To me it seems to be common that many people who post in this forum don't even know what science is. So no surprise they haven't heard of "peer review" or "falsification" or an idea as simple as repeating an experiment ...
  10. sonitus mirus
    Yes. 12 of 15 is what is normally accepted, my apologies.  My steps were not meant to be taken as gospel  Pray indeed.  Who is the fanatic in this discussion?
  11. Dillan
    I don't know, sound science is very different from science in general. I think it really isn't fair to say, "Well you heard a difference, I must also hear a difference." Yes it is good to ask questions, and use the same samples from the people who make their claims, but to have a negative tone and pretend like their making things up really isn't called for. Not to mention, some people have very admirable hearing. Just an example of differences, I myself find that my own hearing can change from time to time. Not even to scrape the surface of how separate peoples hearing is different at any given point of time with any given set of persons.

    Look. The point is, it doesn't matter how many times you ask someone for evidence in their claims, sometimes you have to accept the fact that you can either test their claims and call them a liar, or do the right thing and admit that perhaps someone else situation - whether it be their superior hearing, or perhaps their gear made the difference - provided them with different results.

    In the end a few examples aren't going to prove anything. You can either settle on what you believe, or accept that there are possibilities.
  12. julian67
    @sonitus mirus

    It isn't me that wants to discard a test with a valid statistical method because it allows for valid results which don't confirm my bias.

    It's not me who is taking it personally. I accept that maths does work.

    What I've done in the last few posts is point out an error of such magnitude that it completely invalidates every assumption that rests on it, that error being the belief that subjectivity and bias are the same thing (also expressed as the belief that a subjective test is actually objective). It isn't possible to have any meaningful understanding of the abx tests with that error in place. It really is that significant and not some minor semantic side issue.

    Discarding the valid method while misunderstanding the test that uses it, all in order to propose a new test with invented bogus standards, which can be faked in exactly the same way as the original test by anyone acting in bad faith, does not meet the description of being "pretty normal in science"!

    To crown this by characterising objections as "taking this..personally" almost has a kind of unknowing glory.
  13. Dillan
    Besides, I think I've made my point clear on my situation. I do not care about space or portability, so I chose FLAC, because with that I have my piece of mind, I can sit there listening and knowing it will sound as good as it can. Otherwise I am the kind of person who worries that maybe it wouldn't sound as good if its an MP3. Maybe that is why some people get passionate about this topic, because their situation calls for MP3s and they can't really make the choice otherwise. The fact of the matter is a couple of people on an audio forum aren't going to prove any scientific breakthroughs. So trying won't do anything. I can understand if you just want to undergo research so that you have a basis for your own opinions, but I chose a long time ago that I can easily just choose the lossless file, and not have to worry about anything. I don't have to defend myself, I can rest easy - Because I listen to my music the best way in my situation. Sometimes the easy, safe way is the best way.

    If your situation calls for something different, who cares? That is what you have to deal with not anyone else. Your research can be done on your own, it's your ears and your opinion.
  14. sonitus mirus
    I don't want to make anything up or pretend.  We are clearly not even in the same book, much less on the same page in our thinking. I respectfully bow out.  I suppose an honor system on the validity of an ABX test is as good as I can expect.  I really believe there are differences between the best lossy and lossless files that some people can distinguish between with consistency.  I also believe that life must exist somewhere else in the vast universe.  I am simply looking for tangible proof that is more satisfying that what I have now.
  15. julian67

    I think it's the person who rejects sound method on finding it allows valid results he doesn't like.

    Or it's the person who presumes dishonesty or incompetence in people who use the valid method and produce results he doesn't like.

    Or it could be the person who fundamentally misunderstood what the test is and what its results mean but seeks to replace it with a method more complex, less likely to be completed (or even used), which requires multiple participants to achieve the same results as a single user test, and whose results can be falsified in much the same way. All this to try to avoid results that might not match a particular expectation.

    That is fanaticism.

    I still think in the context of an online discussion it's better to assume good faith unless there is a substantive reason not to do so. If you don't assume good faith it doesn't get you anywhere in an environment where people are not accountable. Assuming bad faith and trying to hold people to unusual standards (especially ones of dubious worth) also invites people to regard you in that same light. If you've made up some whacky rules and stacked the deck and called people names because they debated you it doesn't work out too well unless you're the board admin or moderator and can do magic edits and secret deals/threats/sanctions (I'm not referring to this board, just thinking of a couple of different places that I've seen run that way).
22 23 24 25 26 27 28 29 30 31
33 34 35 36 37 38 39 40

Share This Page