Maty, please don't make inflammatory statements about deafness. It's an exaggeration and can make people with perfectly normal hearing think you are brainless.
See what I mean? It's not nice.
See what I mean? It's not nice.
Pavel, the idea for using forced choice is that the research subject may have to rely on gut feelings arising from System 1. That should produce some statistical result that would show there is some difference to be heard and some preference one way or the other, even if the preference is not strong enough for System 2 to be aware of it. Otherwise, the statistics should show random preference.
Last edited:
All -
what do you think about an idea to post one more test sample that was recorded through another part that is not involved in the poll. Is it a good idea, or it might just spoil the test and I should wait until the test is closed tomorrow afternoon?
what do you think about an idea to post one more test sample that was recorded through another part that is not involved in the poll. Is it a good idea, or it might just spoil the test and I should wait until the test is closed tomorrow afternoon?
I'm not sure tbh. Do you mean simply another opamp used in the identical set up ?
It might be a little late in the day to get meaningful responses. You could perhaps post it later and perhaps with just one of the current files to compare against.
Its up to you though 🙂 You are in the driving seat.
It might be a little late in the day to get meaningful responses. You could perhaps post it later and perhaps with just one of the current files to compare against.
Its up to you though 🙂 You are in the driving seat.
I will post it after the poll is closed. Just one file, that can be compared with the existing ones, if anyone would be interested to do so and to find if it matches by listening to some of the already posted ones.
An experimenter can´t know if particpants really aren´t detecting something or only believe they can´t detect it.
The idea behind the forced choice is indeed to overcome this problem. If the participants are really only guessing then the effects will probably average out although no "tie" (i.e. "no difference/preference") answer is provided.
But if they only believe to hear "no difference" then the forced choice scheme probably will show an effect.
Wrt additional sound samples, what do you think about a "straight wire" version that could be used as a reference?
I know that this is often suggested, but I am not sure if it would be a good idea to include such sample into the test. IME most detectable (please note I do not say how) differences belong to linear issues, like slight differences in the LF amplitude response (coupling capacitors in the D/A -- A/D chain, phase shifts) and some members tend to perform sophisticated file analysis that discover these small differences and they then vote according analysis results. There were several of my tests in the past here that were destroyed by this analytical approach and following discussion in the thread. This happened to Karl (Mooly) as well. I took a break then, but decided to return with these new opamp tests, but very carefully re music samples data, I tried to make it as difficult for the "file analysts" as possible.
Anyway, I will think your recommendation over.
It seems to me that in a visible poll like here, people tended to vote for the longest slider, in this case for the "uu". I am not sure if there is any rational behind.
What I want to know from these test is the correlation between my subjective feeling and the measurement. I wish PMA would post the measurement after the poll ends.
Pavel, Not sure what you mean by "longest slider."
Sometimes people choose may choose, for example, the first or last item in a list. To average out such effects, it is necessary to randomize the order of choices for each participant.
Hopefully, it is clear that designing a good test or research project is nontrivial. Often researchers from multiple disciplines are needed in order to minimize possible errors.
EDIT: Okay, looks like were are referring to the incremental poll results. Yes, people are often influenced by what other people think.
Sometimes people choose may choose, for example, the first or last item in a list. To average out such effects, it is necessary to randomize the order of choices for each participant.
Hopefully, it is clear that designing a good test or research project is nontrivial. Often researchers from multiple disciplines are needed in order to minimize possible errors.
EDIT: Okay, looks like were are referring to the incremental poll results. Yes, people are often influenced by what other people think.
Last edited:
And, a few people probably voted for "I can not hear a difference" just to see the result, I guess.
At the lower right of the poll it says "View Poll Results." Clicking on that will show the results without voting.
It is possible that people were attracted to the appearance of uu for some reason. Have to experiment to find out. But, if they are, it could be because system 1 is probably quite aware of the pun uu = You You. And System 1 is always looking out for references to one's self. Wouldn't be surprised if there were some effect like that, the literature is full of similar findings.
Another brief question: If people prefer the sound of a particular single opamp, would they necessarily prefer the same opamp if there were 10 in series?
What about the same opamp in a different circuit?
What about the same opamp in a different circuit?
I would change this to "would they prefer it in a circuit that is sensitive to EMI interference"?
It's all fun, you will see tomorrow.
- Status
- Not open for further replies.