And I don't see that Sighted subjective listening tests have any substance where audio is concerned.
Sighted subjective listening test seem to make most/many users happy. That aint bad 🙂 Happy is good.
Personally, I think we are masters of fooling ourselves. When I develop my own system, I don't have the stamina to test blind. I try to tell myself to be honest - don't know if I succeed at all times - probably not. On the other hand I don't harm anyone else than myself. I lean on technology rather than on guessing and random twiddling as I want to go as far as possible to the highest fidelity. When I change a component to a technically better one and I find the sound to have been worsened, I search the fault in the other components. I evaluate during a couple of weeks. This strategy has helped me to improve things in the long run.
//
Personally, I think we are masters of fooling ourselves. When I develop my own system, I don't have the stamina to test blind. I try to tell myself to be honest - don't know if I succeed at all times - probably not. On the other hand I don't harm anyone else than myself. I lean on technology rather than on guessing and random twiddling as I want to go as far as possible to the highest fidelity. When I change a component to a technically better one and I find the sound to have been worsened, I search the fault in the other components. I evaluate during a couple of weeks. This strategy has helped me to improve things in the long run.
//
Most people I know who are using controlled "blind/doubleblind" listening tests (including me) are usually listening "sighted" and try to find corrobation in the controlled experiments for the impressions of these sighted listening tests.
The procedure surely wouldn't make any sense if "sighted subjective listening" had no substance .
If you think about it, it would be waste of resources if one would rely on controlled listening conditions for each and everything.
In other the sensory tests (food for example) it is also common to ask experts and switch over to "blind" experiment only if the experts say the differences were really small.
It is a long time since I haven't read anyting about the AK4499EQ here...
Perhaps time to stop gentlemen, or carrying on in the pub?
Very sad this thread is drifting
Claude
Perhaps time to stop gentlemen, or carrying on in the pub?
Very sad this thread is drifting
Claude
ABX tests are indeed a difficult kind of tests (which was already known shortly after its invention in the 1950s) compared to other test protocols; further it delivers the answer to the question of difference which usually is the least interesting one for listener.
So, doing A/B tests instead is the more natural choice and doing these is an interesting activity as one learns a lot about his perception this way.
But most people need some time getting used to these special/artificial conditions so we shouldn't expect to find the ultimate truth within the first attempts.
Maybe all technically important was said? Hence, "the drift"...
Why don't you give it a push in the right direction?
//
Why don't you give it a push in the right direction?
//
Are you all saying I need to provide evidence to support my clam?
Really the situation is simple.
If I need to provide corroborating evidence to support my claim, then so do others.
If others don't need corroborating evidence, then nor do I.
Take your pick!
Really the situation is simple.
If I need to provide corroborating evidence to support my claim, then so do others.
If others don't need corroborating evidence, then nor do I.
Take your pick!
You constisantly miss the point.
Double blind testing is not used to confirm the null hypothesis. The null hypothesis is initially established in order to reject it. It is not logically possible to "prove" that an hypothesis is correct when that same hypothesis has aready been assumed to be correct at the start of the study.
In this respect you are correct, but not because of problems with design, but because the test wasn't designed to do that in the first place. However, after many negative results it would be reasonable to at least doubt the significance of any difference that does exist, but of course, that isn't "proof".
The aim of double blind testing is to reject the null hypothesis, and show beyond reasonable doubt that the difference is in fact audible. Therefore, double blind testing exists as a tool for people who perform subjecctive tests to verify their results. So why is it so rarely used for that purpose in Audio?
The test doesn't have to be perfect, just good enough to reveal the difference. If that difference is really "huge" and "A" sounds "way better than "B", then surely it shouldn't be difficult to design a test to reveal that difference. However, if such a test cannot be designed, then isn't it reasonable to doubt whether, in fact, the difference is "huge" and whether. "A" is really "way better" than "B"
Fine; your last sentenced illustrates one of the problems with the usage of "DBTs" wrt multidimensional evaluations in audio, but how on earth should I've missed this point - even "consistantly"?
Which is more or less the same what I've wrote before and is in quite sharp contrast to your assertion about the senseless "sighted" listening.
My hypothesis is, that the people who tried to establish this tool in the mentioned multidimensional evaluation using music as stimulus, unfortunately were more interested to show so-called audiophiles that there perception was just based on delusion than in finding the truth.
Which is kind of moving the goalpost - after for example Les Leventhal pointed to the huge risk of beta-errors in 1986 - or kind of ad hoc argument to rescue what was done before.
And of course it depends on the mutual understanding what "huge" or "big" or "of relevance" exactly is supposed to mean.
As written in other threads the experiments wrt inattentional blindness/deafness have shown that it is quite easy to provide enough distraction so that this difference (i.e. gorilla):
remained undetected by (depending on the specific conditions) 50 - 70% of the participants.
See above,it is called "subjective evaluation" for a reason and describing vocabulary reflects that. And of course we all know that, so please let's not pretend we don't.
Further, if people are designing experiments where the known confounders/distractions are considered, then usually the results give some evidence for the EUTs.
The experimenter must be able to show that his experiment was objective, valid and reliable and quite often they can't.
Fine; your last sentenced illustrates one of the problems with the usage of "DBTs" wrt multidimensional evaluations in audio, but how on earth should I've missed this point - even "consistantly"?
Because, all of your writing is written from the position that DBT is being used to prove no difference, which as I have already pointed out is pointless. However, I don't think you have ever discussed DBT from the perspective of using it to prove differences do exist, which after all is its correct usage.
Which is more or less the same what I've wrote before and is in quite sharp contrast to your assertion about the senseless "sighted" listening.
I'm not sure I can see a connection between this and what I have written. Can you explain further?
My hypothesis is, that the people who tried to establish this tool in the mentioned multidimensional evaluation using music as stimulus, unfortunately were more interested to show so-called audiophiles that there perception was just based on delusion than in finding the truth.
If the methodology is sound, then the results should be insulated from intent.
Which is kind of moving the goalpost - after for example Les Leventhal pointed to the huge risk of beta-errors in 1986 - or kind of ad hoc argument to rescue what was done before.
And of course it depends on the mutual understanding what "huge" or "big" or "of relevance" exactly is supposed to mean.
As written in other threads the experiments wrt inattentional blindness/deafness have shown that it is quite easy to provide enough distraction so that this difference (i.e. gorilla):
remained undetected by (depending on the specific conditions) 50 - 70% of the participants.
This assumes the subjects don't know what they are listening for.
I would be quite happy for the people performing the original listening tests to also be the subjects in the DBT, provided strict protocols are adhered to. Then there can be no question of the subjects missing what they are supposed to be listening for. (from personal experience it can be very unnerving, when something you thought you could hear clearly, simply vanishes under DBT!)
See above,it is called "subjective evaluation" for a reason and describing vocabulary reflects that. And of course we all know that, so please let's not pretend we don't.
Are you saying that "subjective evaluation" is not open to shared, widely understood terms of reference, that facilitate communication, because its all "subjective?
Further, if people are designing experiments where the known confounders/distractions are considered, then usually the results give some evidence for the EUTs.
The experimenter must be able to show that his experiment was objective, valid and reliable and quite often they can't.
Once again you are arguing from the perspective DBTs being used to prove the "null hypothesis", which I have already agreed is incorrect. If DBTs are used correctly (i.e. to reject the null hypothesis), then these problems go away.
(Sorry about the bold above. I missed using colour originally, and edit mode won't allow me to add it later).
Last edited:
It tests the visual capability and yet you added the word deafness. Why? Because you want to use this as part of your audio business marketing.
Don't sweat it, Naaling. He's been repeating (attached image) this gorilla visual test many times in an attempt to build a leverage in his audio theory.
It's interesting how quite those who complained about objective side repeat this and that, are.
Claude,
I am listening to my 4499 setup with OPA1656 in the I/V position.
Naturally I will not waste time in this thread but would be happy to discuss.
Ciao, George
I am listening to my 4499 setup with OPA1656 in the I/V position.
Naturally I will not waste time in this thread but would be happy to discuss.
Ciao, George
Frankly, it shouldn't be much of a search. Either everything sounds the same or not. Perhaps someone should build and test some of the designs that appear in the Japanese magazine MJ and see how that pans out.
Thanks George!
As things are, I am currently experimenting with OPA1656... and yes, I am likely to unsubscribe from this thread, so read you in another place!
Claude
As things are, I am currently experimenting with OPA1656... and yes, I am likely to unsubscribe from this thread, so read you in another place!
Claude
Why not opening a separate thread, "like my very own 4499 experience, informing on what I did and not discussing processes", and kicking all the non desirable people out of it should they turn up? That would be great as this chip deserves it.
If not, and your time permitting it of course, I would welcome more information per PM on what board you use and what mods you did... but what a pity to have to avoid what is still the best Audio forum in my eyes... Moderator?
Oh, and thanks for your post: somewhere it gives great hope again to see a reasonable person back on the topic. That changes from all the "wise boys" that don't even have the wisdom to stop arguing, remote from anyone else and from the topic...
Claude
If not, and your time permitting it of course, I would welcome more information per PM on what board you use and what mods you did... but what a pity to have to avoid what is still the best Audio forum in my eyes... Moderator?
Oh, and thanks for your post: somewhere it gives great hope again to see a reasonable person back on the topic. That changes from all the "wise boys" that don't even have the wisdom to stop arguing, remote from anyone else and from the topic...
Claude
Last edited:
Why not make a seperate thread for "the scientists"?
I was hoping to find a lot of interesting information on the 4495 here, but sadly it is overshadowed by the old and boring argument of scientifically proven facts...
It reminds me of a thread on Gearslutz concerning audible differences in mic cables. A very substantial proportion of fellow soundengineers dismissed the possibilty of such differences on forehand because they believe that ohms law is the only law.
I propose that these scientifical fundamentalists start their own threads on subjects. These threads could be coloured red, and anyone posting unsubstantiated information (not proven in the lab with at least 10 other witnesses present, and at least 5 times double checked in three-double blind tests), will get a life ban from such threads.
Meanwhile the other threads will be open for people who like to share unsubstantiated information, because they are fools that like to experiment freely, and not be bothered by the fundamentalists... They believe that freedom of thought, experiment and preliminarily conclusions (while sometimes leading to erratic outcomes), will bring them more opportunities, creativity and in the end more progress. To each his own...
Meanwhile I would suggest to just continue to add (non proven) information to this thread and ignore the noise from the science fundamentalists.
I was hoping to find a lot of interesting information on the 4495 here, but sadly it is overshadowed by the old and boring argument of scientifically proven facts...
It reminds me of a thread on Gearslutz concerning audible differences in mic cables. A very substantial proportion of fellow soundengineers dismissed the possibilty of such differences on forehand because they believe that ohms law is the only law.
I propose that these scientifical fundamentalists start their own threads on subjects. These threads could be coloured red, and anyone posting unsubstantiated information (not proven in the lab with at least 10 other witnesses present, and at least 5 times double checked in three-double blind tests), will get a life ban from such threads.
Meanwhile the other threads will be open for people who like to share unsubstantiated information, because they are fools that like to experiment freely, and not be bothered by the fundamentalists... They believe that freedom of thought, experiment and preliminarily conclusions (while sometimes leading to erratic outcomes), will bring them more opportunities, creativity and in the end more progress. To each his own...
Meanwhile I would suggest to just continue to add (non proven) information to this thread and ignore the noise from the science fundamentalists.