Not really, there is a vector subtraction going on. At DC the error correction is perfect and this is reduced to zero correction at the unity gain frequency. The net result is the distortion goes down as f squared rather than f with the same no reduction at the unity gain frequency. This is a net improvement.
No need to say that, it is easy to demonstrate the effect it's just that giving it a name has been difficult. I never compared plots of w/wo HEC but that might be interesting. It was just that when a simple cap between two nodes on the bootstrapped current mirror gave this effect I was surprised, there is virtually no added circuitry to get this benefit.
It does rob headroom, and these days rail to rail input and output rules over raw performance. To be be fair the next generation of processes get nearly the same performance without much help.
Part of a good DBT test is in testing the Test. One check is when both A & B are identical samples and another is when A & B are made different in a known way.Is there a way to DBT whether DBT actually works? Just wondering?.
If you missed it Joe, or Kevin... I cannot express enough how much I think this might be meaningful to everyone on both sides of the fence.
Last edited:
You can try to drill this as hard as you want, but when you're operating on your own language about testing, don't expect anyone to take you seriously. Especially when your explanation is designed to essentially weasel your way out of any test one may be subjected to.
DBT is really, really, really simple in concept. ABX is really really simple, although there's plenty of other methods of DBT. Execution can certainly be tricky depending on what you're testing and ensuring you're actually testing what you think you're testing. And the data will obviously be noisy, making interpretation muddy. But gobblygook "beating into your skulls" speculations aren't advancing anything and says far more about the author than the "recipient" of such a rant. There's a certain level of "alls you can do is alls you can do" that needs to be respected.
DBT is really, really, really simple in concept. ABX is really really simple, although there's plenty of other methods of DBT. Execution can certainly be tricky depending on what you're testing and ensuring you're actually testing what you think you're testing. And the data will obviously be noisy, making interpretation muddy. But gobblygook "beating into your skulls" speculations aren't advancing anything and says far more about the author than the "recipient" of such a rant. There's a certain level of "alls you can do is alls you can do" that needs to be respected.
I guess this is directed at me?
My reason for not deploying any tests is because I can't even get someone to come over to tell me if they like my new amplifier design, let alone a group for testing. Many, many people here have much more willing participants. (And it's not as if I wrote a manual on my idea for testing and presented it, btw, just the concept)
The results would either show a positive or negative ability to associate auditory recognization with indicators. You have to run a number of them to find out which types of associations may work, but if there is any at all it shows the ability to have auditory recognization with indicators, that would dispel the validity of all the other tests, explaining why we get nothing but null results. And it can be double checked by simply performing a traditional test prior to the indicator test - and like all of them reported on this forum they'll show a null response.
Your explanation is either we're all too stupid to execute a test that you call really really simple, or we're all too stupid to realize we have convinced ourselves of the difference between all things audio for nearly a century. Very encouraging... Perhaps that sort of belief in humans explains your rude reply.
Perhaps I missed the thrust of your point in terms of your proposed test, which got lost in the whole, "people obviously can hear things they cannot describe". Which is both true and an accepted limitation that (hopefully) allowing the tested individuals time to train before testing mitigates.
I did quite well object to the idea that diligent scientists who regularly use DBT methodologies to determine preference and thresholds of perception need to have "things drilled into their head" that are apparently obvious to you. Perhaps you didn't mean it to sound like "you all are a bunch of idiots, here's how it really is", but it sure came across like that.
Insofar DBT means double blind test. And ABX is a protocol where 2 identified things are presented and then a third is presented and the person is to determine if it's A or B. That's all I meant in terms of it being conceptually simple. I'm clearly not understanding what sort of protocol you're proposing with its association component.
And yes, generally speaking, we're finding out the *un*reliability of humans more than the opposite when we've dug deeper. Which is also great to understand.
I did quite well object to the idea that diligent scientists who regularly use DBT methodologies to determine preference and thresholds of perception need to have "things drilled into their head" that are apparently obvious to you. Perhaps you didn't mean it to sound like "you all are a bunch of idiots, here's how it really is", but it sure came across like that.
Insofar DBT means double blind test. And ABX is a protocol where 2 identified things are presented and then a third is presented and the person is to determine if it's A or B. That's all I meant in terms of it being conceptually simple. I'm clearly not understanding what sort of protocol you're proposing with its association component.
And yes, generally speaking, we're finding out the *un*reliability of humans more than the opposite when we've dug deeper. Which is also great to understand.
Daniel, my point about drilling hard enough is just for recognization. Like planting seed perhaps. It has nothing to do with intellgince, only attention. I think there are a lot of brilliant minds in this thread. But even someone who is brilliant can overlook, well, anything.
You are the first person I recall to even respond to the idea. I think arguements for and against indicator testing would have been a more constructive post, but I am glad to have any response. Negative attention can be worth it, over appearing non-existant.
There is a consistency in people having expectations in audio, to listen for something. This is also true in our dailys. When you do an audio test you eliminate associations of all kind, which has a secondary effect to remove references to tie an expectation to. I am proposing to use very strong indicators that do not reveal the variables themselves. Maybe the testing will reveal that when people can form expectations they can prevail those expectations as the mechansim for rememebering audio differences in order to not get a null.
The whole process is predicated on that we have a audio memeory in a sense, but isnt one that records sound like how we can recall an image, but rather a set of other senses that form expectation for our response to the sound. This relates to how strongly the neuro wiring is implanted in that when we hear something it can replay other emotions and memories as the reverse.
If you also think about musicians, they often know what comes next. But few could write it out, had they not seen it before hand. If they learned it by names they may recall chords. But when you watch a band practice you see them searching for things they expect, not specifically the chord. Their language is fairly revealing, especially as they sound out stuff. They describe what follows what, a list of expectations, not chords. This is anecodotal but I have collected a good number of observations of exactly thus. Anecdotes can lead to hypthosis, that is fair.
You are the first person I recall to even respond to the idea. I think arguements for and against indicator testing would have been a more constructive post, but I am glad to have any response. Negative attention can be worth it, over appearing non-existant.
There is a consistency in people having expectations in audio, to listen for something. This is also true in our dailys. When you do an audio test you eliminate associations of all kind, which has a secondary effect to remove references to tie an expectation to. I am proposing to use very strong indicators that do not reveal the variables themselves. Maybe the testing will reveal that when people can form expectations they can prevail those expectations as the mechansim for rememebering audio differences in order to not get a null.
The whole process is predicated on that we have a audio memeory in a sense, but isnt one that records sound like how we can recall an image, but rather a set of other senses that form expectation for our response to the sound. This relates to how strongly the neuro wiring is implanted in that when we hear something it can replay other emotions and memories as the reverse.
If you also think about musicians, they often know what comes next. But few could write it out, had they not seen it before hand. If they learned it by names they may recall chords. But when you watch a band practice you see them searching for things they expect, not specifically the chord. Their language is fairly revealing, especially as they sound out stuff. They describe what follows what, a list of expectations, not chords. This is anecodotal but I have collected a good number of observations of exactly thus. Anecdotes can lead to hypthosis, that is fair.
Last edited:
Okay, cool, my apologies on misunderstanding where you were going with your post and belligerency, and to be fair, I've read it a couple of times and I still get confused by the colors description. But onwards we go.
If I were to make an analogy, are you proposing something like allowing someone a ball that they could squeeze in some self selected fashion during training/familirization then, when blinded for the dbt, they could cue off a subconscious reaction?
If I were to make an analogy, are you proposing something like allowing someone a ball that they could squeeze in some self selected fashion during training/familirization then, when blinded for the dbt, they could cue off a subconscious reaction?
You have the right idea. I was thinking visual, but maybe giving them a physcial object could be a great way to deploy a segment to evaluate which kind if indicator helps. You would want to try multiple balls primarily. If a singular one worked better that would be interesting...
So let us say you have two balls. You let the person play with one, during each of two variables, swapping as you go. Then after a bit you play one of the two variables and ask them which ball they think they just heard. You could play it twice, once with each ball, and ask them which one it was. Obviously you need repeated positive results with multiple people if it is 50/50. But you can expand with 3 variables too. There is a number of ways to question the person, along with decieving them. For example you could introduce a third variable and ask them if it is ball 1, 2, or if they dont recognize it.
I still like the idea of a large color, like construction sized paper for a classroom. That way they have to close their eyes to not have to it along with the experience (of each color and variable).
So let us say you have two balls. You let the person play with one, during each of two variables, swapping as you go. Then after a bit you play one of the two variables and ask them which ball they think they just heard. You could play it twice, once with each ball, and ask them which one it was. Obviously you need repeated positive results with multiple people if it is 50/50. But you can expand with 3 variables too. There is a number of ways to question the person, along with decieving them. For example you could introduce a third variable and ask them if it is ball 1, 2, or if they dont recognize it.
I still like the idea of a large color, like construction sized paper for a classroom. That way they have to close their eyes to not have to it along with the experience (of each color and variable).
Last edited:
Encouragement for personal insult, post decline of tension - and beginning discussion. Very mature.
I'd have thought you might have some constructive criticism, Jan. It appears that it is considered more appropriate to view me as a pest, and pointless to converse with. It feels like the old days are creeping back in where DIYaudio is nothing but a place for experienced people to tell everyone they're an idiot without a why. I thought we had a nice run for a few years were that wasn't so popular.
I'd have thought you might have some constructive criticism, Jan. It appears that it is considered more appropriate to view me as a pest, and pointless to converse with. It feels like the old days are creeping back in where DIYaudio is nothing but a place for experienced people to tell everyone they're an idiot without a why. I thought we had a nice run for a few years were that wasn't so popular.
For a bit of fun A Binaural Video Report: the Sound of PS Audio’s IRS Vs | Stereophile.com
With headphones you can hear what a pair of IRS-V sound like*. I can't help thinking that PS audio need a bigger room for that monster though!
With headphones you can hear what a pair of IRS-V sound like*. I can't help thinking that PS audio need a bigger room for that monster though!
Not sure if the "really really simple in concept" line helps. Flying to the moon is conceptually quite simple too, but execution can be difficult.
As you said above, being able to show that one is testing what he wants to test isn´t simple anymore and even more important is to clearly express what hypothesis (research question) the experiment should address before starting to think about the design of the experiment.
@ Destroyer OS,
what you probably mean is that transfer of information to long term memory works better if more categorial processing is involved.
That does not mean that you need a complicated test protocol to follow (it is known that a A/B - comparison most often gives the highest proportion of correct answers) but simply spending enough time for learning to identify the effects, getting used to the specific conditions of the test protocol to follow (speedskater already mentioned usage of positive and negative controls; nice to see that those ideas are now more sort of common knowledge
) .Keeping it as simple as possible is usually a good advice; if you later want to explore the underlying mechanism further additional tests could be run.
- Status
- Not open for further replies.