There is a couple of documents (already linked & talked about by Jakob2 & I in another thread) that are the recommended protocol guidelines for running perceptual testing in the auditory field.
They can be distilled down to a number of key recommendations that relate to (Jakob2 can correct any errors or omissions)
- pretest listeners/setup as to ability to clearly show what level of audible impairment can be detected using known audible impairments
- use trained listeners (the previous step will go some way towards eliminating listeners who are not discerning enough, setups that are not sensitive enough
- use hidden controls within the test to evaluate the ongoing performance of testees
But you really need to read these recommendations for yourself & judge how Foobar ABX qualifies
As someone recently said this thread would be better served by scientific inquiry into this would rather than constant debating & trying to win an argument
This has been your tactic every time I ask you for evidence - you present no evidence. Please stop this constant barrage of insults or I will take to reporting your posts
Yes, please do mention it at the time with what you think I'm misrepresenting, not just a general accusation. It may be your not following the discussion adequately, it's going over your head or people are using this tactic to back track what they said. Let's see!
Just a general FYI, there is no guarantee that any software/hardware combination under Windows is doing audio at the expected resolution and sampling rate unless you verify. I have helped several knowledgeable folks here to discover that everything was being resampled at 48 or 44.1kHz and 16 bits.
Well, I consider Widows resampling is very basic knowledge to be very honest & if anyone is doing Foobar ABX testing without knowledge of this then splutter, splutter, cough, cough - it just shows how uncontrolled this 'test' is
Last edited:
Yes, it's probably me that's at fault, funny that you can't remember any of the previous occasions
I always thought it was just a debating tactic by those saying it & you were just their echo chamber - so I don't remember - let's see
Last edited:
Yes, I have designed condenser microphone electronics too. What I was referring to was the added complexity of the measurement microphone's power supply and exotic connector requirements. Most mikes can use some sort of phantom powering, but not the measurement B&K's that we started using 45 years ago. I had to build custom 200V, and 120V power supplies, on occasion. Too much for a typical recording studio to bother with, but we did the Boston Pops with a pair (actually three) of B&K Mikes with some success.
Attachments
Last edited:
"Testees" - What is being tested? Are you testing the listeners or the claim of audibility of a defect?
This is JCs usual fallacy when discussing ABX testng, he says the he "fails", but he is not the DUT.
You are testing the perception of an audible difference - there can't be perception without a working brain (which, AFAIK, requires a person attached to it). As I said before, auditory perception is a tenuous sense, by its very nature & this is what's missing from view of what is being tested.
Again, it would do you & others well to do some reading on auditory perception - anyone with some knowledge in this field knows that valid perceptual testing needs care & attention - it's not just a computer & Foobar ABX - it's a joke to try to assert that there is any results of value - we just have too may uncontrolled variables to be able to derive any conclusions.
So, I would say it's the exact opposite of what you say - many here think, like you. the fallacy that Foobar is testing something of value
I agreed with the throwaway line earlier - it's a bit-of-fun - I would add "& nothing more"
Last edited:
I really don't know why it is so difficult for many to accept what's going on in Foobar ABX testing
Let's make it easier & use a close analogy - spot-the-difference pictures
We all know how long it takes to find the difference between two images & when we find it is seems obvious. Why? It's because we can only examine a bit of the picture at a time - we can't see it all in detail. And what keeps us trying is that we are informed that there is a difference. When the diff is found we can't look at the image without being aware of it. So what is now obvious was hidden before & was only revealed because we persevered
Now imagine you are given two images but are not told whether there actually is a difference between them - do we have the same tenacity in examining all details between the two pics.
Now imagine we are given two 3 minute movie clips & told there is a difference between them in some of the frames & asked to find any one frame with this difference.
Finally what if we are given two short 3 min movie clips but not told if there is a diff in any of the frames between them & asked again to find a frame that shows a difference. We are asked to repeat this a statistically significant number of times (usually minimum is 16)
Does this give anyone a handle on the cognitive task involved in Foobar ABX when dealing with auditory files?
Let's make it easier & use a close analogy - spot-the-difference pictures
We all know how long it takes to find the difference between two images & when we find it is seems obvious. Why? It's because we can only examine a bit of the picture at a time - we can't see it all in detail. And what keeps us trying is that we are informed that there is a difference. When the diff is found we can't look at the image without being aware of it. So what is now obvious was hidden before & was only revealed because we persevered
Now imagine you are given two images but are not told whether there actually is a difference between them - do we have the same tenacity in examining all details between the two pics.
Now imagine we are given two 3 minute movie clips & told there is a difference between them in some of the frames & asked to find any one frame with this difference.
Finally what if we are given two short 3 min movie clips but not told if there is a diff in any of the frames between them & asked again to find a frame that shows a difference. We are asked to repeat this a statistically significant number of times (usually minimum is 16)
Does this give anyone a handle on the cognitive task involved in Foobar ABX when dealing with auditory files?
Oh, I see - well in all cases I'm talking about Foobar ABX as this is the called for 'proof' requested on forums. This has been stated any number of times now by both me & Jakob2. If you want to change the discussion to ABX testing in general then my same points apply & you need to read the recommended guidelines in the documents published by the recognized standards bodies - I summarized the important ones earlier.
Your original post & this post shows you have not done this & don't understand what perceptual testing is about.
BTW, your attempt to finesse this into my 'putting words in your mouth' is pretty weak - you & scottJoplin need to try harder
Edit: If it's just a semantics issue - my understanding of the word "testee" is the person doing the listening & the "tester" is the adminstrator of the 'test'
Last edited:
I didn't say anything about ABX either, but you are certainly demonstrating some interesting points about perception. You are only seeing what you want or expect to see.
Yes, that is where you are confused. The person who is listening is testing the proposition put forward by the administrator, and therefore is the tester. If the tester is not able to reliably detect an effect, that is not a failure of the tester but of the proposition that there is an audible effect.
I believe Sinatra mainly used a U47 (Telefunken LDC), straight into the console, no compression or EQ.
They knew their $**t back then.
http://www.coutant.org/u47/sinatra.mp3
Just amazing!
It does not matter who is testee and who is tester.
It does not matter if you cannot tell the difference because your hearing is worse than
the electronics or because the electronics is better than your hearing.
G.
- Status
- Not open for further replies.