Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Threshold shifts with blind testing? (Read 9624 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Threshold shifts with blind testing?

An anti-ABX (but very knowledgeable!) person on another forum asserted that studies have shown that non-audio sensory stimuli can lower audibility thresholds.

The implications to this for ABX testing are obvious and large, but so far he hasn't responded to my requests for references. Has anybody heard of this phenomena?

Threshold shifts with blind testing?

Reply #1 –
Here might be a way to test:  Have a spectrogram display running during ABX test.  If you audition "A", show the spectrogram for "A", if you audition "B", show the spectrogram for "B".  But, when "X" is played, always show the spectrogram for "A" regardless of whether "X" is "A" or "B"

Threshold shifts with blind testing?

Reply #2 –
What kind of non-audio stimuli ? Generic (closed eyes vs open eyes) or specific (ex: spectrogram) ?


It might be linked to sensory integration.
I haven't studied it w.r.t. audio perception, but did study and experiment with kinesthesic and visual senses.

Basically the experiment, as I recall it, went like this:
Subjects must learn a new motor skill, repeat the motion a few thousands times (!), demonstrate skill in test and then rest, then transfer test.
Condition 1: Learns motion with visual feedback. Test with visual. Transfer test without visual
Result: Removing visual impairs performance... duh!

Condition 2: Learns motion without visual feedback. Tested without. Transfer test with introduction of visual feedback.
Result: Overall, worse performance than group 1 (because no visual). However, inside group 2 (transfer test vs ordinary test), adding visual feedback makes performance *worse*.

In other words, there is specificity of learning with regard to which senses were used. Both adding and removing sensory information impair performance.
This phenomenon is not present in children. Only adults. (With children: the more the better. Always perform better with vision, even if learned motion without)
I alsho hypothetize that its importance becomes greater in advanced (experienced) people than newbies. Experts (/masters) would be less influenced than advanced, but still more than newbies. (like a bell curve)

I think the spectrogram test proposed by benski would be most evident if performed with audio mastering engineers, for example, who are used to see the sound as much as hear it.


Another complementary explanation might be the optimization of the activation level. Too much activation (stress) --> worse performance. Test it with a stroboscope or similar apparatus. I'm not sure about the specific mechanism involved. Perhaps when dealing with ultra-low amplitude sounds, self-noise (heartbeat, respiration, etc.) becomes an issue, and it tends to increase with stress. Also more difficult to concentrate. One issue with ABX being listener fatigue. Not enough activation, as in without visual perception (eyes closed), also worsens performance. Cognition is less efficient when underactivated, processing takes more time, fatigue kicks in sooner (in relative terms), worse performance.

Finally, an unrelated note, I recall reading that if you put two identical sound systems, integrate them in a home cinema, one with cheap CRT TV, the other with nice plasma TV in properly lit room, people will invariably say the second sounds better than the first. Only because of the better image !

Threshold shifts with blind testing?

Reply #3 –
Quote from: Axon on
An anti-ABX (but very knowledgeable!) person on another forum asserted that studies have shown that non-audio sensory stimuli can lower audibility thresholds.


Yup, like consuming enough beverage to get your BAC up to say 0.3%. ;-)

I'm sure that there are a ton of other non-audible influences that could distract many listeners. How about a certain glance from your significant other? ;-)

Quote
The implications to this for ABX testing are obvious and large, but so far he hasn't responded to my requests for references. Has anybody heard of this phenomena?


I've only got a problem if the claim is made that these other influences affect *only* ABX.  That's usually the implication that the Hate-ABX fearmongers try to make.


Threshold shifts with blind testing?

Reply #4 –
How does this person reply to the fact that DBT's can show response of the human ear within ->||<- of basic physical limits?

Any "lowered thresholds" would break the Shannon Theorem or various signal detection theorems in a very basic way.
-----
J. D. (jj) Johnston

Threshold shifts with blind testing?

Reply #5 –
Quote from: Woodinville on
How does this person reply to the fact that DBT's can show response of the human ear within ->||<- of basic physical limits?

Any "lowered thresholds" would break the Shannon Theorem or various signal detection theorems in a very basic way.


Um... how do psychoacoustic results have anything to do with Shannon's theorem? And what "physical limits" would be violated?

Threshold shifts with blind testing?

Reply #6 –
Quote from: Axon on
Quote from: Woodinville on
How does this person reply to the fact that DBT's can show response of the human ear within ->||<- of basic physical limits?

Any "lowered thresholds" would break the Shannon Theorem or various signal detection theorems in a very basic way.


Um... how do psychoacoustic results have anything to do with Shannon's theorem? And what "physical limits" would be violated?


I would have to wonder how anyone, no matter what they are looking at, could lower the audibility threshold below the physical limits of hearing -- the ultimate difference of 'heard' vs 'not heard'.  That may be was Woodinville is getting at in the first sentence.

Now, if your correspondent is only talking about signals safely above those thresholds, that changes the argument.

Threshold shifts with blind testing?

Reply #7 –
Quote from: krabapple on
I would have to wonder how anyone, no matter what they are looking at, could lower the audibility threshold below the physical limits of hearing -- the ultimate difference of 'heard' vs 'not heard'.  That may be was Woodinville is getting at in the first sentence.

Now, if your correspondent is only talking about signals safely above those thresholds, that changes the argument.
A psychoacoustic threshold is not a physical limit. There is nothing in the idea of a frequency limit of human hearing that relates to the physical properties of the ear. In fact, to rely on such limits as justification here seems a little circular.

Now, if by "physical limit" jj means "the physics of the human ear can be proven not to be able to do this with numerical certainty", then sure. But I don't think that was what he meant.

Threshold shifts with blind testing?

Reply #8 –
Quote from: Axon on
An anti-ABX (but very knowledgeable!) person on another forum asserted that studies have shown that non-audio sensory stimuli can lower audibility thresholds.

The implications to this for ABX testing are obvious and large, but so far he hasn't responded to my requests for references. Has anybody heard of this phenomena?
So I take it this person agrees that the more control needs to be applied to listening tests?

Of course not. I don't see how you get an anti-ABX support out of this line of argumentation. ABX goes out of the way to try to remove the critically skewing biases generated by non-audio sensation.
elevatorladylevitateme

Threshold shifts with blind testing?

Reply #9 –
Quote from: Axon on
Now, if by "physical limit" jj means "the physics of the human ear can be proven not to be able to do this with numerical certainty", then sure. But I don't think that was what he meant.



You are, I think, aware that the noise due to the discrete nature of the atmosphere is about 6dB SPL, give or take, white noise, at the eardrum.

Given the primary integration time of the ear, which is at best 200 milliseconds, and the bandwidth of each critical band, we can very quickly discover the detection limits (mathematically).

Guess what? Both Fletcher and Stevens come within ->||<- of it for signals around ear canal resonance.
-----
J. D. (jj) Johnston

Threshold shifts with blind testing?

Reply #10 –
Quote from: Woodinville on
You are, I think, aware that the noise due to the discrete nature of the atmosphere is about 6dB SPL, give or take, white noise, at the eardrum. Given the primary integration time of the ear, which is at best 200 milliseconds, and the bandwidth of each critical band, we can very quickly discover the detection limits (mathematically). Guess what? Both Fletcher and Stevens come within ->||<- of it for signals around ear canal resonance.
That's swell. What does mathematics have to say about the theoretical limits of temporal masking? About frequency detection limits outside canal resonance? About natural distortion modes inside the ear? About loudness contour shifts?

Threshold shifts with blind testing?

Reply #11 –
Quote from: Axon on
Quote from: Woodinville on
You are, I think, aware that the noise due to the discrete nature of the atmosphere is about 6dB SPL, give or take, white noise, at the eardrum. Given the primary integration time of the ear, which is at best 200 milliseconds, and the bandwidth of each critical band, we can very quickly discover the detection limits (mathematically). Guess what? Both Fletcher and Stevens come within ->||<- of it for signals around ear canal resonance.
That's swell. What does mathematics have to say about the theoretical limits of temporal masking? About frequency detection limits outside canal resonance? About natural distortion modes inside the ear? About loudness contour shifts?


Eh? We're talking about a place where we show that DBT's do not elevate thresholds.  With that, we can show that at least one kind of listening does not suffer threshold impariment in a bind test.

-----
J. D. (jj) Johnston

Threshold shifts with blind testing?

Reply #12 –
Ah! Gotcha. That statement is a conclusive counterexample against the notion that a pervasive threshold shift in blind testing cannot exist. (Of course, arguing that the shift occurs in some situations and not others is ad hockery).

Threshold shifts with blind testing?

Reply #13 –
Quote from: benski on
Here might be a way to test:  Have a spectrogram display running during ABX test.  If you audition "A", show the spectrogram for "A", if you audition "B", show the spectrogram for "B".  But, when "X" is played, always show the spectrogram for "A" regardless of whether "X" is "A" or "B"


I'm sure that would mess up some people. Of course they could cheat and do the test blind! ;-)