No. Different brain segments are involved, in case of "difference" there's no "pleasure" involved. Neurochemical stimulation of pleasure centers is different than rationalizing a black/white decision.
"No 'pleasure' involved..."?
Sounds like materialism. That is not very far away from being a religion in its own right. I am just saying it, I have no interest in discussing that here, but if religion is off topic, then philosophical materialism maybe should be as well. So I for one will leave it there. Don't respond.
No, it's like building a bridge with all the engineering measurements being ticked off only to find that when in use it oscillates with wind or footsteps that correlate in time & is unusable!This is one of those "intuitive analogies" that are no more accurate than saying the flow of electrical current is like water flowing down a creek
Again, no - for one thing, you are using measurements of non real world signals & stating that we should believe these represent the actual signals in the far more complex, real world music signals - sorry, does not compute. For another, you are trying to tell us that our far more complex analysis of the signals (auditory perception) is not more revealing than the above measurements. And finally, you dismiss some of the measurements as you claim they are in the inaudible range & you base this on audiology tests which again you try to extrapolate from simple tone tests to complex signals when even the auditory research recognizes this as a mistake from the pastFrom source to final transducer the information is contained in the electrical signals. Adequate measurement technology exists and claims of "hidden" or unmeasureable information have no basis rather than idle speculation by people that generally don't understand the technology. As for differences in speaker technology or cone treatments, I think that is for another discussion.
Yes, you believe it is the crux of the matter because you are stuck in the frequency/amplitude paradigm of measurements & this is not what concerns auditory processing & perception as many have said already here!The crux of the matter, even the most benign processing (such as a little a equalization) will leave large input to output differences when measured and compared in the most SIMPLISTIC sense.
It's the interpretation and analysis of the measurements not the measurements themselves. In each case propose a hypothesis and some experiments. This is hard as Dave's work on speaker IM distortion showed. There is little reward in this and I don't see a large unserved market of folks ready to pay money for this "accuracy" whatever it is.
KSTR, suggested the differential measurements needed but I didn't see you or anybody asking him to elaborate on the approach to these measurements
They all exist in the real world and are treated the same by the circuitno - for one thing, you are using measurements of non real world signals & stating that we should believe these represent the actual signals in the far more complex, real world music signals - sorry, does not compute.
No. Different brain segments are involved, in case of "difference" there's no "pleasure" involved. Neurochemical stimulation of pleasure centers is different than rationalizing a black/white decision.
At least because, as you just said, a reliable "difference" test is required before the "preference" test. Doubles the cost, also I'm suspecting (intuitively, don't have any direct experience) a reliable "preference" test requires a much larger sample to draw any statistical valid conclusion.
Preference tests of sufficient statistical rigor by definition means that there is a recognizable difference favoring one device A or B. The red herring that ABX has to be done first is illogical & by & large a forum ruse to achieve the null result that such forum tests give & thereby deny any further blind testing.
As you have also noted preference testing exercises different brain processes than ABX testing
What is widely criticized by the other camp is the "difference" tests are delivering unjustified null results. Do you have any open source references that shows "hidden mental processes" are responsible for hiding "differences" in a well designed test (and I mean, WELL DESIGNED, with controls, training, etc... to Mr. Jakob's satisfaction, whatever that means).
No.
However, there is some interesting research about how people's preferences and preexisting beliefs affect what they are willing to accept as 'well designed' research. It can be amazing to see how far people will sometimes go to find reasons to reject research they don't like, or to justify acceptance of research they do like.
Regarding the other camp and concerns about difference tests delivering unjustified null results, I have had some forum conversations with Earl Geddes about hearing research. He seems to think we have already have a pretty good idea of what about 95% of the population can hear. But, he also seems to think the other 5% has never been studied and it doing so would require the development of some new tests. Fair enough seems to me. I would interpret the need to develop some new tests to include work to determine if ABX as it exists now is as reliable as some would like to believe for use on the 5%, or if there are some problems with it that may or may not be fixable, as others seem to believe. My guess would be that to be accurate as possible for the 5%, ABX will need at least a little refinement. But, it might need much. Probably easy to make it at least somewhat better. We should just try it and move on, would be my vote.
No, complex signals are not necessarily handled by the DUT the same as simple test signals. Here's Don from AudioScienceReview saying exactly thisThey all exist in the real world and are treated the same by the circuit
A few years ago my job was designing data converters, both conventional and delta-sigma. RF delta-sigma designs are an interesting challenge. That's a polite way of saying they are a PITA with a number of quasi-hidden "gotcha's" it took a lot of work to ferret out.
Most any signal will show the noise modulation but DC or very LF signals tend to excite some of the filter modes and multitone or NPR (noise power ratio) tests will also show it (though may take a long'ish record). Impulse or step response tests, or square waves with fast edges, will show transient issues. Testing for things like this can get complicated and the causes of problems can vary (in any design, delta-sigma or not).
You also seem to have missed the message in Mallinson's presentation linked to earlier in the thread?
Last edited:
The test signals that Scott and others use are not simple
What signals? Do you know what Scott & others use that would exercise the DUT in the same way as music or is that just your belief?
Last edited:
He and BV have said what they use, and yes, I believe them
OK, it's a matter of belief, got it.
What do you think of KSTR's post citing differential measurements (using music) - not trying to extrapolate from simple (compared to music) test signals & measurements to real world usage with music signals?
Scott posted in 2016 that he was now using multitone tests on DAC modules & it sorted out sheep from goats. Before he started using these tests was he not sorting sheep from goats? What was he seeing now that he didn't before?
Why is a multitone test signal a more revealing test signal & considered closer to what we hear? Is it because it exercises the DUT in ways that simpler test signals don't?
How close is multitone test signal to music signal? Is it dynamically changing in frequency, amplitude, timing? Ar the crest factors varying within the mulitone test signal?
Is there a test signal used regularly that better matches the characteristics found in music signals?
There are many examples of engineering measurements which are found lacking when real world use is experienced.
Last edited:
There was a short discussion recently on the Blowtorch thread about doing real time difference tests with music signals and how they should be possible
Thanks, I'll have a look.
Preference tests of sufficient statistical rigor by definition means that there is a recognizable difference favoring one device A or B.
No, this is a logical fallacy. It doesn't take more than a little logical analysis (you know, of the type A -> B is equivalent to non B -> non A) to figure it out. Exercise left to the interested reader.
What signals? Do you know what Scott & others use that would exercise the DUT in the same way as music or is that just your belief?
Did you pay attention to all of Martin's presentation. Walking voltage slowly through A/D codes is hardly my idea of "the same way as music".
You know how fast things move there, perhaps Scott, Mark or Bill could tell you more
Do you mean the discussion revolving around Diffmaker from March 2018 - see KSTR's post John Curl's Blowtorch preamplifier part III
No, this is a logical fallacy. It doesn't take more than a little logical analysis (you know, of the type A -> B is equivalent to non B -> non A) to figure it out. Exercise left to the interested reader.
Here's the logic - if people statistically show a blind preference for A over B, by definition this automatically shows there is a difference, even if they can't achieve a non-null result in ABX - as you said different processing is used in both these tests
How close is multitone test signal to music signal? Is it dynamically changing in frequency, amplitude, timing? Ar the crest factors varying within the mulitone test signal?
More of the same things that are true because they sound good when stated to an audience that does not have a deep understanding of a subject but wants to believe in something.
I'm pretty sure it was more recently than that, I asked a simple (stupid) question about doing a difference measurement between input and output
Hafler did bridge null tests with real music on amplifiers, everyone complained that the results were too complicated to interpret. This was over 35yr. ago.
- Status
- Not open for further replies.
- Home
- Member Areas
- The Lounge
- Is there more to Audio Measurements?