Yeah, I did read what he said. The subjective comments on the bass response from your excerpt are irrelevant now as we're talking about using the full range driver as mid-tweeters in a two-way.
The bass damping issue has technical merit (AFAICT), but zero technical argument has been put forward to explain the subjectively improved high end performance.
And I did note the complaint about the crappy midrange performance also. If anything, driven with a transimpedance amplifier, the rising HF impedance of the driver would have boosted the high frequency response. Perhaps that was the reason for the subjectively improved highs. Who knows?
Last edited:
Dave,
All of this is OK with me, it's just that I wouldn't want a speaker that works fine with one amp but not with another. Or buy/build an amp that shines with one speaker but not another. This 'mix-and-match' stuff is great for playing around (and I don't mean that negatively) but 99% of the power amps around are voltage sources (or aspire to be) and 99% of speakers around are meant to be driven from voltage sources. Deviate from that and you limit your choices.
Another example: if you have a competent DAC, you should hear no difference between digital interconnects. That doesn't mean you can't 'play around' with DACs and cables that produce all sorts of audible differences. It's just not my idea of competent engineering. YMMV.
jd
You are making to many uninformed generalizations. I just use what sounds best.
Here is an example of the rising impedance of my favorite mid-tweeter.
dave
PS: you have that XO schema?
Here is an example of the rising impedance of my favorite mid-tweeter.
dave
PS: you have that XO schema?
There are no such beasts Jan. For every amp you can find a speaker it won't work with. With every speaker you can find an amp that doesn't work with it.
Whatever choice you make you are limiting yourself to a subset of the other.
If you try to maximize the set, you usually end up with homogenized & boring.
There is much to be made of synergy.
dave
LOL! That's a nice evasion. I'm commenting on the majority of the impedance and frequency response plots shown for the drivers tested in that NP paper, not your speakers, whatever they may be.
And what do the x-over schematics of any speaker I own have to do with anything?
"A rising impedance at the top and bottom of a FR can be used to advantage with a highZ amp to boost a drooping top and bottom"
Well, that agrument for high-z drive obviously does not apply to your favorite driver in mid-tweeter service.
If you can expect those 4 as a random product depends on the specific test that was used. And in this case it was an 20 trial ABX. So it depends on the probability to get 16 or more correct answers by chance. (AFAIR two got 18 resp. 19 correct answers)
As syn08 has written, there are several ways to look at the data to conclude on the distribution of the population. In our case the distribution is already known as we compare to random guessing.
The expectancy value for every result is given by the number of participants multiplied with the probability value for the specific result.
As the probability for 18 or 19 correct answers by chance is extremely low, the expectancy value is it too, as for this low probability the group of ~150 participants is still small.
Otoh to find 4 within 10000 wouldn´t be surprising.
Additional tests against the whole data set confirmed too, that the probability for this outcome due to chance was low.
Wishes
the probability of any person getting 18 correct "by chance" is around 1 in 5535
of getting 19 correct by chance, is 1 in 52632.
Eh? Well if I have been generalising on the HF impedance characteristics of full range drivers, then that specific argument of yours for the desirability of high-z drive can be generally disregarded too.
And come to think of it, the claim about reduced power compression with high-z drive doesn’t hold water here either.
Last edited:
On the face of it, a true* current-out amplifier will "ignore" the load impedance (to a limit; clipping) and actually continue to deliver the same currrent to the voice coil on high impedance peaks and resonances**, while a voltage-out amplifier will deliver less.
**After all, that's why a current-out amplifier gives you a LF boost at the resonant peak and a top-end boost with full range drivers that exhibit a rising HF impedance - more power is delivered to the voice coil when the impedance goes up.
* And that most definately does not include low-zout SS amplifiers with a big resistor bunged in series with the output to simulate "current drive".
Got any hard data on the percentage of heating due to back EMF for voltage driven full range drivers?
**After all, that's why a current-out amplifier gives you a LF boost at the resonant peak and a top-end boost with full range drivers that exhibit a rising HF impedance - more power is delivered to the voice coil when the impedance goes up.
* And that most definately does not include low-zout SS amplifiers with a big resistor bunged in series with the output to simulate "current drive".
Got any hard data on the percentage of heating due to back EMF for voltage driven full range drivers?
Last edited:
I guess the extremists that run out of the room when the "propeller-heads" arrive, are a straw man. But they do exist. I was responding to the argument that when misbehavior and frequency response deviations are eliminated it is reasonable to expect what remains is subtle at best. For me this follows from an intuitive feel for simply how things usually work gained from years of actually doing them. There are others who find science unspiritual and "boring". A healty bit of tolerance for ambiguity goes a long way.
While that conclusion sounds reasonable it depends on the knowledge as big a difference must be, to be detected in an audibility test.
It is surprising that informations about this point were given extremely seldom, yet nearly all experimentators (and readers as is offen presented here) tend to the conclusion that you expressed above.
Of course it is not exactly the same, but just to give an impression what kind of differences could be overlooked in cognitive tests, look at some of these:
http://viscog.beckman.illinois.edu/djs_lab/demos.html
Nearly all experiments are stunning and by no means any of these difference could be classified as "small" or "irrelevant" .
Does one really want to argue that nothing equivalent could happen in listening tests?
Unfortunately there are no explanations physics or engineering could give, beside that two DUTs couldn´t be tell apart by measurements, which is nearly never the case.
All other points were not addressed by physics or engineering but by psychoacoustics. And a simple but reasonable hypothesis would be that our hearing sense is just better, than one would expect from single psychoacoustic experiments done partly a long time ago.
Whenever it is deeper researched, quite often the modern experimental result is, that our hearing is (sometimes a lot) more sensitive than thought before.
And as long as one will use untrained listeners (means untrained under blind test conditions) and will omit positive controls, probability is quite high to produce negative audibility test results.
Wishes
Last edited:
Of course.
Please present some arguments for circular logic in my posts, i´ll be more than happy to correct.
I´ve honestly tried to answer your questions; last time counting there were at least 5 open questions directed to you. Maybe we could agree on "quid pro quo" ? 🙂
Wishes
You still have to show a single *******' workable example of what you call "positive control" when it comes to cables audibility testing. Namedropping 'MUSHRA' is not good enough.
You now also have to define "trained listeners" and how to select and use them in what you consider an acceptable test.
And finally to explain why none out of those promoting cable audibility (that is, at least manufacturers and salesforce) accepts to run such tests, and/or publish the complete procedures and results. That would certainly improve their credibility and would be a very good opportunity to finally 'make a difference'.
Such as Kunchur's recent result that human hearing can detect >5uS time differences.
dave
- Status
- Not open for further replies.