So, coax and big speakers with very low crossover? Did someone even make blind testing of this? I mean, how important is it compared to all the other things we should care dearly about in audio reproduction?
Yes, at least there has been research on what we prefer in sound reproduction (let's mention Toole...). Strong variations in power response is not what we prefer. Funny thing is that a lot of things we do care about (cables and caps or super material drivers) are not that essential when blind testing is involved.
But how much of a variance? I read Toole's book and know that an even power response is important, but still there has to be compromises. How did you do it, if I may ask 🙂
Yes, there is entirely too much speculation and academic posturing on this topic. If someone is at the point of being able to hear and recognise the response of the ceiling reflection then it is likely they wouldn't have the need to ask these questions or speculate on the solutions.
You mention Geddes. He uses ceiling absorption and so do I.
You mention Geddes. He uses ceiling absorption and so do I.
Not only in this topic, but in general.
Question about variances is a very good question.
As far as I know, Toole never really investigated different kind of Q factors.
Also not for the low end. The difference between a 2-way with a Q of 0.5 vs 1.2 is very obvious.
But there are also nuances.
After seeing quite a few interviews with Toole, it seems to me that he is not totally aware about those parameters.
(something he seems to acknowledge himself actually). Also was never part of their research.
@markbakk
Correct about non essential things, although a lot depends on context.
20% distortion at 1kHz is something every person is able to hear.
Erin also has shown that some local resonances can be very audible.
However, often it seems the only thing people hard focus on, and forget about the rest.
Someone asked about the differences between quite theoretical crossover topologies. When diving into the science of designing a sound field established by different sources spaced apart, such questions cannot be isolated from others, it seems to me. IOW you have to define the conditions for such a specific question.
I agree that there is no definitive answer (yet) about the relevance of an even TOTAL power response (that is including what's going on above and below the listening axis). There seems to be a good consensus that an even horizontal dispersion - or in most cases it's rather about: a uniformly increasing directivity with frequency - is desirable in order to get the least detrimental effect from (very) early reflections. To what extent the same reasoning applies to those (very) early reflections from floor and ceiling is unclear (to me at least). Studies seem to indicate that these types of reflections (in the median plane [correct English?]) would primarily result in colourations. (Whereas the lateral ones primarily influence imaging and spaciousness.)
However, when we talk about '(total) power response' it's not any more just about early reflections. Now we tend to consider ALL reflections occuring in a listening room ('reverberation'). And since these consist mainly of secondary, tertiary and-so-on refections, the direction from which they arrive no longer really matters. (Or does it?)
The psychoacoustic effect of these "later reflections" should therefore concern both imaging/spatiality and timbre. And, as I said, the radiation of the loudspeakers in the vertical plane should be just as relevant as that in the horizontal plane.
However, when we talk about '(total) power response' it's not any more just about early reflections. Now we tend to consider ALL reflections occuring in a listening room ('reverberation'). And since these consist mainly of secondary, tertiary and-so-on refections, the direction from which they arrive no longer really matters. (Or does it?)
The psychoacoustic effect of these "later reflections" should therefore concern both imaging/spatiality and timbre. And, as I said, the radiation of the loudspeakers in the vertical plane should be just as relevant as that in the horizontal plane.
Reflections from ceiling and floor induce coloration because of their timeframe: in most rooms which is the smaller dimension?
Smaller meaning short delay time you have more chance to experience comb filtering than once you are in delay time past haas effect where our brain is able to separate direct sound from the rest with ease. It is clear to anyone into room acoustic (for reproduction) and more than widely accepted as such.
The issue is not the direction from which they* happen to come ( up to a limit of course) but the fact they are specular. Once you change this ( use diffusor) you trick our brains, this is much less an issue.
I agree about your 2 last sentence.
* Early Reflections
Smaller meaning short delay time you have more chance to experience comb filtering than once you are in delay time past haas effect where our brain is able to separate direct sound from the rest with ease. It is clear to anyone into room acoustic (for reproduction) and more than widely accepted as such.
The issue is not the direction from which they* happen to come ( up to a limit of course) but the fact they are specular. Once you change this ( use diffusor) you trick our brains, this is much less an issue.
I agree about your 2 last sentence.
* Early Reflections
Last edited:
But LR vs BW, simply let you chose wether we want good directivity in either the horizontal or vertical plane. Without coax, we cant have both - or?
Or we can at least minimize it with lower crossovers and waveguides 👌
Or we can at least minimize it with lower crossovers and waveguides 👌
For Linkwitz-Riley versus even-order Butterworth, rather whether you want a 3.01 dB dip in the total power response or a 3.01 dB bump in the on-axis response.
I'm not sure to follow you about difference between horizontal or vertical plane, the kind of filter used isn't the only parameter at play. You talk about coax, but there is other way to achieve 'control' over vertical directivity.
I think i agree with MarcelvdG pov in general. There is some other details which i would check too to make a choice between them but it is application related.
Last edited:
A 3 dB bump in the on-axis is a lot - depending on the Q of course. I think I prefer the LR 4 filters in general. I'm fond of smaller midranges and waveguided tweeters in 3 ways. I could simply try and make a filter with BW instead to see if I like that better.
On the thread topic the even Butterworth and LR are no different if you EQ them globally for a flat response.
You can use the global equalization to correct for the 3.01 dB peak in the on-axis response and cause a 3.01 dB dip in the power response, or the other way around.
It could be that this discussion in itself is way too theoretical to be practical for the OP.
Looking at a schematic that is developed for a constant impedance of 8 ohm and giving it a name would be somewhat of a gamble.
What about the roll off of the driver itself? Shouldn't we at least mention that? An electrical crossover, or even an active one with a
name does not guarantee the wanted acoustical crossover. That would be the first thing I would like to make clear to the OP.
Yea, in theory the LR crossover would have a dip in the power response. But what would happen if we create (named) crossovers with on line
calculators and slap it on real drivers? I bet you'd have more to worry about than that power response dip...
Isn't it our job to warn about these kind of differences? Be it with passive or active crossovers?
Looking at a schematic that is developed for a constant impedance of 8 ohm and giving it a name would be somewhat of a gamble.
What about the roll off of the driver itself? Shouldn't we at least mention that? An electrical crossover, or even an active one with a
name does not guarantee the wanted acoustical crossover. That would be the first thing I would like to make clear to the OP.
Yea, in theory the LR crossover would have a dip in the power response. But what would happen if we create (named) crossovers with on line
calculators and slap it on real drivers? I bet you'd have more to worry about than that power response dip...
Isn't it our job to warn about these kind of differences? Be it with passive or active crossovers?