How do you get good imaging?

[SY]

Jim, by doing that you also change the frequency response, so beware of that unintended variable.

[jim1961]

Originally Posted by SY

Jim, by doing that you also change the frequency response, so beware of that unintended variable.

Quite right. I will have to see how it affects other things. Not only could it affect freq response, but also the ETC data.

Remarkably, everything is connected. If I have learned anything, it is that if you change one thing, many other things can change also. Often times, I have to make a choice or compromise.

Case and point:

I think the starting point in any room is determining speaker and listening locations. If this is wrong, its hard to make anything right later.

Given that I started here as well, I observed that the listening position in terms of frequency response wasnt the best place in terms of controlling early reflections.

I would also get substantial changes in the freq response itself moving the listening position forward or back a mere 2 or 3 inches. Sometimes I had to choose whether to deal with a lull at 800hz vs 500hz, and so on.

With reflections, it was the same song. A few inches forward might diminish a reflection at 2k, while making one louder at 700hz, and so on.

I am not going off on this too far in this thread, but parameters such as the height and angle of your ceiling cloud, the thickness and position of your diffraction and absorption panels also affect freq response and reflections as well, also forcing still more choices.

So in my experience, you compromise, you prioritize, and you do a lot of listening at each and every juncture as to not rely on readings alone.

[SY]

If you want to be a total madman, have different crossovers for the tilted vs nontilted, each optimized to give the flattest on-axis response for the driver and listener position. Now reverberant response will still be different, but at least on axis response won't be.

[DrDyna]

Originally Posted by SY

If you want to be a total madman, have different crossovers for the tilted vs nontilted, each optimized to give the flattest on-axis response for the driver and listener position. Now reverberant response will still be different, but at least on axis response won't be.

Oh, a total madman would be having the speakers set on linear actuators attached to a circuit that varied the crossover based on the exact angle you set the actuators to

[jim1961]

Originally Posted by SY

If you want to be a total madman, have different crossovers for the tilted vs nontilted, each optimized to give the flattest on-axis response for the driver and listener position. Now reverberant response will still be different, but at least on axis response won't be.

I intend to use a parametric Eq to adjust quickly different configurations as to optimize each. Obviously, if you build a speaker to sound best on axis, and you then play around with the toe-in or lean the speaker to where it is off axis, you might be tempted to reach the conclusion that all configurations where the speaker is on axis to the listener are the best ones. So, I do indeed plan to be a madman.

I have already done this in regards to playing with toe-in. I change the angle I re-measure everything. If I change the height of the speaker, I remeasure everything. If I move the speakers closer together, I remeasure everything, and so on.

Eventually though, it is my hope to dial in the crossover into whatever physical configuration works best after everything I can think of has been tried. This leads to what I feel will be a very controversial conclusion. But here it is.

I plan to dial in the speaker to the listening position criteria. That is, while getting a flat speaker response at 1m at 5 or 10ms is a general aim, the more specific aim is to get things flat (or how I like them) at the listening position. For the listening position in all its reverberation (room reflections) is what I will be hearing, not from 1m away and a controlled gate time. If I have to color the speaker a bit to accomplish this, I am not only willing to, but prefer to, as opposed to hearing unwanted coloration at the listening position or having to rely on EQ to compensate for a flat speaker in a unflat room. Now, most would say, well fix the room. Well, its just not possible, thats why. Perhaps on a unlimited budget and a dedicated room built the ground up with this in mind might get close, the average (my room) wont.

Ive gone a bit off here but I think it very much relates to the thread topic, for the room is critical for good imaging. Certainly, there is more than one right answer, and more than one approach. Again, this is only mine. And as I learn new things, my approach my change. But this is where I am at based on what I know, or what I think I know.

[jim1961]

Originally Posted by DrDyna

Oh, a total madman would be having the speakers set on linear actuators attached to a circuit that varied the crossover based on the exact angle you set the actuators to

Do you have one? I may want to buy it

[DrDyna]

Originally Posted by jim1961

Do you have one? I may want to buy it

Oh, I wish. I'm sure if I did, I'd have a bunch of folks from around here buying plane tickets to come check it out, lol.

[daniel]

I plan to dial in the speaker to the listening position criteria. That is, while getting a flat speaker response at 1m at 5 or 10ms is a general aim, the more specific aim is to get things flat (or how I like them) at the listening position. For the listening position in all its reverberation (room reflections) is what I will be hearing, not from 1m away and a controlled gate time. If I have to color the speaker a bit to accomplish this, I am not only willing to, but prefer to, as opposed to hearing unwanted coloration at the listening position or having to rely on EQ to compensate for a flat speaker in a unflat room. Now, most would say, well fix the room. Well, its just not possible, thats why. Perhaps on a unlimited budget and a dedicated room built the ground up with this in mind might get close, the average (my room) wont.

Sorry to say this (someone has to?), but you are doing it wrong. All the reflections from various room surfaces superimpose with the direct signal to create spatial lobing, witch manifest in the freq. domain as comb filtering. EQ-ing the direct signal is futile fighting this issue. Move an inch and you have different comb-response.
What you need is a source illuminating all the surfaces consistently - no offaxis anomalies from speaker and controlling the early arriving high gain reflections. This gets you undistorted images, good intelligibility and low coloration. Later reflections don't cause image smear/coloration if they are dense, don't protrude too much from the overall response and you utilize a Haas kicker in the 10-20ms region. Remember - there are no absorbers in a RFZ room. (well except the inherent absorbtion in diffusers)

Also you have very early reflections 3ms after direct signal. Agree with everyone else to first zoom at the first 20ms of the individual sources ETC and don't band-limit! (caveat you need a speaker with consistent offaxis performance and treatment that is effective over the whole bandwith above Schroeder freq.)

[jim1961]

Originally Posted by daniel

Sorry to say this (someone has to?), but you are doing it wrong. All the reflections from various room surfaces superimpose with the direct signal to create spatial lobing, witch manifest in the freq. domain as comb filtering. EQ-ing the direct signal is futile fighting this issue. Move an inch and you have different comb-response.
What you need is a source illuminating all the surfaces consistently - no offaxis anomalies from speaker and controlling the early arriving high gain reflections. This gets you undistorted images, good intelligibility and low coloration. Later reflections don't cause image smear/coloration if they are dense, don't protrude too much from the overall response and you utilize a Haas kicker in the 10-20ms region. Remember - there are no absorbers in a RFZ room. (well except the inherent absorbtion in diffusers)

Also you have very early reflections 3ms after direct signal. Agree with everyone else to first zoom at the first 20ms of the individual sources ETC and don't band-limit! (caveat you need a speaker with consistent offaxis performance and treatment that is effective over the whole bandwith above Schroeder freq.)

By RFZ, what I meant was keeping the reflections at the listening position to -15db or less for the first 20ms if possible. I am using a good amount of 4" Owens Corning 703 at the wall and ceiling 1st reflection points to minimize combing. Thick rugs on the floor.

As far as band limiting. I found running individual ETC tests at several frequencies to be more revealing than one overall ETC across the whole spectrum. I DO cover the entire 250-3.6K range, but in 1/3 octave pieces. Their is no reason for me to measure ETC above 3.6k because I dont have any reflection issues above this freq.

The 3ms early reflection you see I believe is a cabinet issue. I am in the process of modifying the cabinets, in part, to try and address this. It may also be due to driver overlap whereby the woofers and mids (both active at the test frequency I illustrated) having different impulse / time arrivals. This issue also I have some ideas on how to address.

[gedlee]

Originally Posted by daniel

Sorry to say this (someone has to?), but you are doing it wrong. All the reflections from various room surfaces superimpose with the direct signal to create spatial lobing, witch manifest in the freq. domain as comb filtering. EQ-ing the direct signal is futile fighting this issue. Move an inch and you have different comb-response.
What you need is a source illuminating all the surfaces consistently ...

Sorry to say this (someone has to?), but you are doing it wrong.

What is required, and I think that Jim does have this mostly correct, is a flat direct field, that is reflection free (or at least surpressed) for 5-10 ms at least. a

source illuminating all the surfaces consistently

will not do this and will in fact make the situation far worse. This is completly the wrong thing to do. You must deal with the early reflections through a narrow directivity in the source that illuminates the listener with a flat direct signal, mnimizes the early reflections through its narrow coverage angle, and has a flat power response so as not to color the reverb field.

I am not sure that I can agree with Jim when he says:

Certainly, there is more than one right answer, and more than one approach.

because I have not found that to be the case. I have found that there is a fairly narrow range of design choices that lead to fufilling the criteria stated above. Jim stated his goal early on and that was to have good imaging with good spaciousness, which is my goal too. But he alluded to this being "having his cake and eating it too". But the fact is that this is not a tradeoff situation if done correctly.

The classic approach of using uncontrolled directivity in a room with high damping will not do it that much is certain, but a narrow controlled directivity speaker in a room which is very live forward of the loudspeakers will. The early reflections will be surpressed by the speakers directivity (good imaging) and the spaciousness will come from the high reverberation in the back portions of the room. Problem solved - except that there are not a lot of options for the speaker or the room design in this scenario. But it does work and it works well as years and years of doing this have shown.