It makes sense, if it's coherent it is. Small preferences could be made after that. 🙂
I might just have to start, take my hands of the railing sort of speak, and take it from there in the steps I find necessary. 🙂
I might just have to start, take my hands of the railing sort of speak, and take it from there in the steps I find necessary. 🙂
What needs to happen next is not a mystery, however it takes considerable time to learn it, even in a forum full of knowledgeable people....Using a without measurement kind of adjustment will get you as close as you can without going to those lengths.PSA78 said:
Now..
Take this example of time from the last post. These are phase plots for two drivers that I was crossing. I gave them complementary response slopes (not shown), so phase should be able to line up.
Next I've shown three examples where phase meets at the crossover frequency. In each case I made an adjustment to time, and to polarity. The correct one is easy to see and now they are in time.
Yes it's time consuming work more than anything I guess, well with a passive crossover it can get expensive and 'a bit' more complicated as well.
I just realized that I can work it all out in the beginning with one speaker, so I have wiring over now and will have one ready to night. I need to check the program in Sigma Studio as well, as I will use 3 channels but need them unfiltered to start measuring.
I don't believe there's any weather the next couple of days to be outside anyhow.
I just realized that I can work it all out in the beginning with one speaker, so I have wiring over now and will have one ready to night. I need to check the program in Sigma Studio as well, as I will use 3 channels but need them unfiltered to start measuring.
I don't believe there's any weather the next couple of days to be outside anyhow.
They say passive is expensive, I hadn't noticed. What can matter in the future is the other parts of doing a crossover, like designing the box.
I guess it's a matter of experience, I would have a great big bag of passive components before I was done. lol
But I have to say that it would be fun with a more traditional build in the future, but I would probably use a DSP to try different things before I order passive parts. Maybe I'll build a complete surround system in the future, it would be interesting.
I live alone as well, so no WAF to consider. 😀
And I have thought about the box (if we talk about the same thing), I don't have any grounded outlets in my house so I had to build it out of non conducting materials. So I'll have to see now if I get problems with EMI, I might have to buy shielding and install in the box.
But I have to say that it would be fun with a more traditional build in the future, but I would probably use a DSP to try different things before I order passive parts. Maybe I'll build a complete surround system in the future, it would be interesting.
I live alone as well, so no WAF to consider. 😀
And I have thought about the box (if we talk about the same thing), I don't have any grounded outlets in my house so I had to build it out of non conducting materials. So I'll have to see now if I get problems with EMI, I might have to buy shielding and install in the box.
I mean the speaker box.. the baffle, the curves. The first part of a crossover is creating the acoustic behaviour. It makes sure there is a point to joining the separate ways together and sets the place to do it.
Oh the speakerbox, yes I guess it could be explored differently in the future. I wonder how far off we are from designing things backwards, ears, room, speakerbox, advanced measurements of drivers (sound signatures), etc.
I would say we are halfway there with things like VituixCAD, Klippel analyzing etc. It would basically be what the self-adjusting/correcting FIR is doing, just the other way around.
Just a crazy thought...
I would say we are halfway there with things like VituixCAD, Klippel analyzing etc. It would basically be what the self-adjusting/correcting FIR is doing, just the other way around.
Just a crazy thought...
Back to reality here, trying to find ways to arrange the measuring...
I figure planing is the key to success, so that's why I'm bothering you here, as I rather take my time as well as looking like a fool up front, rather than later. 😛
Do I need to worry about objects that's more than 90 degree off axis for the higher frequencies?
The reason I ask is that I have a table outside (really not the most solid, but smallest footprint) that I could put on the stairs and it would give me 865mm (34") to the first step on the stairs from the bottom of the speakers (and 1200mm/47" to the bottom of the stairs).
The front door would be just behind the baffle on one side though (but a lot of open space in front).
The drivers center from the bottom are: Woofer 520mm (20,47"), Midrange 820mm (32,28"), Tweeter 950mm (37,4").
It would be necessary for me to arrange a way to move the microphone in increments instead of the speaker that most people seem to do (turntable).
I have made a tool for my router that makes it possible to route about 1 meter radius, so I could theoretically make a halfmoon cut in a piece of wood that I move a staff/stick/microphone stand in, instead of turning the speaker. Or I could just use a 1 meter stick attached to the center of the speaker at the bottom I guess...
As I see it, and please(!) correct me if I'm wrong, the drivers I use are well known/used, the design is straight forward, it's just not anything special about it.
So crossover frequencies for the different drivers isn't a great Big mystery really, as it's DSP I'm running 4:th order filters and it should sort out cone breakup much above crossover frequencies that is sensible.
So what do I have left to adjust?
SPL level/sensitivity between drivers and the fact that I'm using different amps.
EQ for baffle (and possibly uneven respons from the driver itself, but I'll never know the difference between the two at this stage).
And last baffle step compensation.
So besides a near field measurement of the woofer, how many off-axis measurements do you think it needs to get it to where the room is a bigger problem?
I could always go back, but it would help a great deal to 'know' if I should spend weeks building tools for it, or fight the elements of bad setup and pull off just a handfull of good measurements. 🙂
Ps. I did find a tutorial for basic measuring at another forum, but rather than chasing answers there I figured it could be a of future help to other beginners on this forum, as I've searched but not found that much to be honest. 🙂
I figure planing is the key to success, so that's why I'm bothering you here, as I rather take my time as well as looking like a fool up front, rather than later. 😛
Do I need to worry about objects that's more than 90 degree off axis for the higher frequencies?
The reason I ask is that I have a table outside (really not the most solid, but smallest footprint) that I could put on the stairs and it would give me 865mm (34") to the first step on the stairs from the bottom of the speakers (and 1200mm/47" to the bottom of the stairs).
The front door would be just behind the baffle on one side though (but a lot of open space in front).
The drivers center from the bottom are: Woofer 520mm (20,47"), Midrange 820mm (32,28"), Tweeter 950mm (37,4").
It would be necessary for me to arrange a way to move the microphone in increments instead of the speaker that most people seem to do (turntable).
I have made a tool for my router that makes it possible to route about 1 meter radius, so I could theoretically make a halfmoon cut in a piece of wood that I move a staff/stick/microphone stand in, instead of turning the speaker. Or I could just use a 1 meter stick attached to the center of the speaker at the bottom I guess...
As I see it, and please(!) correct me if I'm wrong, the drivers I use are well known/used, the design is straight forward, it's just not anything special about it.
So crossover frequencies for the different drivers isn't a great Big mystery really, as it's DSP I'm running 4:th order filters and it should sort out cone breakup much above crossover frequencies that is sensible.
So what do I have left to adjust?
SPL level/sensitivity between drivers and the fact that I'm using different amps.
EQ for baffle (and possibly uneven respons from the driver itself, but I'll never know the difference between the two at this stage).
And last baffle step compensation.
So besides a near field measurement of the woofer, how many off-axis measurements do you think it needs to get it to where the room is a bigger problem?
I could always go back, but it would help a great deal to 'know' if I should spend weeks building tools for it, or fight the elements of bad setup and pull off just a handfull of good measurements. 🙂
Ps. I did find a tutorial for basic measuring at another forum, but rather than chasing answers there I figured it could be a of future help to other beginners on this forum, as I've searched but not found that much to be honest. 🙂
In my oppinion (and Kimmo will disagree with me), Starting off simple and making sure the measurements you do get are good, will allow you to design a pretty good crossover. By simple, I'm talking about just doing on axis measurements and designing with those.
My acid test of whether the measurements are good or not, is to take the individual driver measurements, and also a measurement of the speaker with a crossover (can be one of your early iterations, doesn't matter) and compare the simulated response to the measured response. If there is good correlation, then you are probably doing well with the measurements, if it's way off you have a problem 🙂
Knowing how your drivers perform off axis is important for working out crossover points, and avoiding big directivity changes, but you should be able to design a competent crossover with just the basics 🙂
Don't approach it as you will be designing the final crossover on the first attempt. Look at it as a learning journey, As you are going DSP there is no expense of buying new components for each thing you change, start off simple, and get something working, then try the next level of complexity, and see what you can improve 🙂
One of the important things with doing off axis measurements is to ensure the distance from the driver to the mic remains constant. Traditionally you do this by making sure the speaker is rotating on an axis through the front baffle. When I first started I wasn't doing that and well the results weren't good 😀 I was rotating around the center of the speaker, not the baffle.
I like the idea of backwards measurement, Kind of like raytracing, I wonder if anyone has done that for sound...
Tony.
My acid test of whether the measurements are good or not, is to take the individual driver measurements, and also a measurement of the speaker with a crossover (can be one of your early iterations, doesn't matter) and compare the simulated response to the measured response. If there is good correlation, then you are probably doing well with the measurements, if it's way off you have a problem 🙂
Knowing how your drivers perform off axis is important for working out crossover points, and avoiding big directivity changes, but you should be able to design a competent crossover with just the basics 🙂
Don't approach it as you will be designing the final crossover on the first attempt. Look at it as a learning journey, As you are going DSP there is no expense of buying new components for each thing you change, start off simple, and get something working, then try the next level of complexity, and see what you can improve 🙂
One of the important things with doing off axis measurements is to ensure the distance from the driver to the mic remains constant. Traditionally you do this by making sure the speaker is rotating on an axis through the front baffle. When I first started I wasn't doing that and well the results weren't good 😀 I was rotating around the center of the speaker, not the baffle.
I like the idea of backwards measurement, Kind of like raytracing, I wonder if anyone has done that for sound...
Tony.
Sounds intriguing, what is it?
I'd add tonal balance..
Through the acoustic centre would be good, though that's not always one point. It shouldn't matter in theory, in practice you might see some inconsistencies.
I believe he's referring to moving the microphone in an arc around the speaker, as described starting in the third paragraph of #318, rather than rotating the speaker.
I have done that. Corner speakers, in instances where they need to be in-situ for example. In any case it shouldn't make a difference if done right, even as that's difficult to do.
I thought Tony was talking about tracing modes back to their source from out in the room.
I thought Tony was talking about tracing modes back to their source from out in the room.
The comment on raytracing was in referrence to computer graphics, where you mathematically define a scene and render it by sending out rays from the pov to every point in the room. reflections and refractions are taken into account. It is done backwards because then you only need to draw one ray to each pixel in the scene ie only things that would reach the eye get computed...
I may have completely missed the point 😀
edit: I was thinking of it as a way of simulating accoustics of a room and the objects in it... Though I guess now thinking about it more, rays of light are very different to varying frequency sound waves 🙂
Tony.
I may have completely missed the point 😀
edit: I was thinking of it as a way of simulating accoustics of a room and the objects in it... Though I guess now thinking about it more, rays of light are very different to varying frequency sound waves 🙂
Tony.
Last edited:
Thank you for the answer!
Yes I guess I got caught up in several things here, and should just move on sort of speak. lol
Thoughts that crossed my mind, in no particular order:
How did they DIY design crossovers when we didn't have fancy software and cheap microphones.
If some of the most expensive/well known speakers on the market measure "like crap", how important is it on an individual basis when all we have to do is to like it, or vice versa.
Doing countless installs in cars over the years (as well as some cheap DIY speakers as a teenager) where you use the same crossover no matter how you install the drivers, and all you do is EQ and reflect over that it sounds a little different.
And then reality check; Boy, you Got a microphone and a DSP, get cracking and do something, after all this time the least you can do is listen to tones. *face-palm* 😛
Yes I'll start out with on (tweeter) axis measuring to get the SPL set for each driver independently, use the simulated crossover to see how it looks together. That I can do without moving the microphone.
After that I guess I'll cross the bridge with off axis measuring, to see if I can expect problems with directivety when listening. And I'll do a near field measuring of the woofer as well.
By the time I've rearranged my living room, and positioning the speakers, I should have a fair idea of what to look for in more in depth measurements to make them sound the way I want to.
I'll see what I can get done today before the rain comes, might be a week before I can do anything outside after that if the forecast is right. 🙂
I used to calculate my filter responses using a pocket calculator, and plot them on paper. During the '90s I began writing computer programs for whatever processes and plotting I needed.
Oh pen and paper, yes I've heard of that. 😀
The kit I bought and built back in the 90's was terrible, by any standards. I sold them after a couple of months and used headphones. lol
The kit I bought and built back in the 90's was terrible, by any standards. I sold them after a couple of months and used headphones. lol
That has been discussed somewhat often in the past. And i it used a lot in audiology research.
dave
By the way, is there anyone else here that use XTZ microphone?
It seems to work now, but only with REW and there's only a universal calibration file for it.
It seems to work now, but only with REW and there's only a universal calibration file for it.
Apparently it uses a USB soundcard. Does the mic connect to it through a line level stereo input?
It's powered through the USB; so there's USB, RCA (connection for amplifier) and the cable for the microphone (3 in total going into the box).
It took some time, but the computer recognizes it as both microphone and soundcard (out) now, and it works everywhere but in their own software. lol
It took some time, but the computer recognizes it as both microphone and soundcard (out) now, and it works everywhere but in their own software. lol