just finished reading an interview of Dr. Gedlee in voice coil magazine. would like a brief expansion of what he said, possibly by the Dr. himself, if available.
"...often too much midrange performance is sacrificed for an extended low end... Better... is to make the main speakers have reasonable low end ( but not extended at the sake of the mids), and then to extend the low end, when required, through the use of subs - the more the better."
i agree with the concept of matching the mains to the subs but i am curious about what frequency range(s) in the mids are affected by modeling a box for different bass extension and by how much?
thanks
blue934
"...often too much midrange performance is sacrificed for an extended low end... Better... is to make the main speakers have reasonable low end ( but not extended at the sake of the mids), and then to extend the low end, when required, through the use of subs - the more the better."
i agree with the concept of matching the mains to the subs but i am curious about what frequency range(s) in the mids are affected by modeling a box for different bass extension and by how much?
thanks
blue934
blue934 said:
I was referring to the fact that bass extension can only come with cone excursion increases. I believe that crossovers are evil things to be avoided at nearly all costs (obviously you have to have at least one IMO). This necessitates taking the woofer very high in frequency by many standards. The wide bandwidth of this choice aggravates problems related to cone excursion. So, to me, its not a good idea to extend the bass, especially since multiple subs work better anyways, if its going to muddy the mids near the crossover.
This is what I meant by the system design making nonlinearity irrelavent.
Is this clearer?
A common point of confusion is "GedLee" versus "Geddes".
GedLee is the company name which comes from Geddes/Lee, Dr. Lidia Lee being my business partner as well as my wife. My name is Earl Geddes.
Lidia had an academic publishing trail and changing names can be a problem. Also, in the Asian culture, wifes do not take on the husbands family name, although all the children do. So we found a compromise which mostly works.
GedLee is the company name which comes from Geddes/Lee, Dr. Lidia Lee being my business partner as well as my wife. My name is Earl Geddes.
Lidia had an academic publishing trail and changing names can be a problem. Also, in the Asian culture, wifes do not take on the husbands family name, although all the children do. So we found a compromise which mostly works.
gedlee said:
Earl,
When you say 'crossovers are evil', are you referring to passive only, or all crossovers? As the subwoofers will also have at least a low-pass - and should have a high-pass to prevent over-excursion (IMO
).I've definitely heard (and seen) the problem. In a supposed hi-end demonstration, using a turntable
)), it was obvious that the woofer was moving in and out in sympathy to the record movement. And it wasn't a small movement (over half an inch in an 8" woofer) which was definitely having an effect on the rest of the sound, even in a three-way. The owner, when asked why he had no rumble filter, said that it was detrimental to the sound!?! Must have been a pretty bad rumble filter that he had previously...
No, all crossovers are bad. The two sources are not coincident and the spatial location moving with frequency always causes problems. The higher the frequency the bigger the problems. (I know use coincident drivers - but that has bigger tradeoffs than the problems it solves.) Sharp filters have rapid phase shifts and deeper off axis problems. It's all a tradeoff that can be avoided by just not having the crossovers. One is all you need and no crossover should ever be in the critical 1 kHz - 8 Khz region where our hearing is hyper-accute to these kinds of problems.
I don't use "crossovers" per-se for the lows, because I have them overlap each other and the mains. And I don't HP filter the mains (I don't use a turntable so rumble is not an issue). With closed boxes that are fairly small for the largish drivers used, cone excursion is well controlled in all situations.
I don't use "crossovers" per-se for the lows, because I have them overlap each other and the mains. And I don't HP filter the mains (I don't use a turntable so rumble is not an issue). With closed boxes that are fairly small for the largish drivers used, cone excursion is well controlled in all situations.
I have one crossover at between 800 Hz and 1 kHz, depending on the system. Thats it. No other crosovers at all. One driver from 900 Hz down and the other from 900 Hz and up.
The sub setup is multi-sub with different LP, gain and phase for each sub depending on the room. There can't be any fixed settings for this since the room dominates the situation and only "in the room" can one determine these parameters. On rare occasions I use some LF parametric EQ, but never above about 200 Hz. Usuallly the LP points on all the subs end up being different and sometimes the subs themselves cover different frequency ranges, but no attempt is made to keep them from interacting with each other and the mains, in fact interaction between all of them, and the room, is the goal.
The sub setup is multi-sub with different LP, gain and phase for each sub depending on the room. There can't be any fixed settings for this since the room dominates the situation and only "in the room" can one determine these parameters. On rare occasions I use some LF parametric EQ, but never above about 200 Hz. Usuallly the LP points on all the subs end up being different and sometimes the subs themselves cover different frequency ranges, but no attempt is made to keep them from interacting with each other and the mains, in fact interaction between all of them, and the room, is the goal.
Hi Earl,
Please excuse my late entry into this thread.
Don't multiple acoustic sources covering the same frequency range make setting up a system very difficult? I understand that the frequencies covered by subwoofers have a very long wavelength, but are you not going to confuse the levels at the very least? Rolling off the lower frequencies on the smaller main bass driver would greatly reduce the distortion it creates. If your main speakers had a low enough range, there would be no reason to have a sub in the system to begin with.
Last query for you. What do you use to cover the very high frequencies? Not many tweeters have a resonance below 500 Hz and a passive at it's crossover region destroys the damping that an amplifier would normally have. This implies a much lower resonant frequency for your high frequency driver. The only way I can see getting around this is to use an electronic crossover so each driver can take advantage of the available damping from the amplifier.
-Chris
Please excuse my late entry into this thread.
Don't multiple acoustic sources covering the same frequency range make setting up a system very difficult? I understand that the frequencies covered by subwoofers have a very long wavelength, but are you not going to confuse the levels at the very least? Rolling off the lower frequencies on the smaller main bass driver would greatly reduce the distortion it creates. If your main speakers had a low enough range, there would be no reason to have a sub in the system to begin with.
Last query for you. What do you use to cover the very high frequencies? Not many tweeters have a resonance below 500 Hz and a passive at it's crossover region destroys the damping that an amplifier would normally have. This implies a much lower resonant frequency for your high frequency driver. The only way I can see getting around this is to use an electronic crossover so each driver can take advantage of the available damping from the amplifier.
-Chris
There seem to be several reasons actually and Earl has answered this one a few times. I guess it's a relatively new idea. There's an article that's been posted that has a lot of information (http://www.harman.com/wp/pdf/multsubs.pdf) but in a nutshell the ideal room location for the main speakers is not the ideal location for making even bass and having the bass coming from more than one location in the room allows you to alter how the room modes are driven so more than one sub is even better. The first reaction to this is naturally "what about localization, now the bass isn't coming from the same place as the left and right speakers?" but under 80-100 hz this apparently doesn't matter.
Setting up a multiple sub system is a constant point of discussion. In principle it should be difficult to "optimize" and there are boxes from JBL, eyc. that will do this automatically. But you'd be amazed at what you can do by ear just by listening to a LF noise signal.
The idea is like throwing a paint ballon at a canvas. The more ballons you throw the more even the coverage of the canvas will be. Sure there will be high and low spots, but on-average the more sources the smoother the response.
Rolling off the mains does not improve the situation, it only degrades it. What one wants are multiple sources of LF sound spread arround the room. Even if the mains do go low, they are fixed in location by the needs of the mids and highs. Rooms act quite differently above and below the modal transition and have to be dealt with differently.
I use a compression driver whose resonance is actually within the passband of the tweeter - not a problem for a compression driver. Amplifier damping is simply part of the crossover design so its effect on the tweeter is acounted for.
Hi Earl,
Thank you for your responses. I've just learned something about integrating LF sources. We did this with night clubs because sound quality was less important and they never did have enough bass. The approach seemed less suited for a home environment.
Some clarification if you don't mind. Please excuse me if you have covered this before.
When you attempt to drive a speaker system below it's LF cutoff point, the woofer can become unloaded (ported, B4 for me) or less damped (sealed). The distortion of that transducer is increased and you may also experience frequency doubling. My question was focused on this aspect alone. This also suggests the use of a sub woofer system to cover frequencies below cutoff.
Would you not roll off the frequencies that you are driving the higher cutoff "main speaker" in this case?
Most speakers are designed assuming amplifier damping factors over 40. Therefore, the source resistance should be dominated by connection resistances. Since a crossover works by increasing the impedance to the drivers in the transition, this reduces the damping factor to unity or lower. From what I know, drivers operating around their resonant frequency have vastly higher impedances. It's not uncommon to see impedances around 40 ohms at resonance. That is what the zobel is for in a crossover, otherwise you would still have high output in the stop band. Never the less, the driver depends on a low resistance to allow damping with a low Qes. A passive crossover destroys this possibility if resonance and crossover point are remotely close.
I'm having trouble grasping what you can do to ensure good high frequency driver damping (or any driver for that manner) using a passive crossover. The only way I know to ensure good damping is to use active crossovers and amplifiers for each driver.
I'm not trying to be funny Earl, I honestly do not understand how to reconcile these needs (roll off and good damping) using a passive network. Is it a case where one factor is simply not that important?
Thanks again, Chris
Thank you for your responses. I've just learned something about integrating LF sources. We did this with night clubs because sound quality was less important and they never did have enough bass. The approach seemed less suited for a home environment.
Some clarification if you don't mind. Please excuse me if you have covered this before.
I guess this would be easier with an electronic crossover. You can then reduce the power to each radiator to get a balance. The rest is a coverage thing - yes?
Here your experience is running counter to mine. Perhaps my question was not clear. Let's see ...
When you attempt to drive a speaker system below it's LF cutoff point, the woofer can become unloaded (ported, B4 for me) or less damped (sealed). The distortion of that transducer is increased and you may also experience frequency doubling. My question was focused on this aspect alone. This also suggests the use of a sub woofer system to cover frequencies below cutoff.
Would you not roll off the frequencies that you are driving the higher cutoff "main speaker" in this case?
That I can understand. I am using PSB Stratus Golds right now (early version). They are a two way design with a built in sub if you want to look at it that way. You might be familiar with these.
I didn't know that. I've always tried to watch my horn cutoff frequency so it doesn't "honk". In doing that, you would rarely get yourself into the resonant frequency range of the compression driver. Question then. Most horns I have seen have a lower high frequency limit as you lower the lower cutoff of the horn. This pushes the design into an upper horn, or (shutter) a piezo as a super tweeter.
I'm a little unclear on this and don't fully understand. Question then-
Most speakers are designed assuming amplifier damping factors over 40. Therefore, the source resistance should be dominated by connection resistances. Since a crossover works by increasing the impedance to the drivers in the transition, this reduces the damping factor to unity or lower. From what I know, drivers operating around their resonant frequency have vastly higher impedances. It's not uncommon to see impedances around 40 ohms at resonance. That is what the zobel is for in a crossover, otherwise you would still have high output in the stop band. Never the less, the driver depends on a low resistance to allow damping with a low Qes. A passive crossover destroys this possibility if resonance and crossover point are remotely close.
I'm having trouble grasping what you can do to ensure good high frequency driver damping (or any driver for that manner) using a passive crossover. The only way I know to ensure good damping is to use active crossovers and amplifiers for each driver.
I'm not trying to be funny Earl, I honestly do not understand how to reconcile these needs (roll off and good damping) using a passive network. Is it a case where one factor is simply not that important?
Thanks again, Chris
Anatech
What you say about woofers below cutoff is true for ported woofers, but not closed box. A closed box has a constant cone displacement below its HP, its never "unloaded" as a ported design is. I would not do this with a ported design, but I don't use ports in any of my systems.
As to driver damping, this is fully wrapped up into the frequency response. As the series resistance increases the output at the high impedance peaks - like resonance - has a relative increase in output, i.e. it is being undamped. But if one takes this into account in the crossover design and compensates for this effect then its not really an issue. I use a fairly small resistor across the Compression Driver to flatten its impedance curve. and then the whole crossover has about 16 parts, just to get it smooth and flat despite its rather irregular impedance.
The damping factor argument is more readily appied to the woofer, where we don't have components at its HP to correct for a series resistance loss of damping, but in a compression driver we do and the damping can be made to be whatever we want it to be. Rememebr that the damping and the frequency response are one and the same thing. If I correct the frequency response then I am simultaneously correcting the damping.
What you say about woofers below cutoff is true for ported woofers, but not closed box. A closed box has a constant cone displacement below its HP, its never "unloaded" as a ported design is. I would not do this with a ported design, but I don't use ports in any of my systems.
As to driver damping, this is fully wrapped up into the frequency response. As the series resistance increases the output at the high impedance peaks - like resonance - has a relative increase in output, i.e. it is being undamped. But if one takes this into account in the crossover design and compensates for this effect then its not really an issue. I use a fairly small resistor across the Compression Driver to flatten its impedance curve. and then the whole crossover has about 16 parts, just to get it smooth and flat despite its rather irregular impedance.
The damping factor argument is more readily appied to the woofer, where we don't have components at its HP to correct for a series resistance loss of damping, but in a compression driver we do and the damping can be made to be whatever we want it to be. Rememebr that the damping and the frequency response are one and the same thing. If I correct the frequency response then I am simultaneously correcting the damping.
Hi Earl,
Okay, that makes some sense to me. Don't you still increase distortion when driving a sealed system below it's resonance?
Thank you for explaining some of this Earl. I feel like a kid trying to learn how to make speakers again. I've always considered the enclosure as a high pass filter. Where you are is outside my comfort zone.
-Chris
Okay, that makes some sense to me. Don't you still increase distortion when driving a sealed system below it's resonance?
... or sealed midrange driver? One can argue that a small bass driver can be considered a midrange driver. Then you have the entire midrange product lines. I know you design 2-way systems. I do as well, or used to about 20 years ago.
This confuses me somewhat. I'll have to think on this idea.
Again, I'm a little confused here. You are looking at a system in a different way than I am used to.
Okay, now that is pretty clear. You are swamping out the Q and losing efficiency at the same time.
This is the concept I'm having the greatest trouble understanding. I'll have to do some reading to figure out how to make sense of it.
Thank you for explaining some of this Earl. I feel like a kid trying to learn how to make speakers again. I've always considered the enclosure as a high pass filter. Where you are is outside my comfort zone.
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.