Ok, so I now finished my amplifiers and played the first music through the speakers. And it is a letdown. When I'm close to the speakers, about a meter or so, then it is all fine. But once I get halfway the room I have just a bit of tinny mid left. If I play it loud (neighbourg disturbing loud) then it is ok. But I rarely play that loud.
Room is the living room of the appartment: 7.5m x 4.1m x 2.5m, Speakers are along one of the 4.1m walls, back about 0.3m from the wall. Room has a wooden floor with thin rugs under the sofa's and table. 2 closed bookcases and 2 cupboards. One of the 7.5m walls has 2 windows with curtains over them in it.
Speakers are my interpretation of the Visaton Monitor 890. I had the TL-16H and (DR-45 + M300) already. But I could not see myself build those octogonal bass cabinets. So I build some rectangular boxes of 90l and put two pieces of Visaton W-250S in them.
Made them active, each speaker an LM3886 (eight in total) and used a Behringer DCX2496 as crossover. Used 4-th order at 800Hz and 5000Hz (like the Monitor 890). And put in delay on the bass and tweeter as that midhorn driver sits about 350mm back.
Now I have no clue how to start debugging this. I see the following points as a possible source:
I'd like to measure a bit but I'm not sure that REW with its swept measurement is going to give me good results inside the room.
My feeling goes to the room being the issue. Before I went with these speakers I tried a couple of Fostex FE-208EZ but that wasn't a success either. A bit the same in fact, ok-ish close, dropped away before you even got halfway the room, better at high volume. There is no space to put a couple of sub's.
Room is the living room of the appartment: 7.5m x 4.1m x 2.5m, Speakers are along one of the 4.1m walls, back about 0.3m from the wall. Room has a wooden floor with thin rugs under the sofa's and table. 2 closed bookcases and 2 cupboards. One of the 7.5m walls has 2 windows with curtains over them in it.
Speakers are my interpretation of the Visaton Monitor 890. I had the TL-16H and (DR-45 + M300) already. But I could not see myself build those octogonal bass cabinets. So I build some rectangular boxes of 90l and put two pieces of Visaton W-250S in them.
Made them active, each speaker an LM3886 (eight in total) and used a Behringer DCX2496 as crossover. Used 4-th order at 800Hz and 5000Hz (like the Monitor 890). And put in delay on the bass and tweeter as that midhorn driver sits about 350mm back.
Now I have no clue how to start debugging this. I see the following points as a possible source:
- the speakers are just too small for the room (ok, don't think it but I can't exclude it firsthand)
- the bassbox is "wrong": too small or too large or reflex pipe not ok (I just followed one of those online calculators, they all seemed to agree)
- the box isn't damped enough. I put felt mats 15mm thick against the sides in the upper 2/3-th and a wedge of rockwool in the lower 1/3-th. Won't be easy to change that
- the room just sucks any bass out. Possible, there is stuff everywhere, mostly books 🙂
- I have a phase issue giving me a hole in the upper bass, lower mid. Could be but.
- that 800Hz crossover is too high for the W-250S and too low for the DR-45. I'm missing a midrange and go 4-way.
- those horn mid and high are playing tricks. Because they are beaming the sound changes much more at a distance.
- I'm just not used to the sound. I spend most of my day behind the pc listening to a couple of PMC DB-Ss next to my monitor. (old non-active version)
I'd like to measure a bit but I'm not sure that REW with its swept measurement is going to give me good results inside the room.
My feeling goes to the room being the issue. Before I went with these speakers I tried a couple of Fostex FE-208EZ but that wasn't a success either. A bit the same in fact, ok-ish close, dropped away before you even got halfway the room, better at high volume. There is no space to put a couple of sub's.
Can you download an app for your phone to measure the audio output in the room, a spectrum analyser type thing that has an averaging or RTA function. Others will suggest suitable Apps's I am sure.
Could you listen with the mid and tweeter channels muted to verify your perception of the bass and Room issues.
As you unmute the mids does that upset the balance, or is it the tweeters.
Getting the overall balance right between the differing sections is crucial. You have implemented Xover frequencies, but have you verified gain or attenuation on the channels being correctly configured for the different sensitivities of the drivers?
As you have DSP you should be able to reach a useable compromise with a little more investigation and work.
Good luck.
Could you listen with the mid and tweeter channels muted to verify your perception of the bass and Room issues.
As you unmute the mids does that upset the balance, or is it the tweeters.
Getting the overall balance right between the differing sections is crucial. You have implemented Xover frequencies, but have you verified gain or attenuation on the channels being correctly configured for the different sensitivities of the drivers?
As you have DSP you should be able to reach a useable compromise with a little more investigation and work.
Good luck.
You haven't applied any baffle step compensation? That might be the answer.
If you only crossover those 2 bass drivers at 800Hz, and haven't applied any other filter, you probably have -3 to -6dB loss somewhere in low mid region.
If you only crossover those 2 bass drivers at 800Hz, and haven't applied any other filter, you probably have -3 to -6dB loss somewhere in low mid region.
That's the key of the whole design.
Once the bass section is ok, which I think it can be done, no!? Two 25 cm woofers ( 4 in total) should overwhelm the room with bass frequencies...
So, once the bass section is ok, maybe you should look at what comes out of some recent threads, such the 'Budget Monkey Coffin" or similars, or go on your own.
Two woofers in parallel would reprezent a 4 Ω load...
First if all, is your baffle the same width as the original project? You only wrote 90 litre rectangular.
If you have narrower baffle, than this makes part of problem.
Are speakers originally itended to be corner loaded?
You won't be able to solve the problem without measuring. Here is a thread on a Visaton forum. Someone had simmilar problems.
https://forum.visaton.de/forum/galerie-der-projekte/29221-visaton-monitor-890-mk-iii
If you have narrower baffle, than this makes part of problem.
Are speakers originally itended to be corner loaded?
You won't be able to solve the problem without measuring. Here is a thread on a Visaton forum. Someone had simmilar problems.
https://forum.visaton.de/forum/galerie-der-projekte/29221-visaton-monitor-890-mk-iii
I tried to start REW but it doesn't start anymore on the pc. Another thing to check.
With mid and high muted I find it a bit "meagre". When the mid is added, voices are added. Adding high doesn't make much change.
No, no compensation so far.
Correct but see 1: REW is failing at the moment.
Well, they do not overwhelm the room, on the contrary.
The woofers are not in parallel, they each have their own amplifier.
The part the speakers themselves are mounted on is the same. But I do not have the "wings" that in the original slope away at 45°. Don't think those were intended for corner loading. as the BR port is at the front they were likely intended for "close" to the wall use.
I'll go through the german forum posts but that will take a while as my german isn't very good.
Well, it seems that your interpretation has a little naive approach. I read 2 x Visaton TIW 300, which are 400 € beasts.
Then there's the compression driver with that ugly horn and a horn supertweeter. The power should be in the 500 W range.
Now, you choose to put two economic speakers (Visaton W-250S ) and with less total Vd than TIW 300, and a lesser motor strenght. Even if the system is active, the bass and the mid-treble sections are not balanced.
Just a read of the recent Monkey Coffin threads should put you in the right direction, even if designing goes in circles...you see, choose a 12" woofer then the midrange should not change abruptly the kind of emission ( dispersion at crossover frequency ) and the same goes for the tweeter.
Then there's the compression driver with that ugly horn and a horn supertweeter. The power should be in the 500 W range.
Now, you choose to put two economic speakers (Visaton W-250S ) and with less total Vd than TIW 300, and a lesser motor strenght. Even if the system is active, the bass and the mid-treble sections are not balanced.
Just a read of the recent Monkey Coffin threads should put you in the right direction, even if designing goes in circles...you see, choose a 12" woofer then the midrange should not change abruptly the kind of emission ( dispersion at crossover frequency ) and the same goes for the tweeter.
The problem isn't in the cheaper and smaller woofers. I'd take advantage of the original passive crossover
and adjust the sensitivity of the mid/supertw to the woofers. Passive parts are not expensive, leaving mkp,mkt
out of the equation.
and adjust the sensitivity of the mid/supertw to the woofers. Passive parts are not expensive, leaving mkp,mkt
out of the equation.
Beauty is in the eye of the beholder, I like that midtone. But to each his own. My original plan was to use a front loaded horn and 4x a 250mm woofer. But at about 1m³ cabinet that was a bit too much.
Now those woofers aren't as bad as the price makes out. The TIW may have a bit more than double the flux in the gap but it also has two-and-a-half times the moving mass. So concerning acceleration that is very similar. Main difference is surface. But the room isn't that large either at about 75m³.
I have adjusted the levels of the mid and high (around 6-9dB) and added 350mm delay to the woofers and high.
Really need to get some app running to do some "measurements".
I think you could make it work with those cheaper drivers. But without measurements...hardly.
You probably have baffle step loss around 380 Hz, if your baffle is 30cm wide.
Do you have separate volume control for each amplifier?
If you do, you could add inductor on lower bass driver, which would serve only as bass driver up to that frequency.
Don't know how many ohms do your drivers have, you will easily calculate it.
You probably have baffle step loss around 380 Hz, if your baffle is 30cm wide.
Do you have separate volume control for each amplifier?
If you do, you could add inductor on lower bass driver, which would serve only as bass driver up to that frequency.
Don't know how many ohms do your drivers have, you will easily calculate it.
Baffle is 34.4cm wide. There is separate volume control on each channel bass-mid-high.
But I agree, I need to get something to measure up and running. Sadly REW is out of action. And I don't know about anything else on Linux.
Can you cut it actively? If you can, cut the lower driver on 400Hz. Try a bit lower, and listen to results.
Hi Havoc, measure the system to weed out all humanly errors, which can be expected. Normal procedure for any speaker built. Then, if everything is as it should, you'd tweak "voicing", boost the bass / reduce the rest for example. This is very important to get right, not sure how sensitive others are but even ~1db difference in the balance is quite audible and could mean listening fatigue or not, and is also partially a personal preference and so on, just make it to your liking. Also, on the other hand anything goes, so it's just something you have to do for yourself. Too little bass? boost it. Too shrill top? cut it. Measurements are your friend here.
And, since you are in a room mind about room modes, listening middle of the room is about worst position, I have it too 😀 Add more sources to help with it, for which you need to measure and tweak the DSP again. For very least, cut the worst modal peaks with high Q filters, and rebalance the system if needed. A high Q room mode bass peak will mask and make the sound wrong, and when you cut it you might notice your balance is now off. Same for all resonances/peaks like possible box resonances, high Q peaks in the frequency response are the worst. I hope you get it sorted, have fun!
And, since you are in a room mind about room modes, listening middle of the room is about worst position, I have it too 😀 Add more sources to help with it, for which you need to measure and tweak the DSP again. For very least, cut the worst modal peaks with high Q filters, and rebalance the system if needed. A high Q room mode bass peak will mask and make the sound wrong, and when you cut it you might notice your balance is now off. Same for all resonances/peaks like possible box resonances, high Q peaks in the frequency response are the worst. I hope you get it sorted, have fun!
play with EQ. my Bose headphones take almost 30 decibels of correction to sound best.
real crossovers don't just divide frequency bands - they also shape the frequency response of the drivers.
in your case you will have to do that with EQ.
but keep in mind that depending on room acoustics you may not have enough power to overcome some of the suckouts in bass response.
you may have to add subs around the room.
without EQ most speakers will sound thin because of baffle step diffraction and other effects such as cone breakup and back EMF giving speaker a rising response.
furthermore target response of speakers isn't flat - it's actually slight rise in the bass, called "house curve"
play around with EQ. it's called "voicing" which means dialing in the sound of a speaker until it sounds right to the ear ...
real crossovers don't just divide frequency bands - they also shape the frequency response of the drivers.
in your case you will have to do that with EQ.
but keep in mind that depending on room acoustics you may not have enough power to overcome some of the suckouts in bass response.
you may have to add subs around the room.
without EQ most speakers will sound thin because of baffle step diffraction and other effects such as cone breakup and back EMF giving speaker a rising response.
furthermore target response of speakers isn't flat - it's actually slight rise in the bass, called "house curve"
play around with EQ. it's called "voicing" which means dialing in the sound of a speaker until it sounds right to the ear ...
Actually, one should never boost EQ trying to fill in the modal dips, because it's sound destroying itself (superposition) so no matter how much power you throw into it there is a dip. If you move yourself a bit out from the suckout you now get modal peak tere + the boost you just applied and made things way way worse any where else in the room, while not doing nothing for it a the listening position.
Simplest DSP trick one can do is narrow band dip the modal peaks, which helps at the listening spot quite a lot, and don't matter much at all anywhere else in the room. True fixes for modal problemos (the dips) include change of positioning, adding more bass sources to make the modal peaks and dips reduce statistically or just "counteracting" a mode with another source otherside of the room as in Double Bass Array, or manipulate room acoustics which is very expensive and tough job unless the building is specially built to manage bass modes, combination of any of these. DSP could at least in theory help with modes with some delay tricks and so on, but not sure how well these work, the dirac and stuff, perhaps they work oke.
It's a true issue for all small rooms though, so use any tricks that help with it. It's important to understand that this is not specifically a speaker issue but issue of a room. Speaker could have issues in addition.
Last edited: