Why wouldn't a ribbon based MTM be ideal for home theater

[JonMarsh]

The LCY has less than 1% distortion in a range that the ears are not sensitive toward. I wouldn't call that high distortion. It's higher than the domes, but having heard the LCY's myself, I question the meaningfulness of the distortion tests. Subjectively it sounds wonderful and has amazing dispersion. one might say that it sounds good because of its higher distortion, but i know how real instruments sound like and when i hear the cymbals from a ribbon that don't sound quite right from even the best domes, then that's a cymbal not a distortion!

Have you ever compared (measurring AND listening) the LCY to a comparably priced high performance dome tweeter in a similar design like the D3004/66000? Or AB'd them against electrostatic headphones? Now, the LCY's are much better than the lower cost planar leaf tweeters as regards distoriton, but while 2nd harmonic is rarely detectable at low levels, non-harmoniclly related IM such as this tweeter is prone to is another matter. It's general "noise floor" is not state of the art or representative of what I think a $200 tweeter should be.

Keep in mind I have experience with Ribbon tweeters since the Decca ribbon of the 70's and quite a few things since then, including the big Fountek's in arrays.

Nevertheless, regardless of my opinion, if you fancy building a system with the LCY's, no reason not to take a stab at it just because of my opinions.

Understand that when people come on the board seeking advice or opinions, I tend to err on the side of conservativism, and don't push my own notions or favorites. There are some fairl nice ribbons out there, but they're more than $300, and they have issues too which may render them unsuitable expect for carefully controlled music listening in a small seating position - witness the Stereophile review a year or so ago of the top of the line Areial system.

[Dennis H]

http://www.bobgolds.com/SmX720/20060708/home.htm

Interesting, Chris. Frankly I find it confusing as hell -- way too much scrolling involved. He's testing several things at once but it looks to me that his 'baffle' is doing more harm than good -- aside from reducing combing, it's really mucking up the tweeter's frequency response. But I could easily be wrong about that.

FWIW, some say the best answer for screen combing is a horn tweeter pressed right against the screen -- no parallel surfaces to reflect off.

[Dennis H]

Read and digest my earlier posts in this thread..

Thanks for the condescension. I believe I have a firm technical grasp of the subject at hand. You are certainly welcome to disagree and build a ribbon-based MTM if you think it will work well. Let us know how it sounds.

[Marzen]

I'm in a similar mindset as Greg; I want that airy ribbon sound. I'm going with Ruben's SMX screen material & I'm planning a ~24" false wall to hide the speakers. I would have bought the LCY 110 short ribbons today if I hadn't run into a PS3 last night by accident.
I'm working with dual RS225 (8"), RS52 (2" dome), and currently using a donated YAG20 planars (700hz & 4khz xo points). This 3 way has been a huge collection of compromises & problems like keeping sensitivity high (~90dB) due to a 4ohm nom/2.5 ohm min impedance used with an inexpensive AVR rated for 8 ohms. I chose a Qb of around .550 which is excursion limited as a full range, but works well with dual front subs crossed at 80hz as it's very close to a 2nd order Butterworth rolloff. Polar response has been a bear as well due to my screen size is 59" tall in a basement with an 84" ceiling. I turned the TMWW upside down to keep the tweet at ear level.
I can say that there's enough I like about it so far to keep plugging away at a solution. It was never meant to be a reference grade system so distortion was of limited concern. Just my 2 cents anyway... It will be interesting to see what pans out with Greg's theater.
Regards,
-Ward

[Piotr]

Thanks for the condescension. I believe I have a firm technical grasp of the subject at hand. You are certainly welcome to disagree and build a ribbon-based MTM if you think it will work well. Let us know how it sounds.

Maybe if you remember the thread started with a question of.. (guess..) a ribbon based MTM. h:

Your reply to me indicates that you lack technical grasp of the subject or that you did not read and understand the content of my posts. Even though you claim otherwise.

Besides that, the Von Schweikert and Burmester speakers seems to get good reviews.. as does the Griffin studio speakers.. Not that I have any plans on building something like this myself or asked someone else to do it either, but once again, the thread started witht this subject so maybe it's not so strange when you think about it that I adress design issues regarding this very subject.... or?

A MTM ribbon design will have LIMITED vertical dispersion and that is a fact, at least above 300-400Hz or so where the mids start to cancel each other of axis. Also a ribbon have a limited vertical dispersion depending on the height of the ribbon. Then why do you suggest that one should try to get good vertical dispersion around the crossover area?? Only to get narrow dispersion once again an octave or two up.

Maybe you are not familiar with the big designs of Dynaudio and therefore did not understand my earlier post? Dynaudio uses A WWMTTMWW layout in order to (you get it..) control the vertical dispersion. It's a design goal and you don't want a sudden change in the polar response as you touched earlier without really seeing the big picture.. Not that the Dyn's are super smooth but the example serves well to illustrate these types of designs.

The two tweeters in the big Dyn's are used in paralell for the lower part of the tweeter range guess why? Yup to REDUCE vertical dispersion even in the tweeter range. Both Esotar domes work in tandem at the crossing to the dual 5" mids and one is rolled of at the top.

Hope you get it now, otherwise let me know and I'll help some more.

/Peter

[cjd]

you're saying that you believe the differences between tests 5-11 should be audible?

No, didn't say that either, though I can understand how that could be implied. Simply that 8" is where the influence of the screen vs. distance is what seems to me to be the most balanced - screen influence is the most benign without making the required distance rediculous and too far impractical. I mean, felt diffraction control, rounded baffle edges, and so many other things we do when putting together a good design often make differences that aren't that different than what is being measured here.

C

[Dennis H]

Piotr, once again, thanks for the condesension. I suggest you model your proposed design and see how smooth the vertical polar response really is. When I do it with the Assymetrical Response Pattern Estimator from the FRD site, I get a really ugly (uneven) curve for the off-axis response when crossing as high as a ribbon needs. The so-called 'limited' vertical response is only at certain frequencies. If you can build one with a smooth curve, I'd love to see it and learn from you.

[Dennis H]

Picture, 1000 words, blah blah. MTM with 7" woofers and a 5" long ribbon, 2.5K LR4 XO. To answer the question posed by the OP, no a ribbon MTM is not 'ideal.'

[Jim Holtz]

Picture, 1000 words, blah blah. MTM with 7" woofers and a 5" long ribbon, 2.5K LR4 XO. To answer the question posed by the OP, no a ribbon MTM is not 'ideal.'

Hi Dennis,

I don't really want to get into the skirmish between you and Peter but you made a comment about ribbons and MTM's. I'm designing a W-M/T/M-W using, you guessed it, a ribbon as the tweeter. Now, no one that has any understanding of ribbons would use a 5" ribbon on any point source speaker unless you want the "head in the vice" listening experience. I'm using a Fountek NeoCD3.0 which has a ribbon element 2.3" long and has excellent vertical dispersion.

I did a little more research on the M-M vs M-T spacing conversation we had yesterday. You were exactly correct that moving the mids close together would in fact enhance off axis vertical dispersion but I also found out that doing that created an issue with horizontal off axis dispersion.

I listen off axis and horizontal is way more important than vertical to me so I'm back to a conventional MTM spacing format. I'm using a 4" mid in this design so it will be much easier to work with than the typical 6" or 7" woofer design. However, I've heard any number of those designs that sounded really excellent with both ribbons and domes as tweeters.

This speaker will be at the Iowa DIY event next fall so we'll all get to find out what 30+ people think of the sound. That's the true test, how they sound.

Jim

[Dennis H]

Hi Dennis,

I don't really want to get into the skirmish between you and Peter but you made a comment about ribbons and MTM's. I'm designing a W-M/T/M-W using, you guessed it, a ribbon as the tweeter. Now, no one that has any understanding of ribbons would use a 5" ribbon on any point source speaker unless you want the "head in the vice" listening experience. I'm using a Fountek NeoCD3.0 which has a ribbon element 2.3" long and has excellent vertical dispersion.

I did a little more research on the M-M vs M-T spacing conversation we had yesterday. You were exactly correct that moving the mids close together would in fact enhance off axis vertical dispersion but I also found out that doing that created an issue with horizontal off axis dispersion.

I listen off axis and horizontal is way more important than vertical to me so I'm back to a conventional MTM spacing format. I'm using a 4" mid in this design so it will be much easier to work with than the typical 6" or 7" woofer design. However, I've heard any number of those designs that sounded really excellent with both ribbons and domes as tweeters.

This speaker will be at the Iowa DIY event next fall so we'll all get to find out what 30+ people think of the sound. That's the true test, how they sound.

Jim

Hey Jim,

The short ribbon is a good choice if you want it to emulate a point source, i.e. not 'control vertical dispersion' as Piotr was suggesting. You may be able to pull it off well with the small mids, keeping in mind that a single small ribbon needs to cross pretty high. Compare the output of a couple inches of ribbon to the several feet you have in your line array and it makes sense to cross higher with the MTM.

I don't understand why moving the drivers vertically would affect horizontal response. Can you explain? Are you shifting the tweeter to the side to move the woofers closer? I can see how that could be a Bad Thing (Martha) for horizontal dispersion.

[Habs4life]

Are you shifting the tweeter to the side to move the woofers closer? I can see how that could be a Bad Thing (Martha) for horizontal dispersion.

I am planning to do exactly this with a horizontally oriented MTM center channel in an attempt to improve the horizontal off axis response.
What effect will placing the tweeter beside (above)the mid/woofs have on the vertical off axis response.The tweeter is a 1" dome in a small 2" dia chassis and the woofers are 6.5" with an LR4 at 2k.The woofer frames will be within 1" of each other.

All the info in this thread has me rethinking the MTM idea because a lot of listening will be from angles greater than 20 degrees.

[Dennis H]

I am planning to do exactly this with a horizontally oriented MTM center channel in an attempt to improve the horizontal off axis response.
What effect will placing the tweeter beside (above)the mid/woofs have on the vertical off axis response.The tweeter is a 1" dome in a small 2" dia chassis and the woofers are 6.5" with an LR4 at 2k.The woofer frames will be within 1" of each other.

All the info in this thread has me rethinking the MTM idea because a lot of listening will be from angles greater than 20 degrees.

I think that's a good thing to do with an MTM turned on its side. That's what Jon did with his MTM center. Horizontal response is more important than vertical response and the vertical response won't be all that bad. 2K is still a bit high with 6.5" woofers but it may work okay. I think Jon pushed his down around 1.4K with his fancy Cauer filters.

[Jim Holtz]

Hey Jim,

The short ribbon is a good choice if you want it to emulate a point source, i.e. not 'control vertical dispersion' as Piotr was suggesting. You may be able to pull it off well with the small mids, keeping in mind that a single small ribbon needs to cross pretty high. Compare the output of a couple inches of ribbon to the several feet you have in your line array and it makes sense to cross higher with the MTM.

I don't understand why moving the drivers vertically would affect horizontal response. Can you explain? Are you shifting the tweeter to the side to move the woofers closer? I can see how that could be a Bad Thing (Martha) for horizontal dispersion.

Hi Dennis,

Yes I was actually creating a triangle with the mids closely spaced and the tweeter to the inside also closely placed to the mids. I was told by my crossover guys that it was a matter of trade offs. When you offset the tweeter to get close spacing on the mids, this enhances the vertical dispersion but creates a similar dip in off axis horizontal response. Since I'm not a crossover guru, I can't provide any more explanation than that but I take it to be fact, considering the sources and their experience level.

Now, closely spacing the MTM in a vertical line is a good thing and helps integration etc. The small 4" drivers should integrate well with tight spacing.

I know the higher the crossover for a small ribbon is better but the NeoCD3.0 is quite robust and will cross comfortably at 2.5K I'm told. The 4" drivers can easily go beyond that so we'll see what the design ends up crossing at. There are always trade offs.

Stimulating and informative discussion for me. Thanks for your participation.

Best regards,

Jim

[Lindahl]

No, didn't say that either, though I can understand how that could be implied. Simply that 8" is where the influence of the screen vs. distance is what seems to me to be the most balanced - screen influence is the most benign without making the required distance rediculous and too far impractical.

Can you explain to me how you came to that conclusion from those tests? I really don't see it. The only tests I see, that compare distances, are tests 5-11 - which, as we both agree, shouldn't have audible differences.

OP, sorry for this hijack, that wasn't my intention.

[Piotr]

Picture, 1000 words, blah blah. MTM with 7" woofers and a 5" long ribbon, 2.5K LR4 XO. To answer the question posed by the OP, no a ribbon MTM is not 'ideal.'

And you think a dome MTM will produce a smoother over all response?

Actually the results on this graph are pretty good. I don't understand what the problem is? The average vertical response is actually pretty good. Especially the angles that will give the first reflection points in the ceiling and floor. I also doubt that the graphs are really realistic. The mid will not be as peaky/dippy in a real speaker and the ribbon response at 15degrees also looks strange.

Thanks for the modelling, it shows that I was right.

Oh, what listening/mic distance did you use in the modelling? Fiddling around with M-M distance and listening distance it should be possible to achieve even better results.

A MTM solution should really be listen at seated. So if the on axis curve are smooth and the average vertical curves are smooth, then all is good. Average or rather power response WILL be better/smoother than a dome MTM and most important is the spectral balance towards the first reflection points (vertical) since thoose will have the strongest effect on the perceived tonality.



/Peter

[cjd]

Can you explain to me how you came to that conclusion from those tests? I really don't see it. The only tests I see, that compare distances, are tests 5-11 - which, as we both agree, shouldn't have audible differences.

OP, sorry for this hijack, that wasn't my intention.

Sure. And hopefully such hijacks are good for all involved, including those just reading - ya learn something of what all everyone thinks about when putting something like this together, etc.

So.

First thing to note here is that the use of the rock wool (i.e. making sure the front wall is treated to absorb) is really important to some of the reflective issues that can occur, though I don't know that the specific implementation was ideal. You can compare 26 (treated) and 4 (reflective baffle) (they're placed adjacent in the first section, making this comparison easier). In both cases there is a little bit of SPL loss as well, hard to really be sure but I think ~4dB maximum rolloff starting around 5k on up, with only a couple dB loss up to 10k and another couple above that - not bad at all when you consider the circumstances.

From here I start noting specific areas that are still less than ideal - at 2" there is a suck-out centered around 5k, another at 10k - these seem to be 3-4dB, so theoretically very much audible if you know what you're listening for. The drop at 5k never really goes away - at 3" it's less specific, 4" it doesn't really change much behond that. However, the behavior around 10k is still pretty peaky, and that settles down at the 8" distance and doesn't really change significantly beyond that. It's those peaks that bother me, as they're up 4dB or so, I think. 2-3dB at 4".

Of course, there are issues around 5k in this setup no matter how this is spun, but I think that may be something between implementation in the speaker itself, and the specifics of what the rock wool application is doing to response and driver interaction.

I went and read the preamble just now, which actually supports some of the observations I made that led me to my conclusion. Also explains a bit about the reflective issues and why the rock wool was used. I might suggest a 3/8" wool felt application and the diamond shaped tweeter cutout for this, see previous work of Thomas and Jon for more on that.

C

[Paul W]

Greg,
My “ideal” HT speakers would be faceted ESL panels about 7’ tall and at least 3’ wide. For me, anything less than full range ESLs simply represents a different set of compromises. Unfortunately, without a dedicated room to “hide’ them, the ESLs would be unacceptably large…a domestic compromise that will not float.

Are ribbon MTMs ideal? No, they are just a combination of advantages and disadvantages that works well in my room. Though vertical lobing is a fact, reducing ceiling and floor reflections help tame a lively room. Horizontal pattern symmetry produces a very convincing lateral soundstage; vertical symmetry provides very stable vertical positioning. Do I believe ribbons are the ideal transducer? No, ESL is probably the best practical transducer.

I may replace my ribbon speakers, not for performance, but because they are too large for a revised room layout. Several different configurations have been tried, each with a different set of compromises…no “perfect” solution. (The one thing I have not been able to equal is vertical positioning.)

So, depending on the compromises you make, you could do far worse than a ribbon MTM.
Paul

[Lindahl]

Sure. And hopefully such hijacks are good for all involved, including those just reading - ya learn something of what all everyone thinks about when putting something like this together, etc.

I just realized why we came to such different conclusions. I'm analyzing the graphs on the far right, you're analyzing the graphs down the middle. The amount of smoothing done for the graph on the far right is similar to what the human ear hears (1/3rd octave smoothing). When comparing those measurements, you can't come to the conclusion whether the differences are bad or good - just different, and so small that they would be inaudible. In fact, the bass response is smoother the closer the speaker is to the screen. However, this is probably due to the changes in distance from the front wall. Looking at the middle graphs, I'd agree with you - except for the suckout at 5khz, which appears to be the least prominent at 3", but has a bit more combing in the highs (which the ear filters out, anyway).

Also, I didn't notice this until I was typing this reply, but the distance he's measuring from changes as the speaker's distance from the screen changes - which will change the combing characteristics, for sure. Overall, it was a good attempt, but needs a bit more control to truly be representative of the real waveform manipulations the screen performs at various distances.

Finally, looking at the far right graph in test 27 and 28, you can see that the use of a baffle doesn't improve the response, and only kills the bass response - which can probably be eliminated by using thinner absorption. He's using an absorptive baffle to eliminate combing that the ear can't hear anyway, at the expensive of the bass response!
:E

[cjd]

I understand the differences between the graphs. I don't agree that it's what I should be using to evaluate.

The ear/brain certainly doesn't take in data quite the same way a computer does, and has its own process that certainly can result in something akin to a smoothed response, etc. But the ear can also be sensitive to anomalies that do not show up in 1/3 octave smoothed response graphs, but DO show up in raw response. Particularly in the ranges we're talking about, where we derive a lot of enunciation detail, etc. So, while the far graph is useful for a brief overview, it doesn't really tell the story I want to understand, and I'm not sure it is the graph to be using to evaluate.

There are a lot of things Bob did well, and a lot of things he would probably do differently on a second or third attempt. Certainly things I would do differently (among them, I would probably get absolutely completely bored :lol: ) so we take the data we have, we ponder, we hem and haw, and... who knows.

C

[Lindahl]

The ear/brain certainly doesn't take in data quite the same way a computer does, and has its own process that certainly can result in something akin to a smoothed response, etc. But the ear can also be sensitive to anomalies that do not show up in 1/3 octave smoothed response graphs, but DO show up in raw response.

While I do agree that such anomalies do exist, combing is not one of them. The ear does an excellent job eliminating combing, at any frequency range. 1/3rd octave smoothing shows what the ear does in the presence of combing. If we depart in agreement here, so be it. However, you're correct that it does hide other problems the ear will pick up, such as deep nulls or huge peaks (the preamble of your link touches on this). I enjoyed the mental exercise.
:T

[cjd]

Yeah, I'm not sure the issues are combing induced, truth be told. More likely bounce issues.

[Lindahl]

Yeah, I'm not sure the issues are combing induced, truth be told. More likely bounce issues.

Umm... that is combing... very early reflections (bounces) creating a complex of series of peaks and dips.

[Piotr]

Which is audible..

/Peter

[cjd]

well then, I guess I think it's comb. Suppose that's how it would work if the reflection re-merges 180 out of phase on rebound. I was thinking bounce the way room nulls appear, which is probably still comb effect then, eh?

[indygreg]

http://www.bobgolds.com/SmX720/20060708/home.htm
You won't achieve the same results with the Modula and the behringer crossover as with the passive, though you'll definitely get something that works. If you're going that route, by all means go for it all around. You can then say "I want to go with ribbons, help me pick a good driver complement" and we'll dive in.
C

a noob question - why is this? i would think that a processor that has xover, parametrics, time delays for phasing would be able to do anything a passive xover would do.

i might reach a point where i say i want to do ribbons but first i want to hear the basics. i ordered parts to build some mtm's today (my inwall speaker) and i want to hear them. maybe the 29tbfc tweets will blow me away and this will be simple. at least i will have a reference point to start from.

greg

[Lindahl]

Which is audible..

Can be audible. Heck, even speaker grills cause combing. The audibility all depends on the size of the peaks and dips and the width of them. The larger the peaks and dips, and the larger the distance between them, the more audible it is. 1/3rd octave smoothing approximates what the human ear (or brain) does to combing, and therefore is a good graph manipulation to visualize whether or not combing will be audible.

I was thinking bounce the way room nulls appear, which is probably still comb effect then, eh?

Well, technically it is. However, I've usually seen the term "mode" used to refer to peaks and nulls caused by a room. These are (almost?) always in the bass and low midrange, due to the size of the wavelengths and are very difficult to treat without good bass trapping. Comb filtering is generally used to describe the same effect at the higher frequencies, which is rarely caused by room boundaries in a room that has decent acoustics. Due to the size of the wavelengths, it usually occurs when a material is placed between the listener and the speaker (as we have seen). A room with nothing in it except for speakers will exhibit serious comb filtering ("echoy"), but most rooms have couches, carpets, pillows, blankets, etc. that will eliminate these problems.

[Piotr]

Ok!

/Peter

[cjd]

a noob question - why is this? i would think that a processor that has xover, parametrics, time delays for phasing would be able to do anything a passive xover would do.

Well, you can probably get close, but it takes an awful lot of mucking about to get a zero sum filter into play. I think folks have come close stacking two notch filters, but I seem to remember them running out of filters to work with and needing a lot more crossover units to handle a simple speaker. You can get 8th order pretty simply, but Jon (and an ever increasing group of the rest of us) employs those nifty cauer-ellpitic zero sum whatsit thingamabob filters.

i might reach a point where i say i want to do ribbons but first i want to hear the basics. i ordered parts to build some mtm's today (my inwall speaker) and i want to hear them. maybe the 29tbfc tweets will blow me away and this will be simple. at least i will have a reference point to start from.

A great way to start. Something simple, just to hear what you hear and all that. I did a lot of tweeter listening at the local DIY event a couple years ago and it helped me immensely.

If you're up north of Chicago ever, give a shout, I may have time in my schedule and you can give a listen to whatever projects I happen to have around, if you're so inclined. Might even manage to bring a larger crowd along for the fun, a couple more speaker models to listen to...

C

[cjd]

Well, technically it is. However, I've usually seen the term "mode" used to refer to peaks and nulls caused by a room. These are (almost?) always in the bass and low midrange, due to the size of the wavelengths and are very difficult to treat without good bass trapping. Comb filtering is generally used to describe the same effect at the higher frequencies, which is rarely caused by room boundaries in a room that has decent acoustics. Due to the size of the wavelengths, it usually occurs when a material is placed between the listener and the speaker (as we have seen). A room with nothing in it except for speakers will exhibit serious comb filtering ("echoy"), but most rooms have couches, carpets, pillows, blankets, etc. that will eliminate these problems.

Comb issues can exist with nothing between the listener and the speaker as well. When I think of comb filtering, this is how my brain categorizes it. Boundary interactions (being room-sized when it comes to bass) don't generally slip into that category in my mind, whether they should or not. This is where my brain was making a distinction, whether accurate or not. The dip at 5k is very different than the bouncy output seen without the rock-wool treatment. And it's that bouncy bouncy trait that is most commonly equated with comb filtering specifically. Probably a muck-up of terms being far more complex than they're general given credit for being.

C

[Piotr]

Can



Well, technically it is. However, I've usually seen the term "mode" used to refer to peaks and nulls caused by a room. These are (almost?) always in the bass and low midrange, due to the size of the wavelengths and are very difficult to treat without good bass trapping. Comb filtering is generally used to describe the same effect at the higher frequencies, which is rarely caused by room boundaries in a room that has decent acoustics. Due to the size of the wavelengths, it usually occurs when a material is placed between the listener and the speaker (as we have seen). A room with nothing in it except for speakers will exhibit serious comb filtering ("echoy"), but most rooms have couches, carpets, pillows, blankets, etc. that will eliminate these problems.

Those are not the same effects. Comb filtering is indeed caused by room boundaries. Any reflection of sound is delayed and summed with direct sound and casues combfiltering. Normal decour of a room will not eliminate combfiltering from reflections, merely decrease the level to some extent by absorbtion and diffusing. You need to have 100% absorbtion to eliminate combfiltering i a room.

Room modes are a resonance that builds up between two (or more) surfaces. Even though this results in peaks and dips in the FR it is not the same thing as a reflection. A room mode has a Q, a decay time which a reflection does not have. A reflection only need one surface, a mode need two.

Combfiltering is almost always audible.

/Peter

[Lindahl]

Room modes are a resonance that builds up between two (or more) surfaces. Even though this results in peaks and dips in the FR it is not the same thing as a reflection. A room mode has a Q, a decay time which a reflection does not have. A reflection only need one surface, a mode need two.

Combfiltering is almost always audible.

Gotcha. What makes you think comb filtering is almost always audible? All the acoustic literature I've read has never made such a claim. Everything I've read has said that audibility is very dependent on the amplitude and Q of the combing. However, I have read that comb filtering from room boundaries is almost always audible, but only in lower frequencies due to the longer wavelength (Q of the combing). Is that what you mean?

[Piotr]

Comb filtering in a room is always due to delayed reflections and the ear/brain is sensitive to transients. If several reflections, especially early ones, are added to the direct sound the resolution is reduced since we end up listening to a mess of spikes arriving at the ears. Also the resulting peaks and dips will affect the balance between the harmonics in the recorded sound and thereby change it's timbre and tonality.

I thinks it's safe to say that combfiltering normally is audible. When reading books and articles it's important to really understand what the goal was with the study and how it was performed and so on.

Delay and reverb used in studios on voices and instrument is all about combfiltering. Part of the signal is delayed one or many times (like one or many reflections in the room) to create a new exciting sound with more feeling of space and dimension.

/Peter

[Rick Craig]

One of the misconceptions is that ribbons have less vertical dispersion solely because of the ribbon length. It also is a function of how the faceplate and motor design affect the response. Units like those from LCY, Fountek, and Aurum Cantus have horn loading which narrows the vertical coverage. I tested a tweeter from ATD and the faceplate was very shallow - vertical dispersion was quite good.

The models for driver interaction in a MTM don't always measure as predicted. Some of this I think can be attributed to the cone profile (shape) and voice coil inductance. While I have done a longer ribbon with a dual 7" MTM it was designed for a surround speaker with the vertical coverage not being a priority. The smaller ribbons can be crossed close to 2.5K but even with a 5" driver you may have some response anomalies.

Also keep in mind that a stated crossover point can be acoustical or electrical. One thing I like about LsPCad is that you can look at the electrical gain / transfer function of the filter to see how much energy is going into the tweeter. This is important to see when trying to cross a ribbon low. hope this helps.