What Makes a Loudspeaker Musical?

[Spannko]

Thanks for your input Rutger 👍

Do you mind me asking if what you’re saying is based on personal experience, or on something you’ve read or been told?

To update my previous post, I’ve been experimenting with different enclosure materials, and the results have been a bit of an eye opener, to say the least! Identically sized baffles made out of different materials have their own unique timbre and effect on the music. The best material I’ve found so far is poplar. It’s not a material I’ve seen used for loudspeakers, possibly because although it’s classified as a hardwood, it’s actually quite soft and easily dented. The second best is maple, followed by American walnut, although I’d possibly say that walnut is bordering on not being quite good enough and even maple is bordering on just good enough. Fortunately, poplar is quite musical. So what’s the very worst material I’ve tried? MDF!!! As used, it has absolutely no musical qualities whatsoever. This may be a slight over exaggeration, because it actually gets worse as its thickness increases, so thinner mdf must have a degree of musicality, but in comparison to any other material, it’s ‘musicality’ is absolutely inaudible. Second worst? Birch plywood!!! I can only conclude (so far) that mdf and plywood are used for loudspeaker enclosures because they’re regular, easy to machine, stable over time and cheap.

Previously, I suggested that the crossover and drive units were possibly more important than the enclosure. The results of my latest experiments suggest that they’re all equally important. As is usual in hifi, everything matters!

[lindsayt]

If only 18mm void-free birch plywood was cheap! Or even available (in the UK)!
One advantage of birch plywood for DIY projects is that, in my opinion you don't need to veneer it to look acceptable. I think that birch plywood has a natural beauty to it.

The joint most musical sounding speakers I've heard, had cabinets made from (painted) 25mm chipboard.

I agree on MDF being unmusical.

Something that's worth trying, from the reports I've had, is lining the speaker cabinets with steel plates. I guess aluminum plates might also be worth a try (and easier to machine).

There may also be a horses for courses element. With the optimum cabinet materials depending on the drivers, the type of cabinet, the details of the cabinet construction, including bracing.

[Spannko]

You can get 18mm birch ply here: https://mdfdirect.co.uk/product/birch- ... size/18mm/

Joe Akroyd used steel sheeting in his top line 7 litre loudspeaker (Eden?). It produced a very refined sounding speaker, but sadly lost the musicality of the Coniston R’s.

In my experiments, chipboard was slightly better than mdf and on a par with birch ply. However, decent chipboard in multiple thicknesses is very difficult to find and I’m not convinced that the sample I tried was that good. A decent, branded flooring grade might be worth trying though, so I’ll get some and give it a listen.

[Rutger]

Thanks for your input Rutger 👍

Do you mind me asking if what you’re saying is based on personal experience, or on something you’ve read or been told?

To update my previous post, I’ve been experimenting with different enclosure materials, and the results have been a bit of an eye opener, to say the least! Identically sized baffles made out of different materials have their own unique timbre and effect on the music. The best material I’ve found so far is poplar. It’s not a material I’ve seen used for loudspeakers, possibly because although it’s classified as a hardwood, it’s actually quite soft and easily dented. The second best is maple, followed by American walnut, although I’d possibly say that walnut is bordering on not being quite good enough and even maple is bordering on just good enough. Fortunately, poplar is quite musical. So what’s the very worst material I’ve tried? MDF!!! As used, it has absolutely no musical qualities whatsoever. This may be a slight over exaggeration, because it actually gets worse as its thickness increases, so thinner mdf must have a degree of musicality, but in comparison to any other material, it’s ‘musicality’ is absolutely inaudible. Second worst? Birch plywood!!! I can only conclude (so far) that mdf and plywood are used for loudspeaker enclosures because they’re regular, easy to machine, stable over time and cheap.

Previously, I suggested that the crossover and drive units were possibly more important than the enclosure. The results of my latest experiments suggest that they’re all equally important. As is usual in hifi, everything matters!

Its from personal experience , I have tried many things :)
Nice to read about your findings.
One can easily hear musical differences with different enclosure material using tunemethod.
The same clear differences can be found using different dampingmaterial inside the box. I have found that sheep wool is the best material, and its because it dont act as a spring, making the airpressure inside the box unlinear.

It was about 25 years ago I did learn that different thickness of the walls in a loudspeaker made a difference for the better in my own DIY speakers, spreading the resonanses at different frequencies and avoiding any tunefork effect. Later I recognized that Rod Crawford had been using this technique many years before that.

[Spannko]

Thanks Rutger. That makes your findings doubly interesting!

Internal damping is something I’ve really struggled to get to work well. However, I think my methodology may not be very good so it’s not been possible to come to any definitive conclusions. Tuning the enclosure volume without filling is relatively straight forward, but as soon as I add even the slightest amount of filling it seems to reduce the speakers musicality. I think I need to do the experiments again, but this time be more rigorous in how I maintain filling density as the enclosure volume changes. It’s not particularly easy, because (as you know) changing filling density alters the enclosures apparent volume which will de-tune the bass unit loading so it’s easy to end up going round in circles! Have you tried making speakers without any filling?

I’ve seen manufacturers use thicker baffles or rear panels, sometimes just screwing them into place. Although I’ve not done any serious experiments with panel thickness and attachment methods, I’ve noticed that glueing a baffle in place can make a speaker less musical, possibly due to more energy being transferred into the main enclosure. Is this anything you’ve tried too?

[Rutger]

Thanks Rutger. That makes your findings doubly interesting!

Internal damping is something I’ve really struggled to get to work well. However, I think my methodology may not be very good so it’s not been possible to come to any definitive conclusions. Tuning the enclosure volume without filling is relatively straight forward, but as soon as I add even the slightest amount of filling it seems to reduce the speakers musicality. I think I need to do the experiments again, but this time be more rigorous in how I maintain filling density as the enclosure volume changes. It’s not particularly easy, because (as you know) changing filling density alters the enclosures apparent volume which will de-tune the bass unit loading so it’s easy to end up going round in circles! Have you tried making speakers without any filling?

I’ve seen manufacturers use thicker baffles or rear panels, sometimes just screwing them into place. Although I’ve not done any serious experiments with panel thickness and attachment methods, I’ve noticed that glueing a baffle in place can make a speaker less musical, possibly due to more energy being transferred into the main enclosure. Is this anything you’ve tried too?

If the baffle is screwed in place it might sound good If you are lucky but than there is always a risk with a main resonanse occuring exactly at the frequency of a real tone that the musicians are playing. Using only glue with pressure seems to be more even in the forces occuring in the cabinet. I have tried to make a whole cabinet heavier putting thin extra walls inside of the cabinet and the sound became less musical.
I then only used extra thickness on each opposite wall , and this time the sound was better than a cabinet without extra thickness on the walls. I was spreading the resonanse frequencies from the different walls.This was repeated in two different DIY constructions with the same result, one was a closed box speaker and the other was a bassreflex loaded.

Then we have the dampingmaterial - some of it just makes the cabinet slightly virtual larger inside ( the usual white wadding ) and some is doing good damping above 100 Hz . There is always a risk of drawing wrong conclusions and go round in circles so one must try in each case. To much sheep wool will kill the musicality so one has to try different amounts of damping material. Metal cones or very rigit ones seems to be more immune to the sound coming delayed from the inside of the cabinet and through the cone, so slightly less damping material might be used in such cases - but again, no rule of thumb .

The resonanses from a cabinet is often very real and audible and If they come exactly at the same frequency as one of the left hands tones played by a piano , one will hear it and not in a good way.

I have also come to the conclusion that for 2 channel playback, the loudspeakercabinett shouldnt be totally dead because the very primitive stereosystem with all its faults. Some resonanses in the right places can make the sound more believable. Rigid spikes screwed into the base plate of the cabinet and with the sharp tip of the spikes into the floor often makes it easier to follow the tune of the bassplayer in a recording, you get somewhat ” less amount of bass” with spikes but in reality the bass tones are less smeared and more in tune . This is definitely true if you have a very rigid floor ( concrete )

Unrigid floors are more problematic and Linn sheets seems to lessen the bad effects of a springy floor.

Im happy to read about your findings because there are not much written about the art of making a good loudspeaker cabinet. I also believe that there is a lot of knowledge about how to do it at Linn, but also at the store tonlaeget where the constructor seems to have tested many things with his Klongedang speakers. Maybe he is active on this forum and can give us some more information ?

[Spannko]

I agree Rutger, there’s practically nothing written about how to build a musical loudspeaker. I was asked to recommend a book, and in all honesty, I couldn’t recommend one. They’re all just a collection of loudspeaker research, none of which focuses on a loudspeaker’s ability to play something as simple as rhythms and tunes. Similarly, all of the diy loudspeaker forums focus on a loudspeaker’s technical measurements. As tuneistas, we have to do our own research.

Unfortunately, it gets worse! Drive unit manufacturers don’t appear to design for musicality, and despite what their marketing departments claim, crossover component manufacturers don’t either. Every single drive unit, capacitor, inductor and resistor has its own level of musicality, and from what I’ve heard, nearly all of them are pretty poor. It’s no wonder that practically all loudspeakers on the market fail to reproduce music in a wholly convincing manner.

From a source first perspective, the crossover could possibly be the most important element, yet there’s practically nothing written about crossover design. The general approach to designing a loudspeaker is to pick drive units with good looking specs (whatever they may be!), use a computer program to design the “optimal” enclosure (whatever that is!) and then use a crossover simulator to find a combination of components which will somehow compensate for drive unit and enclosure errors and miscalculations. From a source first perspective, this appears to be total madness. If the crossover was designed first, it appears to me that there would be only one topology which would let the signal through without destroying its harmonic integrity: a series connected, first order dual filter network, consisting only of one capacitor and one inductor. I appreciate that this is a bold statement and that it places considerable demands on the drive units and may produce the type of sound people are unaccustomed to, but it seems to me that conventional loudspeaker design methodology only works very occasionally, and only then by accident. Even the more knowledgeable manufacturers struggle to consistently produce musical loudspeakers. Time will tell if I’m on the right track, but the initial results have been pretty encouraging!

[matthias]

If the crossover was designed first, it appears to me that there would be only one topology which would let the signal through without destroying its harmonic integrity: a series connected, first order dual filter network, consisting only of one capacitor and one inductor.

Completely agree, Spannko, this is my "dream XO" as well. With such a XO there is no need for "active" solutions at all.

[Rutger]

I agree Rutger, there’s practically nothing written about how to build a musical loudspeaker. I was asked to recommend a book, and in all honesty, I couldn’t recommend one. They’re all just a collection of loudspeaker research, none of which focuses on a loudspeaker’s ability to play something as simple as rhythms and tunes. Similarly, all of the diy loudspeaker forums focus on a loudspeaker’s technical measurements. As tuneistas, we have to do our own research.

Unfortunately, it gets worse! Drive unit manufacturers don’t appear to design for musicality, and despite what their marketing departments claim, crossover component manufacturers don’t either. Every single drive unit, capacitor, inductor and resistor has its own level of musicality, and from what I’ve heard, nearly all of them are pretty poor. It’s no wonder that practically all loudspeakers on the market fail to reproduce music in a wholly convincing manner.

From a source first perspective, the crossover could possibly be the most important element, yet there’s practically nothing written about crossover design. The general approach to designing a loudspeaker is to pick drive units with good looking specs (whatever they may be!), use a computer program to design the “optimal” enclosure (whatever that is!) and then use a crossover simulator to find a combination of components which will somehow compensate for drive unit and enclosure errors and miscalculations. From a source first perspective, this appears to be total madness. If the crossover was designed first, it appears to me that there would be only one topology which would let the signal through without destroying its harmonic integrity: a series connected, first order dual filter network, consisting only of one capacitor and one inductor. I appreciate that this is a bold statement and that it places considerable demands on the drive units and may produce the type of sound people are unaccustomed to, but it seems to me that conventional loudspeaker design methodology only works very occasionally, and only then by accident. Even the more knowledgeable manufacturers struggle to consistently produce musical loudspeakers. Time will tell if I’m on the right track, but the initial results have been pretty encouraging!

Yes, its almost that the most musical speaker should be a speaker without any crossover .

One more thing I want to contribute to this thread and its something I have recently discovered.
I have always wondered why the bass quality of the Linn Keltik was so good , and why some of the successors sounded less good at some points. In my opinion, its a very bad idea to have a subwooferdriver placed very near the floor, because you excite the whole room more than having it higher up from the floor. Some might say floor level placement of the driver could be a good thing, but if a clear pitch is the goal, it probably isnt.

So, using two very expensive active loudspeaker with subwoofer, I raised up the subwoofer so it was placed about 50 cm above the floor and the soundquality regarding a clear pitch became much better.
This is a rather good video:

https://youtu.be/S_GwaGeWQOg?si=iQa9Lwc5QR1IOrJN

( he explains why its a problem with floorstanding subwoofers, but he knows nothing about the tunemethod , so he makes some wrong solutions )

Later on, with two stereo DIY subwoofers, I recognized that two subwoofers sounds the best at the exact spot as the main speakers are placed, If using tunemethod in installation. And the subwoofers sounds even better 50 cm above the floor . - exactly the way Keltik or Isobarik was made.
Conlusion : A subwoofer should be inbuilt in the main speaker cabinet and not placed at floor level.

Drawing the conclusion of some of this , my last DIY speakers was a very simple construction that had the basstube on the backside of the cabinett, about 50 cm above the floor. The sound didnt need any roomcorrection at all and this became my best construction yet.

It might be a very good idea also ( in a bass reflex speaker ) to tune the speaker below the musical tones area, ie below lowest E from a bassguitar, 41 Hz . You easily mess up the tune If its right were the musical content is. In this case, its probably better if the tuning is somewhere between 34-36 Hz.
To complicate things, the box volume must follow the chosen tuning frequency, a lower frequency usually needs a bigger a box. And different drivers needs different box volume.

Another thing happened some time ago, were my friend who has a very expensive system ( Linn exakt with klimax source ) had done a tunemethod installation off his speakers and the inside of the speakers front baffle was placed 43,5 cm from the frontwall. This was very strange because I had done the same installation at my place the week before with different speakers and source, and they also happend to be placed 43,5 cm from baffle to the backwall. Coincidence ?

If there is something in this findings, technicaly you do change the SBIR frequency when one moves speakers from the frontwall and 43.5 cm baffle-frontwall might be the optimal frequency to make the stereo system faults as ”good” as it can be and the sound most tuneful , thus making the illusion better ? This might be the wrong conclusion and maybe this is just a coincident.

Conventional wisdom also tells us that we can only hear 20-20000 Hz .

I think this is wrong - we hear from 1 Hz to 20000 Hz , but the lowest frequencies are ”heard” by the body , feeling the rythm of the beat.

A guitar cord thats been struck has a very short beginning of very low frequencies right before the string starts to resonate at a higher frequency. This is probably why one can hear that a HP filtering beginning at 20 Hz in an amplifier is bad for the sound, and why one can experience that a HP filtering of 3 Hz sounds better with less blurring of the guitar pitch. I know this is my opinion and I have heard some loudspeakers (Linn ) that sounded better with a HP crossover right below the tuning point , but If one use driveunits that can take the punishment without sounding bad below the drivers fs, the sound might get better If the amplifier is close to DC in response.

As you already know, a passive loudspeaker, unconnected and placed in the listening room will colour the pitch of the tones in the music because of sympathetic resonanses. So when doing fine tuning of damping material in a DIY speaker, its important to put away all other passive speakers in the room.

https://en.wikipedia.org/wiki/Undertone_series

More:

https://en.wikipedia.org/wiki/Sympathetic_resonance

[lindsayt]

Thanks Rutger. That makes your findings doubly interesting!

Internal damping is something I’ve really struggled to get to work well. However, I think my methodology may not be very good so it’s not been possible to come to any definitive conclusions. Tuning the enclosure volume without filling is relatively straight forward, but as soon as I add even the slightest amount of filling it seems to reduce the speakers musicality. I think I need to do the experiments again, but this time be more rigorous in how I maintain filling density as the enclosure volume changes. It’s not particularly easy, because (as you know) changing filling density alters the enclosures apparent volume which will de-tune the bass unit loading so it’s easy to end up going round in circles! Have you tried making speakers without any filling?

I’ve seen manufacturers use thicker baffles or rear panels, sometimes just screwing them into place. Although I’ve not done any serious experiments with panel thickness and attachment methods, I’ve noticed that glueing a baffle in place can make a speaker less musical, possibly due to more energy being transferred into the main enclosure. Is this anything you’ve tried too?

It sounds like you have the answer already. Don't use any damping at all! Or at most, try a "curtain" about half way down the speaker to "break up standing waves".

[Spannko]

Thanks lindsayt. I’m not entirely convinced I’ve got any answers tbh, mainly because I’m not entirely convinced I understand the questions!