Latest diy muddy sounding

[polkymon]

I originally planned on using some BA cabs I had to make a set of speaks using the Seas L18 and 27TBFC/g but I f'd up the cutouts. These were .5cf cabs and I was using the x-over from here;


http://www.zaphaudio.com/audio-speaker17.html


Anyways I always wanted to copy the Sonus Faber Concerto so I made these using the exact x-over as Zaph's project. The cab is .58 before braces or anything. The walls are 1 1/8" mdf and are covered with PE lightweight damping sheet and 1/2 carpet padding.

My problem is that they sound muddy to me. My reference is a set of Sonus Faber Guarneris I built awhile ago using Usher drivers. They sound a lot more live and transparent/airy. The Seas sound muffled.

What could I do to make them brighter?

[angsuman]

Well firstly great work they look amazing. Muffled? That's a very vague term. Well Zaph's design appears pretty flat. Maybe you just prefer a brighter treble. You could try less padding on the tweeter. Perhaps because the baffle is angled that you are off axis to the tweeter. So the treble is shelved down a bit. But I wouldn't expect it to be that much different.

[Zaph]

More people describe that design as bright than muddy.

The obvious things people normally mention are:
Get them away from walls
Check the crossover wiring
Check the tuning frequency and go lower if necessary

Looking at the picture, those are beautiful speakers. But unfortunately, I see a couple issues. (angsuman, you hit it on the head) This design was specifically created to be used on a flat baffle. The design listening axis is on or above the tweeter. If you listen below this axis, (with the baffle tilted back) you position yourself closer to a null that reduces midrange output. Since midbass output is not affected and still at the reference level, the resulting sound could be described as mud.

This image shows the design listening axis on a flat baffle at the crossover frequency. If you tilt the speaker back, you can see how the null will affect the response.



Try tilting the speaker forward and see if that brings up the midrange to match the midbass. If it doesn't, there's probably other issues. But if it does, we know the problem. You may be able to compensate for this by relaxing the tweeter rolloff by one order. The shallower slope, which essentially makes the rolloffs assymmetrical in the opposite direction, should bring the below-axis phase closer to alignment and ease the null.

The other thing I see, which may be minor, is the port on the front next to the tweeter. This normally causes a cavity effect of a sharp peak/dip in the response. The effect isn't normally described as muddy however. It's more edgy / less defined. I couldn't say for sure how much of an issue that is.

John

[polkymon]

I should have mentioned that yes I do prefer a brighter speaker. When I a-b these though to my Ushers the difference is huge. Others who listen agree that the Seas just sounded flat and the Ushers more live.

I know the Ushers cost much more but I am thinking of putting them in this latest project if I can't get them to sound better. The Usher tweeter for $60 is incredible imho.

[polkymon]

First of all thanks John for your response. By previous threads I know your time is limited on these forums.

I will try tilting to see how much the effect is. As far as adjusting the x-over that is beyond my scope. Typically I choose drivers, build an enclosure and give the specs to Madisound or PE for x-over design. So my knowledge is limited in that regard.

The port is probably not in the optimal position but I was trying to stay true to the original (which was an older design).

Also I was surprised at the difference in sound level compared to the Ushers. The Seas seem much less efficient.

Don't think I was doubting your design as I am sure it is excellent. The problem lies in my adaptation. I try to avoid building a simple box-like cabinet (see link) which gets me into trouble sometimes.



http://s37.photobucket.com/albums/e5...al&media=image

[ermes]

My reference is a set of Sonus Faber Guarneris I built awhile ago using Usher drivers.

attention polkymon!
sonus-faber never used and use usher drivers : you have only built a simply usher two ways.

[polkymon]

Originally posted by ermes


attention polkymon!
sonus-faber never used and use usher drivers : you have only built a simply usher two ways.

I never knew or cared which drivers they used (Dynaudio??). I only copied the cabinet design and refer to it by their model name to be more specific.

Not meant to be a clone only a facsimile.

see pic.

[polkymon]

Originally posted by angsuman
Well firstly great work they look amazing. Muffled? That's a very vague term. Well Zaph's design appears pretty flat. Maybe you just prefer a brighter treble. You could try less padding on the tweeter. Perhaps because the baffle is angled that you are off axis to the tweeter. So the treble is shelved down a bit. But I wouldn't expect it to be that much different.

Thanks for the response. I don't design x-overs but I do build them. If you could explain further about changing the padding I'd appreciate it.

[angsuman]

Well padding refers to the resistors that lower the output of the tweeter. By lowering the value of these if its a series resistor you can increase the output of the tweeter. If you look at Zaph's crossover you'll see the R9 and R10 values fiddling with these will change the output of the tweeter...though i'd have to use a simulation program to see what the changes will be.
However Zaph pointed out something different. Its not neccessarily the tweeter's level in general but rather when you go off axis vertically there is a 15dB suckout at the crossover frequency. If you look at that graph. you'll see that when the tweeter is on axis or slightly up axis (basically the tweeters below you) the frequency response is correct and flat. But not imagine tilting that graph counterclockwise...which is probably close to what results from your tweeter being angled back. Which is approx a 15dB dip at 2khz. So this might be what you mean by muddy maybe vocals or strings just don't sound strong and upfront. The ear is very sensitive in this region.

Lowering the slope by one order which will make it 1st order electrical rather than 2nd order electrical will change its vertical off axis characteristics for the better since Zaph mentioned it. I think there was a graph somewhere online that showed how different orders affect the vertical response. The tweeter will also contribute more in this frequency range since the slope is shallower and you might find that you prefer the sound. Since you don't have much experience designing xovers basically you get rid of the tweeter inductor to make it first order...but you'll also have to change the capacitor's value so basically a new crossover.

[polkymon]

Originally posted by angsuman
Well padding refers to the resistors that lower the output of the tweeter. By lowering the value of these if its a series resistor you can increase the output of the tweeter. If you look at Zaph's crossover you'll see the R9 and R10 values fiddling with these will change the output of the tweeter...though i'd have to use a simulation program to see what the changes will be.
However Zaph pointed out something different. Its not neccessarily the tweeter's level in general but rather when you go off axis vertically there is a 15dB suckout at the crossover frequency. If you look at that graph. you'll see that when the tweeter is on axis or slightly up axis (basically the tweeters below you) the frequency response is correct and flat. But not imagine tilting that graph counterclockwise...which is probably close to what results from your tweeter being angled back. Which is approx a 15dB dip at 2khz. So this might be what you mean by muddy maybe vocals or strings just don't sound strong and upfront. The ear is very sensitive in this region.

Lowering the slope by one order which will make it 1st order electrical rather than 2nd order electrical will change its vertical off axis characteristics for the better since Zaph mentioned it. I think there was a graph somewhere online that showed how different orders affect the vertical response. The tweeter will also contribute more in this frequency range since the slope is shallower and you might find that you prefer the sound. Since you don't have much experience designing xovers basically you get rid of the tweeter inductor to make it first order...but you'll also have to change the capacitor's value so basically a new crossover.

Been busy finishing up the project. Ended up with brown hammered finish on sides. They look great now I have to dial in the x-over.

I tilted the speaker and had some improvement but not enough to keep them in that awkward position. They will be on shorter stands anyway due to the canted baffle. I may try swapping out the resistors after I do some research.

Just for giggles I changed the x-over to a generic PE 2500hz I had laying around and ended up with a rather harsh sounding tweeter. Kinda thought I would. I am not adverse to building a new x-over. If I can't figure it out I may let Madisound LEAP design it for me.